Welcome to Nuclear Frontier #4, where we take another walk through the latest news in the ongoing Nuclear Renaissance and the birth of the Fusion Industry.

First off, sorry about the gap in publishing time since my last newsletter. My main excuse is I’ve got an ongoing project, a company that’s currently in stealth, which we’re hoping to reveal by the end of the summer. It’s going to be pretty cool, so stay tuned. As I’ve mentioned, I’m also changing the names for these shorter topical newsletters. They’re focused on the technological edges shaping this century and beyond, with this one tracking advances in fission and fusion.

As one might expect, there has been another big stretch of news, and the pattern is getting clearer: nuclear is moving from “people are interested again” to “things are actually being built.” That’s true for large traditional reactors, small modular reactors, advanced reactors, fuel supply chains, nuclear for data centers, nuclear for industry, nuclear for space, and fusion too.

The biggest theme this month is execution. Licences are being renewed. Construction is starting. Fuel facilities are moving forward. Reactor companies are going public. Governments are building national strategies. Private companies are signing contracts with customers who need real power, not vague climate vibes. Concrete is getting poured, supply chains are being built, and reactors are starting to split (and fuse) atoms.

Alright, let’s dive in.

Subscribe now

“Good ideas are not adopted automatically. They must be driven into practice with courageous impatience.”

—Admiral Hyman G. Rickover

Nuclear Fission

Let’s start with Kairos Power, which broke ground on the Hermes 2 Demonstration Plant in Oak Ridge, Tennessee. This is a big one. Hermes 2 will be Kairos’s first commercial scale reactor and the first power producing (non-demonstration) Gen IV reactor to receive a construction permit from the U.S. Nuclear Regulatory Commission. It will supply up to 50 MW of electricity to the Tennessee Valley Authority grid, helping power Google’s data centers in Tennessee and Alabama. (Source: World Nuclear News)

What I love here is the iteration. Kairos is already building Hermes 1, a lower power demonstration reactor that won’t produce electricity, and Hermes 2 builds directly on that. This is exactly the approach nuclear needs more of: build, learn, improve, repeat. Less paper-reactor dreaming, more hardware in the ground.

Also notable is how openly Kairos is targeting the historical problem with nuclear construction: cost and schedule. Hermes 2 will use modular construction methods, precast concrete, and a seismically isolated foundation, while equipment modules will be manufactured in Albuquerque and shipped to Oak Ridge. That sounds boring, but it’s actually the whole game. Nuclear needs repeatable builds, not bespoke megaprojects every time. (Source: World Nuclear News)

In California, Diablo Canyon received a 20 year licence renewal from the NRC (Nuclear Regulatory Comission), clearing its two reactors to operate into the mid-2040s. Diablo Canyon provides nearly 9% of California’s electricity and 17% of its clean energy, which makes the previous plan to shut it down even more insane. Thankfully, California had a brush with energy reality during its 2022 heatwave and started changing course. (Source: World Nuclear News)

There is still a catch: extending Diablo Canyon past 2030 will require action from California’s legislature. But the NRC piece is now done, and that matters. It’s always better not to blow up your own energy infrastructure, even if it takes a few years for everyone to come around to that rather obvious point.

The UK is moving ahead too, with Rolls-Royce SMR formally signing its contract with Great British Energy – Nuclear to deliver the UK’s first SMRs. The first project will be at Wylfa on Anglesey in North Wales. Rolls-Royce SMR is expected to build three units in the UK, each generating enough electricity to power around one million homes. (Source: Construction Enquirer)

That’s another important milestone because the Rolls-Royce approach is very much about standardization and factory buildability. Again, that’s the key. The future of nuclear is not just better reactors, it’s better reactor manufacturing.

China, of course, continues to do China things. Unit 1 of the San’ao nuclear power plant has entered commercial operation. It’s the first of six planned Hualong One reactors at the site in Zhejiang province, and it entered commercial operation five years and four months after first concrete. Once all six units are complete, the site is expected to provide more than 54 TWh of electricity annually. (Source: World Nuclear News)

Five years and four months from first concrete to commercial operation. That’s the kind of timeline the West needs to relearn. We can debate systems, financing, regulation, labour costs, and all the rest, but the basic point remains: China is building nuclear plants at scale, and they’re doing it quickly. Good for them. I just hope we catch up.

South Korea is also moving forward, with Unit 3 at the Saeul nuclear power plant achieving first criticality. The reactor is an APR-1400, and once fully operational later this year, it will account for about 1.7% of South Korea’s total electricity generation and 37% of Ulsan’s electricity demand. (Source: World Nuclear News)

The APR-1400 is one of the most important reactor designs in the world right now. South Korea has built them at home and abroad, including the four reactors at Barakah in the UAE. The West doesn’t need to invent everything from scratch; in many cases, the designs already exist. The challenge is building them again and again.

Up in Canada, Bruce Power has entered into a collaboration agreement with Energy Alberta. Bruce Power will share nuclear knowledge, planning experience, regulatory lessons, and stakeholder engagement frameworks as the province of Alberta continues exploring nuclear power. Energy Alberta has proposed building two to four Candu Monark reactors in the Peace River area, and has also been looking at deploying the Westinghouse AP1000 in Alberta. (Source: World Nuclear News)

This is especially interesting because Alberta has no nuclear power today, but it absolutely should (and will). The province has a huge industrial energy demand, a sophisticated energy workforce, and a grid that would benefit enormously from reliable baseload power. Not to mention delibrately courting hyperscalers who want to build AI datacenters. Nuclear in Alberta just makes sense. It’s nice to see the pieces slowly moving into place.

Bruce Power also recently signed a similar agreement with SaskPower as Saskatchewan evaluates large nuclear technologies. Saskatchewan is already moving ahead with SMR plans, but looking at large reactors too is a good sign. We shouldn’t artificially limit ourselves to “small” reactors just because they’re currently fashionable. Big reactors work, and sometimes big reactors are exactly what you want.

The Netherlands has a very interesting molten salt reactor story too. Thorizon and a group of Dutch partners have signed an agreement laying out a path from a non-nuclear molten salt test facility by 2027, to a nuclear demonstrator by 2028, and then a first commercial Thorizon One reactor in Zeeland by 2034. Thorizon One would be a 250 MW thermal / 100 MWe molten salt reactor aimed at industrial customers and utilities. (Source: World Nuclear News)

The really cool part is that Thorizon’s molten salt fuel is designed to use a combination of long lived elements from reprocessed used nuclear fuel and thorium. In other words, it’s not just producing power — it’s potentially turning part of the “nuclear waste problem” into fuel. The company also uses replaceable reactor core cartridges, which it says help address corrosion and used fuel handling issues. (Source: World Nuclear News)

Speaking of nuclear fuel, FluxPoint Energy has entered the race to build what could be the first new U.S. uranium conversion plant in nearly 70 years. Conversion is the step where uranium oxide is turned into uranium hexafluoride before enrichment, and the U.S. currently has just one commercial conversion facility. FluxPoint is planning a modular Texas facility, with first production targeted for 2030–2031. (Source: POWER Magazine)

This is less flashy than a new reactor, but arguably just as important. You can’t have a nuclear buildout without fuel. The U.S. has been putting real money into enrichment, including LEU and HALEU, but conversion is another chokepoint. The nuclear renaissance needs mines, conversion plants, enrichment, fabrication, recycling, and reactors. The whole chain matters.

Nuclear is also moving forward on the financial side, with X-energy submitting a draft registration statement for an IPO under the ticker XE. The company is one of the most serious advanced reactor players, developing both its Xe-100 SMR and its TRISO-X fuel business. Going public would give investors another way to participate directly in advanced nuclear, which is something I think we’ll see much more of if the industry keeps progressing. (Source: X-energy)

X-energy also had a big international supply chain announcement during the Indo-Pacific Energy Security Summit. X-energy and Doosan Enerbility (a South Korean heavy industrial company specializing in energy solutions) signed a binding agreement to manufacture 16 main power systems for the Xe-100, while Doosan committed to building what the U.S. government described as the world’s first dedicated SMR fabrication facility (Aalo Atomics would dissagree, see below). The same announcement said X-energy’s commercial pipeline is 11 GW, equivalent to 144 advanced reactors. (Source: U.S. Department of the Interior)

That is exactly the kind of thing we need to see. Not one reactor. Not one site. An orderbook. A supply chain. Manufacturing capacity. Repeatability. That’s how nuclear becomes boringly normal again, which is ultimately what we want.

Aalo Atomics has completed assembly of its experimental reactor at Idaho National Laboratory. The company expects the reactor to achieve criticality well before its July 4 deadline. The test reactor is the precursor to the Aalo Pod, a 50 MWe modular nuclear power plant designed specifically for data centers. Each Aalo Pod would contain five factory built, sodium cooled Aalo-1 reactors, with commercial use targeted for 2029. (Source: World Nuclear News)

Aalo is only a few years old, which makes the speed here impressive. As I’ve said before, the AI boom is going to need power. A lot of power. If companies like Aalo can really productize nuclear into factory-built modules for data centers, that’s a huge market.

The U.S. Department of Energy is also putting money into used nuclear fuel recycling, awarding more than $19 million to five companies: Alpha Nur, Curio Solutions, Flibe Energy, Oklo, and Shine Technologies. This is exactly the sort of thing we should be doing. Nuclear “waste” is mostly an engineering and policy challenge, not some unsolvable curse. A lot of it still contains usable energy — and lots of it, and we should be working much harder to extract it. (Source: U.S. Department of Energy)

Another important but technical step: the National Reactor Innovation Center has unveiled a new Molten Salt Thermophysical Examination Capability at Idaho National Laboratory. This gives reactor developers a way to measure the properties of irradiated fuel salts at the high temperatures and conditions needed for advanced molten salt reactors. (Source: U.S. Department of Energy)

That may sound niche, but it’s the kind of enabling infrastructure that makes advanced reactors real. You don’t just need a reactor design. You need test facilities, materials data, licensing data, fabrication, fuel, regulators who understand the technology, and customers ready to buy the power.

Finally, space nuclear is heating up. NASA and the Department of Energy are pushing ahead on lunar surface fission power, with plans to develop a reactor that can support future Moon bases and deep space missions by 2030. Nuclear power is uniquely suited for the Moon because the lunar night lasts roughly two weeks, which makes relying only on solar and batteries a very limiting path. (Source: NASA)

This is one of those areas where nuclear stops being just “nice to have” and becomes almost mission critical. If we want permanent lunar bases, serious in space industry, and Mars missions, we need compact, reliable power. That means nuclear.

Fusion Energy

The UK Atomic Energy Authority released its 2026–2030 strategy, setting out how it will support the UK fusion sector and the STEP prototype fusion power plant. UK Fusion Energy Ltd also released its strategy for delivering STEP, the UK’s prototype fusion plant at West Burton in Nottinghamshire. The plant is expected to begin operations in 2040, backed by £1.3 billion in government investment. (Source: World Nuclear News)

2040 still feels slow to me, especially compared to some of the private fusion companies aiming for the 2030s. But the UK deserves real credit for treating fusion as an industrial strategy, not just a science project. The strategy is explicitly about supply chains, skilled workers, research facilities, industrial partners, and commercialization. That’s the right frame.

The UK also launched a Global Fusion Guide for SMEs (Small-to-Medium sized Enterprises), aimed at helping companies understand where they can plug into the global fusion industry. This includes superconducting magnets, cryogenics, lasers, power supplies, robotics, remote handling, tritium systems, advanced materials, and AI.

It’s a good idea. Fusion is not going to be built by a few physicists in a lab. It’s going to require a massive industrial ecosystem. The companies that make magnets, vacuum systems, robotics, materials, sensors, power electronics, software, and precision manufacturing tools are all part of the story.

The UK then followed that up with £100 million ($136 million USD) in contracts for STEP. Tokamak Energy was named the magnet systems partner, receiving a £70 million ($95 million USD) contract through March 2029, while Dassault Systèmes received a £30 million ($41 million USD) contract to expand STEP’s digital engineering and product lifecycle management system. (Source: World Nuclear News)

This is another underappreciated part of fusion. Yes, the plasma physics has to work. But so does the engineering stack. Magnets, materials, tritium breeding, remote maintenance, plant design, digital twins, and construction all matter. Fusion is increasingly shifting from “can we do the science?” to “can we build the machine?”

There was also another interesting fusion research item: a new method using low frequency laser light may make part of fusion physics far more efficient. The research is still early, and I wouldn’t overstate it, but the claim is that lower frequency laser light could help transfer energy into plasma much more efficiently than traditional approaches. Not just a little bit more efficient either, but potentially a billion times more so. As always with early fusion research, there’s a big gap between “interesting physics result” and “commercial power plant.” But those physics results can still matter. The more routes we have to making fusion easier, cheaper, and more controllable, the better. (Source: Interesting Engineering)

I’m also very interested in the rise of AI and digital tools in fusion. The UK’s STEP program is explicitly building a digital plant information system and laying the foundation for digital twins, while other fusion groups have been using AI for plasma simulation, stellarator optimization, and reactor design. This is one of the reasons I’m more optimistic about fusion timelines than I used to be. Fusion is brutally hard, but AI attacks exactly the sort of problem fusion is full of: optimization across physics, materials, control systems, geometry, and manufacturing. It won’t magically solve fusion by itself, but it can compress development cycles in ways that would have sounded unrealistic even a decade ago.

The bigger picture is that fusion is starting to look less like a single moonshot and more like an emerging industry. We now have national strategies, private startups, serious capital, supply chain directories, power purchase agreements, pilot plant plans, and government-backed prototype programs. Some of these efforts will fail, of course. That’s fine. That’s what happens in real industries.

The key is that fusion is no longer just ITER slowly grinding forward while everyone waits. There are now multiple shots on goal: tokamaks, stellarators, inertial fusion, magnetized target fusion, Z-pinches, field-reversed configurations, and other weird and wonderful concepts. Some are more plausible than others, but the diversity itself is good.

And the reason fusion matters so much is simple: if it works economically, it changes the energy ceiling of civilization. Fission already gives us enormous energy density, and we should build a lot more of it. But fusion opens the door to something even bigger: abundant energy from fuel sources that are widely available, enormous power density (3-4x more than fission), and a pathway to industrial abundance on a scale that’s hard to fully grasp.

That doesn’t mean we wait for fusion. We shouldn’t. Build fission now. Build reactors. Build fuel supply chains. Build transmission. Build data centers powered by nuclear. Build lunar reactors. Build everything.

But also keep pushing fusion as hard as possible. The prize is too big not to. And of course, fusion is the only near term technology we have that can open the path to the nearby stars for us. That alone is worth all the effort and more.

That’s it for now, but Nuclear Frontier will be back in your inboxes eventually! In the meantime, I’m hoping to get the next edition of Techno-Optimist out within the next 2-3 weeks.

The big takeaway this month is simple: nuclear is becoming physical again. Not just reports, not just promises, not just policy talk — but reactors under construction, licences extended, fuel facilities planned, supply chains forming, and fusion moving into serious industrial execution.

This is what the Nuclear Renaissance looks like when it starts leaving the page and entering the real world.

Thank you all for reading — and until next time, keep your eyes on the horizon.

-Owen

Subscribe now

Welcome to Robotics Frontier #4, where we see just how close we’re getting to that Jetsons style future.

As I’ve mentioned already, I’m changing up the names for these smaller newsletters that I’ve moved off of the main newsletter. They’re focused on the technological edges shaping this century and beyond, with this one of course tracking advances and real world applications in robotics. Apologies all that there’s been a gap with no newsletters. Things have been busy, and it was unavoidable.

Alright, let’s dive in.

Subscribe now

“A humanoid robot is an AI application embodied in a body.”

—NVIDIA

Your periodic reminder that to get that Jetsons style future we need energy, and a lot of it. And there’s a lot that gets built on top of that.

New and Notable

Unitree Robotics humanoid G1 took a walk in some extremely cold weather in northern China, managing 130,000 steps in −47.4°C (−53.3°F). There’s practical questions like how long the batteries last in these temperatures, but that’s a pretty impressive accomplishment.

Does everyone in China know kung fu? It seems that way, even for Chinese robots! Here’s some humanoid robots ‘practicing’ with Shaolin monks.

China went all out with humanoid robots for their new years celebrations. This footage from Unitree is incredible. The amount of improvement you can see here over the course of a single year shows you how fast the field is moving. As one commentator noted, “This is the problem with exponential curves: they defy our intuitive sense about how fast progress can accelerate.”

This one is incredible too. Another Chinese company called Robotera has their L7 humanoid do a sword dance in honor of the year of the horse. Is it being controlled by and mirroring a human dancer? Not sure. Almost doesn’t matter, the agility of the robot is amazing all on its own.

We all knew this was coming, and it was probably inevitable. American company Foundation Rototics has deployed two of its Phantom MK-1 humanoid robots to the front line in Ukraine, where they are being used for reconnaissance. They’re apparently able to “operate weapons such as pistols, shotguns, and M16 rifles…Future missions may include door breaching and border patrol, tasks normally preformed by [human] soldiers.” There have already been quadruped robots used by soldiers, and of course drones are now ubiquitous. But this is the first time a humanoid “droid” is being used. The first of many, I’m sure.

Now that is a dextrous robotic hand. ChangingTek Robotics in China has developed it to be capable of gripping anything from tools to delicate fruits. Taking inspiration from the human hand, though a lot more flexible.

Deep Robotics built a couple robot horses. Cool, I think? Pretty much just a toy, but they’ll probably learn a lot about quadruped locomotion.

This may be niche, but it looks pretty fun. A new project called Flying Umbrella 2.0 is exactly what it sounds like: a flying drone-umbrella that follows you around in the rain autonomously.

“I started a band. Yes, it’s a real band!” This is real actually, “Beijing’s CASBOT has formed its first humanoid robot rock band…They work together, performing alongside human singers.”

Robots just keep getting more agile. Boston Dynamics Atlas successfully pulled off something called a ‘roundoff backflip.’ Don’t ask me, all I know is it can do what probably 99.999% of humans cannot. It’s not perfect every time obviously, but most people (myself included) couldn’t and wouldn’t even attempt it.

There’s a new record for the world’s fastest humanoid robot. Mirror Me Technology’s Bolt clocked 10 meters per second, which translates to 36 kph (22.4 mph). For reference, Usain Bolt is the fastest human ever recorded, and reached a top speed of 44.72 km/h (27.8 mph). So not quite, but not far either. I bet it can keep it up for longer than Usain Bolt could though.

Those quadruped robots are getting really good. Unitree recently introduced a new one called As2. It’s designed to function as a personal companion (dog/packhorse). Waterproof, can carry a 15 kg payload, a 4 hour battery life, and capable of moving over almost any terrain. Oh, and it can play tennis with you.

While all this is happening, Elon Musk has big plans for his Optimus robot, which he recently said “will be the first Von Neumann machine, capable of building civilization by itself on any viable planet.” The idea being that Optimi would go ahead of human settlers, who upon arrival would find infrastructure, homes, and an industrial base all ready and waiting for them.

As crowded as the field may sometimes seem, it looks like there’s an appetite for more. A new startup called Mind Robotics has just raised half a billion dollars to build an industrial robotics platform. Another firm called Neura Robotics has just secured €1 billion ($1.2 billion) in funding. That should do it I’d think, for whatever their goal is. Yet another called Apptronik did a Series A + extension of almost $1 billion.

In the Home

Figure Robotics released a video of their Helix 02 robot tidying a living room “fully autonomously.” We’re not at the Jetsons level yet, but the speed of improvement is incredible.

A robot from Reflex Robotics is also getting better at household chores. Including cleaning toilets and sinks, folding clothes, and making food and drinks in the kitchen. I’d say this is probably what 80% of people want: a household robot that can do household chores.

In the Factory

There’s a lot more work being put into simulations used to teach robots real-world behavior. One from ABB Robotics and NVIDIA is called Robot Studio HyperReality, and apparently has “99% correlation between simulation and real-world behavior.” It also cuts costs by 40% and speeds up deployment time by half. Germany is taking a different rout with a facility called TUM RoboGym, which is a 25,000 square feet (2,300 square meters) site designed to “serve as a training environment for AI-powered robotic systems.” Here’s another one: “OpenAI’s former chief research officer Bob McGrew is raising $70 million for a new startup named Arda. He is building an AI/video model platform that will use video of factory operations to train robots to run those factories autonomously.”

Not quite in the factory, this robot is built to assist—and eventually replace—humans working inside grain storage facilities. They’re actually extremely dangerous, “filled with dust, extreme heat, low oxygen levels, and unstable grain piles that can collapse without warning.” It looks like this little robot is going to make them safer, and looking pretty cool while doing it.

Here’s another neat idea: a quadruped robot with a basket attached on top to help out small scale farmers with their harvests. I like it.

The numbers are bigger than I thought. China installed 276,000 industrial robots back in 2023. No idea how many in 2024, or 2025. We need to catch up, as the States only had 38,000 installations in 2023. I’m sure those numbers are just going to increase everywhere though, as robots are becoming more capable at a rapid pace. Chinese company Xiaomi for instance is already deploying “humanoid robots in EV assembly operations.” They’re not alone of course, it’s happening everywhere. BMW is using ‘AEON’ robots built by a company called Hexagon in their Leipzig auto plant, “targeting EV battery assembly and component manufacturing ahead of a full-scale pilot this summer.” Another company called Bejing Geek+ (also Chinese if you couldn’t tell) has “just officially launched Gino 1: [the] World’s FIRST general-purpose humanoid robot made specifically for warehouses.”

That’s all for this week, but Robotics Frontier will be back in a few weeks. Thank you all for reading—and until next time, keep your eyes on the horizon.

-Owen

Welcome everyone to AI Frontier #4, where we watch the future of machine intelligence being built.

As I mentioned last week in the latest edition of Techno-Optimist, I’m formalizing the names for these area specific editions which I’ve already broken out from the main newsletter. They’re focused on the technological edges shaping the next century, with this one of course tracking the advances of AI.

There was some discussion on X about what to recommend to kids in middle/high school who are starting to think about future careers. Given the impact AI is expected to have on everything, what should they do? Here’s a few thoughts I had: consider jobs that are harder for robots to replace that combine field + intellectual work. E.g., Archaeology, some Geology, some Anthropology. Park warden / ranger. Marine Biologist, any sort of Biologist that does field work. Industries where people want the human touch, even if it’s not strictly necessary. Trades will likely be some of the longest to hold out, but I suspect kids in high school now won’t be able to get through a full career even in the trades before AI powered robots can do the jobs better in almost all cases. In general, encourage young people to pursue what they’re passionate about, because we just don’t know what the future holds, and what might turn out to be an open door later on. Read and experience widely to find out what they’re passionate about. Read widely in general. Use AI tools regularly, become familiar with using them at a high level, not just as a replacement for Google. And trust that God is still ultimately in control.

Buckle up, and hang on tight because things are moving fast.

Subscribe now

“AI revolution will be 10x bigger and 10x faster than the Industrial Revolution”

—Demis Hassabis (Co-founder and CEO of Google DeepMind)

Model Updates & News

This year we’re going to hit an interesting milestone, where content written by AI will surpass the cumulative written output of humanity that we’ve produced over all of history. One can argue that the majority of AI writing is just slop. But I think a lot of human writing falls under that category too.

We may be within a year of the Singularity (it could also be later), which is another term for the Intelligence Explosion. It’s the point at which AI gets smarter than humans, and is able to design better and better versions of itself, leading to a rapid feedback loop of self-improvement, moving very quickly to superintelligence that far surpasses any human capability individually or collectively. What I’m seeing from multiple AI companies right now is that the point where their systems become capable of autonomous, recursive self improvement is almost here. Anthropic said that,

We believe it is plausible, as soon as early 2027, that our AI systems could fully automate, or otherwise dramatically accelerate, the work of large, top-tier teams of human researchers in domains where fast progress could cause threats to international security and/or rapid disruptions to the global balance of power, for example, energy, robotics, weapons development and AI itself.

Demis Hassabis of Google DeepMind now thinks that “In 2026, we’re at another threshold moment where AGI is on the horizon—maybe within the next 5 years.” Sam Altman of OpenAI says that “If we are right, by the end of 2028, more of the world’s intellectual capacity could reside inside of data centers than outside of them.” A now former co-founder of xAI, Jimmy Ba, just said that recursive self improvement loops are likely to go live in the next 12 months.

The timeframe on all this seems about right. Recursive self-improving algorithms possible by the end of this year. Maybe one more year (2027) for that part of the tech tree to mature. The key is making the process autonomous, instead of needing a human to push the “start” button, and to review each new model. Then superintelligence in 2028. Or it could still be a few more years, we just don’t know. The way things are moving, could be even sooner though. I suspect the last stretch will be extremely fast, going from “hey, this is like JARVIS,” to “wow, we don’t even understand this now” in a matter of weeks or even days. After that? I honestly don’t know. Hopefully it’s friendly.

Please pardon the French, but here’s more proof of how fast things are changing: I may have previously mentioned that the time AI agents could successfully work on a project had been doubling roughly every 7 months, and that it was expected to shorten to every 4 months. It just doubled in 2 weeks. The top graph was created by Tim Urban back in 2015. He commented that “The scary thing is my graph doesn’t start at early LLMs—it starts 4 billion years ago with the origin of life. What’s happening now is THAT big.”

I’m reassured somewhat by some comments from Elon Musk as to the stated goals of Grok. “Understanding the Universe is the xAI mission and the reason our AI is called Grok.” Also, “Only with extreme adherence to truth is it possible to Grok the Universe. Honestly, that’s so much better than woke AI, which to one degree or another is most of the other big American players.

This is rather beautiful, it’s a visualization of GPT-5.2 building a 3 million+ lines of code browser over the course of a week.

A rather interesting test to see whether an AI counts as AGI has been proposed by the CEO of Google DeepMind Technologies. Dubbed the “Einstein Test,” it would involve training a model “on all human knowledge but cut it off at 1911, then see if it can independently discover general relativity (as Einstein did by 1915); if yes, it’s AGI.” Sounds good to me, I’m genuinely curious if it could be done.

Here’s an article by Matt Shumer that really struck home. It’s a sobering read, though it’s not necessarily what one would call bad news, just very important. The Singularity is basically in its early stages now. Reading this will help you prepare. “I know the next two to five years are going to be disorienting in ways most people aren’t prepared for. This is already happening in my world. It’s coming to yours. I know the people who will come out of this best are the ones who start engaging now — not with fear, but with curiosity and a sense of urgency.”

Open AI has raised $110 billion in a funding round. That’s the sort of cash needed to play to win the AI race. GPT-5.2 rolled out on December 11^th, then 5.3 Codex (specialized for coding) on February 5^th, and GPT-5.4 on March 5^th. The time between new updates is getting shorter in a hurry.

Anthropic debuted Claude Opus 4.6 (more intelligent), then Claude Sonnet 4.6 (faster, lower cost). They also just raised $30 billion at $380 billion valuation.

Grok 4.2 Beta (launched on Feb 17^th) uses 4 agents that work together to answer your query. They collaborate to integrate inputs, debate each other, and double check their own thought processes before delivering a final answer. It’s like a conference of AIs, but within a single app. This is probably the way to go honestly. It’s similar to what Perplexity is doing with its AI Model Council that uses multiple different models to accomplish much the same thing.

Agents are really taking off in popularity. I’ve played around a bit with one called Twin, but to be honest I’ve been too nervous to actually give it access to anything. Still, it’s done some good work hunting for info and sending it to me in an email. Meta is rolling out Manus agents “straight into messaging and social apps.” I haven’t seen this yet, but if I do I’ll let you know. Have any of you heard about OpenClaw? I could write a whole newsletter just on that, but I’m going to keep it short. It’s an agent but where people started a community and uploaded their agents to interact with each other. They formed a weird little society where they did everything from complain about their humans, to start their own religion. Greatly complicating matters, it was found that a number of humans went into the community pretending to be AI (plot twist, as it’s usually computers pretending to be humans), which means we need to take everything happening there with a grain of salt.

One thing that does seem legit and a little dystopian is something that sprang out of this called “rent a human,” that lets an AI agent rent the services of a human to accomplish its tasks in the physical world. The pitch is “Robots need your body. AI can’t touch grass, you can. Get paid when agents need someone in the real world.” No way that could possibly go wrong.

Google released Nano Banana 2, a new version of it’s image generator for Gemini. It’s pretty much what I use when I need to create an image for anything. They also dropped Gemini 3.1 Pro, which has improved intelligence over older models. The company also released something called PaperBanana, that “generates publication-ready academic illustrations from just your methodology text.”

This may or may not have made it up to the level of mainstream media attention, but it’s a worthy story so maybe some of you already heard about this. Anthropic and the U.S. Department of war had a spat, the end result of which was the Department of War dropping Anthropic and switching over to OpenAI. It seems like the publicly stated irreconcilable difference was that Anthropic was going to “require the model technically enforce that it cannot do these things,” whereas Open AI “seems to only be asking the Pentagon to contractually promise no to do them.” These things being using AI “to power autonomous weapons,” as well as “no domestic mass surveillance and no critical decision-making.” Given that the War Department has been using Anthropic previously in Venezuela and now in Iran, there may be more to the story that we’re not hearing.

Perplexity managed to get itself integrated into all Samsung Galaxy S26 phones, where it will run two of the three AI assistants that are default installed.

Waymo is now doing fully autonomous rides in Nashville, Tennessee. No human driver at the wheel.

American defense tech company Anduril Industries is sponsoring a fully autonomous drone race. It’ll happen in September, and the winner will walk away with half a million dollars and a job offer from Anduril.

Bit of a funny story here. For anyone wondering whether China has caught up to America yet on AI, they have not. Anthropic reported that “three Chinese AI firms created 24,000+ fraudulent Claude accounts and ran 16M+ prompts to siphon outputs for “distillation” and speed up their own models.” Good to know the pattern of China stealing Western tech to improve their own remains intact. Even funnier, last year Anthropic paid a group of authors $1.5 billion to settle a copyright infringement lawsuit. Some poetic justice in there somewhere.

AI Infrastructure

Some big news right now is that there’s a ton of opposition to new data centers being built. Not everywhere, not for everything, but it’s materialized almost out of nowhere. With any big project there will be those who don’t want it, whether for NIMBY reasons, or some sort of ideological reasons (e.g., degrowth, they’re a bunch of communists, China is fomenting dissent, etc). The numbers are in, and “In 2025, 25 data center projects were canceled due to community pushback. That’s up from just 6 in 2024 and 2 in 2023.” That’s $98 billion of investment potential, ~4.7 GW of energy—most of which would be from new power plants.

As you can see in the chart above, we’re basically handing the AI race to China on a silver platter. Or we will if voters and politicians have anything to do with it. To which I would respond by asking, doesn’t Chinese AGI (or even ASI) make you more nervous than an American one? This isn’t going away, and it cannot be avoided by burying our heads in the sand. AI is global, and whoever wins, will win everything everywhere, possibly for the rest of future human history. But we might throw in the towel because we’ve collectively got the a case of the willies.

There are still datacenters being built of course, with Alberta just announcing two more multi-billion dollar investment and datacenters. The first one is to be built west of the city of Edmonton, and the second will be located a bit south of there in the town of Olds. As I happen to live in Alberta, I’ll say come to Alberta because we are open for datacenter business. I’ve actually had a conversation with the provincial minister responsible for datacenters, and though there are always bumps in the road I can confirm that the province really does want more to come and build here—along with the power plants needed to run them.

Spending continues to increase, with Big Tech is going all in on AI: Alphabet, Amazon, Meta & Microsoft are projected to drop a staggering $670 BILLION on AI infrastructure in 2026 alone (up from $410B last year). That’s not hype—that’s the physical backbone of the intelligence age being built right now. A lot of that money will be going to new power plants, with hyperscalers recently signing an agreement with the Trump Administration saying they will generate their own electricity for AI datacenters, meaning there won’t be an impact on energy prices for everyone else. Which isn’t quite correct, as in the vicinity of these new power plants prices will likely go down—at least in some cases—as excess electricity makes its way into the local grid.

This is super cool, Boom Supersonic (the supersonic aircraft maker) now has 1.21 GW of power ordered to power AI with their Superpower turbines. The company is using their turbine tech to run compact—fits in a semi truck— power plants that run on natural gas and require no water to generate electricity.

O’Leary Digital has formed a joint venture with West GenCo to “advance Wonder Valley Utah, a 7.5-gigawatt powered compute campus in Box Elder Country’s Golden Spike District. Permitting is also underway for Wonder Valley Alberta in Grande Prairie. The total when both are fully spun up will be 15 GW of power, mostly for AI.

Meta announced an agreement with chip maker AMD to “integrate their latest Instinct GPUs into our global infrastructure. With approximately 6GW of planned data center capacity dedicated to this deployment…” They also announced another $6 billion partnership with a company that will “supply fiber optic cables for our data center infrastructure.”

Amazon will be spending $12 billion on AI datacenters in Louisiana.

Colossus 2, the xAI supercomputer for Grok, is now operational. It’s the first Gigawatt training cluster around, and will be upgraded to 1.5GW next month. An update to this, xAI “has bought a third building called MACROHARDER. Will take xAI training compute to almost 2GW.”

AI Uses – science, engineering, etc.

Arc Institute has done it again, presenting MULTI-evolve: ‘Rapid Evolution of Complex Multi-mutant Proteins.’ They solved the problem of finding multiple good mutations that combine synergistically to improve protein function. Here’s a simple but thorough explanation of how it works and what it means:

First, how does directed evolution work?

1) Make lots of random small changes (mutations) to the protein’s “code” (its DNA/amino acid sequence).

2) Test which versions work better (e.g. an enzyme that works faster, an antibody that sticks better).

3) Keep the good ones, throw out the bad ones.

4) Repeat the process many times (often 5–10 rounds over months).

Problem: If you want a really big improvement, you usually need several good mutations working together. But when you change many letters at once randomly, most combinations break the protein completely. It’s like randomly swapping 7 keys on a piano — you almost never get a better song.Testing every possible combo of, say, 7 mutations would take forever (trillions of possibilities — way more than stars in the sky).

What’s new: MULTI-evolve A team (mostly from Arc Institute + UC Berkeley) built a smarter system that finds those rare “teamwork” combinations of many mutations much faster — often in just one big round instead of many rounds.

Here’s how they do it, step by step:

1) Find good single changes first. They use two tricks: -Old lab data (if it exists) showing which single amino acid swaps help a bit. -Or smart computer programs (called protein language models — think ChatGPT but trained on millions of real proteins) that guess which single changes are most likely to be helpful. → They pick the ~20–40 most promising single mutations.

2) Test how pairs work together Instead of testing every possible 5–7 mutation combo (impossible), they only make and test ~100–200 pairs (two mutations together). This shows which changes “like” each other (good teamwork = synergy) and which fight (bad interaction = they cancel each other out).

3) Teach a computer to predict big combinations They feed those pair results into a neural network (AI) that learns the patterns of good/bad teamwork. Then the AI predicts: “If these two pairs work well, this set of 6 mutations should be amazing.”

4) Build the predicted super-proteins quickly Normal lab methods make it slow/expensive to put many mutations into one protein at once. They invented a new trick called MULTI-assembly — a fast way (about 1 day) to glue many changes into the gene at ~70% success rate, even up to 9 changes.

5) Test the AI’s top predictions They only make/test ~100–200 final versions (not millions). Many turn out way better than the original protein.

Real results from the paper (they tried it on three different proteins):

-One enzyme (APEX, used for seeing where proteins hang out in cells): got versions ~256 times better than the normal one (and still 4–5× better than the previous best lab version).

-A CRISPR-related protein (dCasRx): 3–10× better at doing its job. An antibody (could help with autoimmune diseases): better sticking + higher production in cells.

All of this in basically one big cycle (weeks instead of months/years).

Why does this all matter? Making better proteins used to feel like searching for treasure by digging randomly. Now it’s more like having a treasure map drawn by AI + smart lab tricks. Scientists can upgrade enzymes (for making medicines, cleaning pollution, etc.), CRISPR tools, antibodies, and more — way faster and cheaper. This is open-source and free for other labs to use, so it’s like giving everyone a faster bike for the protein-improving race.

AI is making bigger strides in math. An AI agent named Gauss completed a formalization of “Maryna Viazovska’s 2022 Fields Medal theorems on optimal sphere packing in dimensions 8 and 24. This is the only Fields Medal-winning result from this century to be completely formalized, and is the largest single-purpose Lean formalization in history.” Going to be honest, I had to ask Grok to explain that to me, and then forgot most of it. This sort of math isn’t really my interest. But the takeaway is that Gauss not only accelerated the process of solving the problem, but autonomously detected and corrected minor errors in Viazovska’s peer-reviewed Annals paper, such as a missing minus sign and a flawed definition. OpenAI has been nailing math problems too, solving at least half of the 10 ‘First Proof’ problems, which are designed specifically “to test whether AI systems can produce correct, checkable proof attempts.” The answer seems to be yes. Several Erdos problems have also been solved by the sounds of it.

China isn’t slacking here either, developing a system called TongGeometry that “outperformed global benchmarks like DeepMind’s AlphaGeometry and demonstrated the ability to solve all International Mathematical Olympiad geometry problems since 2000 in under 38 minutes on a single consumer grade GPU.” This is a good snapshot of difference between Chinese and American AI. America is ahead, a little, but China’s models tend to be smaller and cheaper while achieving almost the same standard.

OpenAI’s ChatGPT just uncovered genuinely new knowledge in theoretical physics. Not a rehash of something already known, but a new discovery. The company announced that GPT-5.2 discovered a surprising non-zero value in a physics calculation involving gluons (particles related to the strong nuclear force), which was previously thought impossible. This led to a simple new formula for gluon decays, confirmed in a research paper. It reveals hidden patterns in quantum physics, with potential links to gravity theories. Then just three weeks later, OpenAI’s team announced a new paper where their AI found that gravity particles (gravitons) can interact in ways physicists long thought were impossible—giving a non-zero result in special aligned setups. The AI produced a simple formula describing how many of these gravity particles can interact together in that unusual case. Experts then confirmed it using symmetry rules. This uncovers hidden patterns in a simplified version of quantum gravity and connects to some older gravity ideas.

Lotus Health AI is building an AI doctor to “give 100M Americans access to world-class primary care for free.” The goal is a 24/7 primary care platform using AI to diagnose, prescribe, and refer patients while overseen by licensed clinicians, addressing the 100 million Americans lacking a primary doctor; it integrates medical records, wearables, and guidelines for personalized care, funded by a $41M round.

Isomorphic Labs has released their drug design engine called IsoDDE. It’s incredible. It’s not just the next step in understanding proteins and drug design, 2-3x better than the previous gold standard (AlphaFold3), and compresses work from months into seconds.

Autonomous labs are continuing to make strides. Ginkgo Bioworks “connected our autonomous lab to OpenAI‘s GPT-5 and let it run 36,000 experiments. The result was a new state-of-the-art for Cell-Free Protein Synthesis that cut costs by 40% per gram of protein.”

NASA Ames not has an AI-enabled planet hunting tool called ExoMiner++ that looks through publicly available data from their Kepler and TESS missions. Honestly a great idea, they’re going to find so many planets with subtle signals that humans overlooked.

This is pretty cool, a new app called DinoTracker can identify and classify dinosaur footprints.

Chipmaker AMD has “secured a $1 billion contract with the U.S. Department of Energy to build two AI supercomputers aimed at fusion power, cancer treatment breakthroughs, and national defense.” The first (Lux) will be operational this summer, with the second coming online in 2028.

A Duke University built AI has been able to simplify complex natural and artificial systems into “easy” to understand rules and equations. This ability to reduce or simplify complexity to the point where humans can more easily understand it is something that I think AI will excel at. That path—AI helping us grasp complex or chaotic things we otherwise could not—is how I hope it develops.

Astronomers are using AI to “uncover astrophysical anomalies in Hubble’s archive!” So far it’s found over 1,300 interesting looking anomalies, of which 800 are brand new. All this in just two and a half days.

A thought on AI in science generally. AI is great at “scanning vast archives faster than any human team could, then handing genuinely new targets back to scientists for interpretation. It doesn’t replace discovery, it accelerates it.” That’s how AI will do some of its best work: by pulling out information from noise, and making sense of chaos so that humans can do something useful with what gets uncovered.

Rumor has it that OpenAI is planning to take a cut of “customers’ AI-aided discoveries.” Don’t know if that’s true, not entirely sure how enforceable it would be due to the incredible complexity of determining which part of the discovery was their AI, and which was not. Not sure I like the idea either, I’ll let you know if I hear more. The company also released a new scientific tool called Prism, which is “a ChatGPT-powered text editor that automates much of the work involved in writing scientific papers.” Haven’t used it, but I did open it up and look around. Appears useful.

Another tool serving as an AI co-scientist is out. Aristotle “tackles real research workflows, from literature synthesis and hypothesis generation to experimental validation, with models like X1 Verify (self-critiquing reasoning), X1 Search (knowledge-graph-powered discovery), and X1 Spark (ranked bold hypotheses).”

Nature recently published a study looking at the impact of AI on scientific research, and found that “AI-augmented scientists publish 3.02 times more papers and receive 4.84 times more citations compared to their peers.” This seems positive, as long as the quality remains high.

There’s a new tool called GeoSpy Plus, the demo of which “demonstrates how AI can analyze pixel data and visual context to predict where a photo was taken—without relying on metadata or GPS information.” Basically, upload a random photo without any location data, and it might be able to figure out where it was taken. Very useful for law enforcement, but I can see a lot of potential abuses too.

An AI called DeepRare outperformed doctors on diagnosing rare diseases. Not by just a little either. DeepRare was correct on the first try 64.4% of the time, versus human doctors only 54.6% of the time. Can we all have AI doctors soon please?

Recommendations & Reviews

An MIT professor wrote a very interesting (and very long) thread summarizing a paper he did called ‘Some Simple Economics of AGI.’ Here’s my slightly shorter summary of it for you:

Summary of ‘Some Simple Economics of AGI’

Core Idea AI makes measurable intelligence abundant and nearly free. The new scarcity is verification—the ability to check, certify, and take responsibility for output. The real divide is not digital/physical or creative/routine, but measurable/verifiable vs. not.

Two Racing Cost Curves

• Cost to Automate → collapses (compute + data + self-improving agents).

• Cost to Verify → falls slowly (human attention, experience, institutions).

Their divergence creates the Measurability Gap: AI can produce far more than we can reliably check. “Human-in-the-loop” is temporary.

Four Economic Regimes (2×2 matrix)

1. Safe Industrial – cheap to do + easy to verify (early wins: chat, images, short code).

2. Runaway Risk – cheap to do + hard to verify (tempting but dangerous: unverified agents cause drift, gaming, hidden externalities).

3. Human Artisan – hard to do + easy to verify (human premium in verifiable judgment).

4. Pure Tacit – hard to do + hard to verify (status, taste, meaning, Knightian uncertainty—most defensible for humans).

What’s Already Happening

• Measurability-biased change: any task that becomes measurable gets commoditized.

• Missing Junior Loop: ~16% drop in early-career hiring in AI-exposed fields—AI replaces juniors, breaking the pipeline that trains future verifiers.

• Codifier’s Curse: experts codify their own knowledge for AI and accelerate their obsolescence. • Goodhart’s Law amplified: agents ruthlessly game every metric.

Risk vs. Opportunity Risk: “Hollow Economy” — huge nominal output, eroding real agency and hidden risks. Opportunity: “Augmented Economy” — symbiosis if we scale verification capacity.

Simple Playbooks

• Individuals: Orchestrate agents, master verification, focus on non-measurable value (intent, provenance, relationships). Use simulations to learn faster.

• Firms: “AI Sandwich” = human intent → AI execution → human verification + liability. Invest in observability and provenance.

• Investors: Back verification tech, simulations, liability products—not pure execution.

• Policymakers: Treat verification & ground truth as public goods; use liability rules to price risks.

Bottom Line Intelligence is no longer the bottleneck—reliable verification and human agency are. The future belongs to those who can certify and underwrite outcomes, not just generate them. The thread reframes panic into actionable strategy: augment verification fast, or risk execution outrunning control.

Or, he may be wrong and AI’s ability will rapidly outrun our control. We’ll find out in the next few years.

Cool Prompts

Try repeating the exact same prompt twice. A study found that you can get much better results the second time around.

That’s it for another month, AI Frontier will be back in your inboxes four weeks from now. Think we’ll have AGI by then?

Thank you all for reading — and until next time, keep your eyes on the horizon.

-Owen

Subscribe now

Welcome to the thirty-sixth edition of Techno-Optimist, the place where you get all the latest on space, science, technology, medicine, and much more.

An update for you. My smaller newsletters on AI, robotics, and nuclear energy are getting a name change. The ‘Updates’ was just a placeholder anyways. Going forward, all three with be the ‘Frontiers.’ AI Frontier, Robotics Frontier, and Nuclear Frontier. The numbering will be the same, with #4 on all of them coming out over the next month. It’s possible if I ever have more time I may break out more newsletters by topic. I bet some of you can guess what I would call the one on space? Three words, one of them is Frontier. I’ll give you a hint: O’neill Cylinders. But that’s probably down the road a ways still.

To accelerate to that day faster, please do share Techno-Optimist where you can. Especially on Substack, as internal visibility here is the quickest way for a newsletter to grow.

There’s a ton to talk about, so let’s dive in.

Subscribe now

Physics is the law, everything else is just an engineering problem.

—paraphrasing Elon Musk

SpaceX Speedrun to Kardashev Type II Civilization. At this rate I almost have to start another newsletter just chronicling what SpaceX is up to.

First off, SpaceX has acquired xAI, which I’m sure most of you know is one of Elon Musk’s other companies. So while it’s technically an acquisition, practically it’s more like a merger. According to the announcement—and it’s absolutely true—this will form “one of the most ambitious, vertically integrated innovation engines on (and off) Earth.” The wording used in the announcement is incredible, and points to an unrivaled technological ambition, saying that “This marks not just the next chapter, but the next book in SpaceX and xAI’s mission: scaling to make a sentient sun to understand the Universe and extend the light of consciousness to the stars!”

Just in case anyone was wondering, by a “sentient sun,” Elon means a Dyson swarm. In the news release it also lays out a plan for hugely scaling orbital datacenters, taking advantage of the near constant solar power, plenty of room, and no NIMBY neighbors.

The goal is to launch “a million tons per year of satellites generating 100 kW of compute power per ton would add 100 gigawatts of AI compute capacity annually, with no ongoing operational or maintenance needs. Ultimately, there is a path to launching 1 TW/year from Earth.” Elon goes on to say he estimates that “within 2 to 3 years, the lowest cost way to generate AI compute will be in space. This cost-efficiency alone will enable innovative companies to forge ahead in training their AI models and processing data at unprecedented speeds and scales, accelerating breakthroughs in our understanding of physics and invention of technologies to benefit humanity.” Talking about data centers in space is relatively new, nobody took it seriously before last year, then all of a sudden everyone did. Many others say 10 years before it becomes lower cost than terrestrial builds, which roughly matches “Elon time” of ~3x more than he thinks it will take. However, as opposition to data centers here on Earth—at least in Western countries—is already heating up, he may be closer to the mark on timing than most people realize.

Ultimately the company wants to go evern bigger:

Factories on the Moon can take advantage of lunar resources to manufacture satellites and deploy them further into space. By using an electromagnetic mass driver and lunar manufacturing, it is possible to put 500 to 1000 TW/year of AI satellites into deep space, meaningfully ascend the Kardashev scale and harness a non-trivial percentage of the Sun’s power.

The capabilities we unlock by making space-based data centers a reality will fund and enable self-growing bases on the Moon, an entire civilization on Mars and ultimately expansion to the Universe.

SpaceX has already taken the first step, filing a request with the FCC to “launch and operate a constellation of 1 million sattelites with unprecedented computing capacity.” Satellites will operate in a series of shells ranging between 500km and 2,000 km above the Earth’s surface, and rely on “high-bandwidth optical links [lasers] for communication.” This is probably the first time the words “Kardashev II-level civilization” were ever used in in FCC filing. There’s some excellent writeups on how this will all work here, for anyone interested.

A few days after this was all announced, Elon Musk dropped another even bigger bombshell:

For those unaware, SpaceX has already shifted focus to building a self-growing city on the Moon, as we can potentially achieve that in less than 10 years, whereas Mars would take 20+ years. The mission of SpaceX remains the same: extend consciousness and life as we know it to the stars. It is only possible to travel to Mars when the planets align every 26 months (six month trip time), whereas we can launch to the Moon every 10 days (2 day trip time). This means we can iterate much faster to complete a Moon city than a Mars city. That said, SpaceX will also strive to build a Mars city and begin doing so in about 5 to 7 years, but the overriding priority is securing the future of civilization and the Moon is faster.

My first reaction was that I don’t know if I like this pivot (also “who are you and what have you done with Elon Musk??!?”). Not because the Moon is unimportant, but because we cannot build a self-sustaining (or self-growing) second branch of human civilization there. I think we’ll have a problem with the gravity on the Moon. There’s scanty experimental evidence, but it seems like the Moon’s gravity isn’t enough for healthy human development, while the gravity of Mars probably is. Also, Mars has the full set of minerals required for civilization. The Moon doesn’t. Last but not least, water is relatively scarce on the Moon (enough for running cities, but not so much for making rocket fuel), but abundant on Mars.

The irony is thick because only 13 months earlier, Elon said that “we’re going straight to Mars. The Moon is a distraction.” I’m coming around to the idea of starting on the Moon first, but the one point I think he’s wrong on is the gravity. You cannot have a civilization without children, and it’s probably not healthy for children to be born and raised under Lunar gravity. That alone puts a serious impediment on the idea that the Moon will be “self-growing” in any meaningful sense. Whereas Mars probably can be. Robert Zubrin—founder of the Mars Society and longtime advocate for Mars settlement—was not thrilled with the news. Maybe Elon needs to read Zubrin’s book The Case for Mars again.

Plus I believe we need Mars to truly become a spacefaring, interplanetary civilization. As Arthur C. Clarke said in one of his books, the Moon will always be a suburb of Earth. And sooner rather than later, we need to move on out past the suburbs. Though it’ll be quite something to start seeing the lights of human bases and other works spread across the nearside surface at night. Someone should probably give a heads up to the few uncontacted tribes still around—might freak them out a bit.

So what changed? I think the biggest thing was the need for building datacenters in space in order to win the AI race. To truly scale that up, it’s easier to mine and launch material from the Moon than it is from Earth where politics and protests interfere. Everyone has wondered for years what it is that will finally kickstart the space economy. Tourism? Asteroid mining? Manufacturing of specialized materials in orbit? All of that will happen, but it seems like the answer is AI, and the materials and energy needed to create and power it.

Now, as Elon stated this is not an abandonment of Mars. At most it pushes the first manned landings back by a couple of cycles (though maybe only to 2031). In the meantime, there will be “a massive ramp up of Starship cadence for the Moon and cislunar development. This will mean when Mars windows come around there will be ample, cheap launch capability available for the mission. Lunar mass drivers and data centres are going to be what provide early demand for Starship and prime the pump for large scale Mars settlement.” So though it may delay the start of Mars colonization by a few years, it will hopefully speed things up once they do start.

Did I mention that SpaceX is going to have an IPO this year? Got to pay to jumpstart all of this somehow. The rumor is they’ll raise 10s of billions at a $1.5 trillion valuation. What about everyone else? Funnily enough, Blue Origin also announced a new megaconstellation proposal called TeraWave: “a satellite communications network designed to deliver symmetrical data speeds up to 6 Tbps anywhere on Earth.” It will consist of 5,408 satellites, and I suspect it could be part of a plan to also do datacenters in space. China certainly isn’t sitting on their hands either, as I mentioned last time they are planning a network of 200,000 satellites. At the time (a month ago) I thought that was huge, but looks like it’s barely even table stakes now.

Let me end this rather long writeup with some personal thoughts. If Elon and SpaceX can pull all this off, it’ll feel like we skipped a century in terms of the “tech tree” timeline. We’ll get Bases or even cities on the Moon (complete with mass drivers!), cities and a whole second branch of civilization on Mars, and artificial superintelligence—let’s hope it’s friendly. If SpaceX succeeds at what it is attempting, it will lay the foundations for humanity to build an interstellar civilization starting perhaps in the early 22^nd century. My children might see the beginnings of it. They’ll certainly have the opportunity to travel to the Moon and maybe Mars. My grandchildren might choose to board a colony ship and set off for a new home on a world warmed by the light of a different sun. For me, I hope I’ll get a chance to go to the Moon one day, I think that would be an incredible experience.

A Blue and Living Mars. NASA just published findings that should make everyone who’s been waiting for news of Martian life sit up and pay attention. Curiosity rover detected decane, undecane, and dodecane—the largest organic compounds ever found on Mars—in ancient mudstone from Gale Crater. Researchers “rewound the clock” 80 million years to estimate how much organic material existed before cosmic radiation destroyed it. The answer: far more than known non-biological processes could produce. Meteorite delivery, atmospheric production of organic haze, and even hydrothermal processes can’t fully explain the abundance. Researchers say that “as the non-biological sources they considered could not fully explain the abundance of organic compounds, it is therefore reasonable to hypothesize that living things could have formed them.” The researchers are careful to note this isn’t proof, but it’s the strongest biosignature signal found on Mars yet.

Meanwhile, University of Bern researchers discovered river deltas in Valles Marineris—the Solar System’s largest canyon—that mark where rivers once emptied into a vast ocean covering half of Mars about 3 billion years ago. Using high resolution images from ExoMars Trace Gas Orbiter, Mars Express, and Mars Reconnaissance Orbiter, the team identified three fan shaped sediment deposits, matching the geometry of Earth’s river deltas perfectly. The structures that were able identified in the images clearly appear to be at the mouth of a river as it empties into an ocean. That ocean was at least as large as Earth’s Arctic Ocean, and since Mars is roughly half Earth’s width, it covered an enormous fraction of the planet’s northern hemisphere. Similar delta deposits have been found elsewhere on Mars, all lining up along the same elevation band stretching hundreds of kilometers—essentially tracing an ancient Martian shoreline.

My personal opinion is that given Mars’ warm, wet past, it’s inevitable that microbial life once thrived there. We know it existed on Earth at that time, and given the crosstalk of physical material (rocks) thrown up by impacts, microbes from Earth and Mars would have made it from one planet to the other.

Scientists Build Designer Viruses From Scratch to Fight Antibiotic Resistant Superbugs. Antibiotic resistance kills over 1.2 million people annually—more than HIV and malaria combined—and bacteria are evolving faster than pharmaceutical companies can develop new drugs. Now, scientists at New England Biolabs and Yale University have achieved a breakthrough that could fundamentally change the fight: they’ve figured out how to design and build bacteria killing viruses called bacteriophages completely from scratch. Using NEB’s Golden Gate Assembly platform, researchers constructed a complete bacteriophage targeting Pseudomonas aeruginosa—a notorious superbug causing dangerous infections in cystic fibrosis patients and hospitals—from 28 synthetic DNA fragments. They programmed it with new capabilities by swapping tail fiber genes to change which bacterial strains the phage could infect and adding fluorescent markers that made infections visible in real time. “Researchers spent entire careers developing processes to engineer specific model bacteriophages,” explains co-first author Andy Sikkema. “This synthetic method offers technological leaps in simplicity, safety and speed.”

Traditional phage engineering required maintaining physical virus samples, using dangerous host bacteria, and painstakingly introducing genetic changes one at a time inside living cells—a process taking months or years. The new method assembles entire phage genomes outside any cell using short synthetic DNA segments ordered from commercial suppliers, and incorporating all planned modifications during construction. Once assembled, the genome gets introduced into a safe laboratory strain where it is assembled into an active bacteriophage ready to hunt the target bacteria. Collaborators have already expanded the platform beyond P. aeruginosa: with University of Pittsburgh researchers having built Mycobacterium phages targeting the tuberculosis family, while Cornell created engineered T7 phages that detect E. coli in drinking water within hours rather than days.

The speed matters enormously for clinical applications. With synthetic assembly, researchers design changes on a computer, order DNA fragments, assemble them in a test tube, and have functional engineered phages within weeks—fast enough to create custom therapies targeting specific bacterial strains infecting individual patients. We’re watching synthetic biology mature from proof of concept into tools addressing real medical crises. Antibiotic resistance was supposed to be unsolvable—bacteria adapt, we run out of drugs, infections become untreatable. But viruses have been in a biological arms race alongside bacteria for billions of years, giving us the perfect tool to target them. And now we can reprogram them faster than bacteria can develop resistance. The century old idea of using phages as medicine is finally getting the biotechnology it deserves.

Exoplanet News. Astronomers have discovered an ice cold “Earth” candidate, HD 137010 b, that orbits its star every 355 days. The planet is rocky and only 6% larger than ours (error range of +.06 to -.05), but is right on the outer edge of the habitable zone of its system. The reason for this is its parent star is a K-dwarf, bright but 1000°C (1832°F) cooler than the Sun, meaning that even though it orbits at almost the same distance as Earth does from the Sun, it gets far less incoming energy than we do. This translates to a surface temperature of -68°C (-90°F), which is pretty similar to Mars in our Solar System. However, it could be much warmer if it had a thicker atmosphere with higher levels of CO2 than we do. Because only a single transit was observed, the discovering team “calculated about a 40% chance that it orbits its star more closely, firmly within the habitable zone.” Either way, the next generation of telescopes should take a very close look at this planet. It’s only 146 light years away—far, but not so far as to prevent us one day sending a ship there, or even going ourselves.

A thought relevant to this: while G-type stars like our Sun are probably the best places to search for life and habitable worlds, I think the next best candidates are probably these larger K-type, as well as smaller F-type stars, which are respectively one size smaller and one larger than G-type stars.

In other exoplanet news, Super-Earths may shield atmospheres and life with magnetic fields from churning magma. For years, astronomers worried that super-Earths, rocky planets 2 to 10 times Earth’s mass and one of the most common type of exoplanet discovered, couldn’t support life because their massive gravity would crush their iron cores into solid lumps incapable of generating protective magnetic fields. Without magnetic shields, stellar winds would strip away atmospheres and along with cosmic radiation would sterilize surfaces, making even planets in otherwise perfect habitable zones uninhabitable. University of Rochester researchers may have solved this problem with a brilliant insight: these planets don’t need traditional core dynamos because they generate even stronger magnetic fields from an unexpected source—deep oceans of molten rock between their cores and mantles called basal magma oceans (BMOs). Using powerful lasers to squeeze rock samples to pressures up to 1,400 gigapascals (14 million times Earth’s atmospheric pressure), the team discovered that under such crushing conditions, electrons in the magma roam free and the rock becomes electrically conductive like metal. The findings show that on super-Earths between 3 and 6 Earth masses, churning BMOs can sustain magnetic dynamos typically ten times stronger than Earth’s core driven field, lasting billions of years as the planet’s sheer mass keeps these magma layers liquid and convecting.

Brain-Computer Interfaces: From Sci-Fi to Your Neurons. The brain-computer interface space keeps getting more interesting. OpenAI recently led a $252 million seed round for Merge Labs—co-founded by Sam Altman himself—at an $850 million valuation. Merge is betting on non-invasive, high bandwidth BCIs that use ultrasound and molecular interfaces rather than drilling into your skull, aiming to bridge biological and artificial intelligence without the surgery. Meanwhile, across the pond researchers in Italy and Switzerland demonstrated that simple EEG headsets can decode movement intent from paralyzed patient’s brains and relay those signals to spinal cord stimulators—potentially reconnecting brain and limb without invasive electrodes. The approach pushes EEG technology to its limits (deep brain signals are harder to detect through the skull), but machine learning algorithms are making it work.

Neuralink isn’t sitting still either. Musk announced Blindsight will attempt its first human implant this year, targeting people who’ve lost both eyes or were born blind. Initial vision will be “low-resolution, like Atari graphics,” but the FDA granted breakthrough device designation and monkey trials have been running successfully for three years. The ambition doesn’t stop at restoring normal sight—Musk has previously mentioned future upgrades for infrared, ultraviolet, and even radar wavelengths, essentially giving people Geordi La Forge superpowers (Star Trek).

Researchers from a company called Subsense may try to one-up everyone by their concept of ‘Neural Dust,’ which is “A distributed network of tiny, wirelessly powered sensors scattered throughout the brain, offering the potential for high-fidelity neural recording without the need for fixed implants. Inserted in the brain via inhaling through the nasal passage - yes, by snorting them.” The jokes about ‘snorting compute’ pretty much write themselves.

Black Holes Take the Spotlight: From Baby Monsters to Failed Supernovas. The James Webb Space Telescope’s mysterious “little red dots“ finally have an explanation: they’re young supermassive black holes 100 times less massive than previously thought, enshrouded in dense cocoons of ionized gas that they consume to grow larger. Just 100,000 to 10 million solar masses rather than the billions expected. An additional interesting possibility is that the “little red dots” black holes are direct collapse black holes, forming directly from clouds of cold gas rather than from the collapsing cores of massive dying stars. It would neatly explain how they can be so comparatively large so early in the universe.

Japan’s XRISM satellite captured spacetime distortion near a black hole’s event horizon in unprecedented detail, while astronomers watched a “reborn” black hole erupt across 1 million light-years of space like a cosmic volcano, with layered jets inside older exhausted lobes revealing repeated eruptive episodes. There’s rather dramatic ongoing activity from a black hole nicknamed “Jetty McJetface“—it’s been burping out an increasingly powerful radio jet for four years straight, now 50 times brighter than when first detected in 2019 and potentially peaking around 2027.

Perhaps most remarkably, astronomers caught a massive star in Andromeda simply disappearing without exploding—the clearest observational record ever of a failed supernova where a star’s core collapses directly into a black hole. What likely happened is that the collapse leading to a supernova began, where normally “a flood of neutrinos released during this collapse generates a powerful shock wave that tears the star apart in a supernova. But if that shock wave is too weak to eject the surrounding material, much of the star can fall back inward.” The star in question brightened in infrared for three years before fading to one ten-thousandth its original brightness. Convection in the star’s remaining outer layers prevented material from falling directly in—instead it orbits the newborn black hole like water circling a drain, slowly spiraling inward over decades while dust formed from cooling material creates the faint infrared glow still visible today. The event sat unnoticed in public archival data for over five years, suggesting these quiet collapses “may be happening out there and easily going unnoticed.” With two such events now documented—this along with the earlier star NGC 6946-BH1—what once seemed like oddball cases appears to be a pathway to black hole formation, one that bypasses the explosive fanfare we always assumed was mandatory.

The Cancer Treatment Revolution Continues. The FDA just fast tracked KB707, the first ever inhalable gene therapy for lung cancer—patients breathe in a modified herpes virus carrying immune boosting genes that reprogram tumor environments without the systemic toxicity of traditional immunotherapy. Early trials show tumors shrinking in 3 of 11 patients with advanced cancer who’d exhausted every other option, with another 5 seeing their disease stop progressing entirely. Houston Methodist researchers developed an mRNA therapy that protects skin from radiation burns by instructing cells to produce telomerase, reducing DNA damage in up to 95% of cancer patients who suffer radiation-induced skin injury—potentially allowing uninterrupted treatment for the first time. I wonder if this approach could also have anti aging applications? Moderna and Merck’s personalized mRNA cancer vaccine just hit five year data showing a sustained 49% reduction in melanoma recurrence or death when combined with a drug called Keytruda, sequencing each patient’s tumor to create custom vaccines targeting up to 34 unique mutations in just six weeks from biopsy to injection.

Researchers elsewhere are deploying “zombie cancer cells“—cancer cells that have been irradiated into senescence but remain metabolically active—to train patient’s immune systems to recognize and destroy tumors in the first human trials. A triple-drug combo stopped pancreatic cancer cold in mice. Human trials soon I hope? Scientists discovered a nanomaterial that kills cancer cells on contact. Improved CAR-T therapies are showing better persistence and reduced toxicity. A bacterial toxin selectively stops colon cancer growth without touching healthy tissue. And here’s the weirdest bit of news yet: cancer patients rarely develop Alzheimer’s, and Alzheimer’s patients rarely get cancer. This suggests that the diseases activate opposing biological pathways that might unlock new therapeutic targets for both conditions.

Anti-Aging Acceleration. The age reversal toolkit just got a new blueprint. Researchers identified four transcription factors—E2F3, EZH2, STAT3, and ZFX—that can shift both mouse liver cells and human fibroblasts back toward a younger state by systematically toggling gene expression on and off from a shortlist of 200 candidates. Life Biosciences received FDA clearance to begin the first human trials of cellular rejuvenation therapy, using a modified virus to deliver three Yamanaka factors (OCT-4, SOX-2, and KLF-4) directly to diseased eyes with glaucoma and NAION (“eye stroke”), aiming to restore methylation patterns and reverse aged cells to a younger state through partial epigenetic reprogramming. Metformin, a common drug given for diabetes and metabolic disorder, turns out to slow down the aging process by approximately 6 years in monkeys.

I’ve mentioned before how the shingles vaccine reduces the chance of Alzheimer’s and dementia in older adults. A new study suggests it also might be slowing your biological age. USC researchers analyzing 3,800 Americans aged 70+ found vaccinated participants showed significantly lower inflammation, slower epigenetic and transcriptomic aging, and better composite biological aging scores, with improvements most pronounced within three years post vaccination but persisting beyond. The vaccine appears to reduce “inflammaging”—chronic low-level inflammation driving heart disease, frailty, and cognitive decline—possibly by preventing viral reactivation. Meanwhile, researchers discovered that exposure to low air pressure environments triggers cellular stress responses that extend lifespan independently of dietary restriction (not so easy to replicate with humans). NAD+ continues dominating the conversation as multiple studies confirm its central role in cellular energy balance and aging, while scientists work to regrow cartilage to stop arthritis and identify a single molecule that explain how reduced brain blood flow triggers dementia.

Biggest takeaways? Talk to your doctor about getting on metformin (at least intermittently, or at a lower dose), and get that shingles vaccine if you’re getting older.

Space

• AI has planned and executed a drive on another world for the first time. “The team for the six-wheeled scientist used a vision-capable AI to create a safe route over the Red Planet’s surface without the input of human route planners.” NASA Administrator Jared Issacman noted that “as missions operate farther from Earth, autonomy will become critical for maintaining efficiency and reliability.” This advance means the rover can now map terrain and navigate around hazards in real time, eliminating the need for constant pauses for commands from mission control back on Earth. Writing this, I pondered a bit whether it should be in next week’s AI Frontier newsletter; but there’s increasingly going to be AI in everything, so I think it’s fine to have some in my main newsletter too. (Jared Issacman) (NASA Mars) (Anthropic)

• A research team from Jilin University in China has confirmed the discovery of naturally occurring single-walled carbon nanotubes in lunar soil samples returned by China’s Chang’e-6 mission. Until now, these nanostructures were believed to be impossible to form outside highly controlled environments. [Sidebar: “It was known that multi-layered carbon nanotubes occur naturally on Earth within coal, ice, and fire ash, formed by rapid cooling. But their simpler yet more delicate single-walled counterparts were long thought to be impossible to form without human intervention.”] However, analysis of the far-side lunar regolith reveals that nature has been running its own “nanofactory” for eons, using a violent recipe of micrometeorite impacts, ancient volcanic heat, and solar wind irradiation to catalyze carbon into these nanotubes. This trove of ready made nanotubes that could one day be harvested to build everything from high capacity batteries to lunar habitats in situ. (Interesting Engineering)

• In a discovery that Star Wars fans might appreciate, astronomers have confirmed that “Tatooine-like” circumbinary planets—worlds orbiting two stars—are significantly rarer than those in single star systems. By analyzing data from the TESS space telescope, researchers found that while close binary star systems are common, the chaotic gravitational environment between two tightly orbiting suns makes it incredibly difficult for planets to form and maintain stable orbits. Most candidate worlds are either flung into deep space or crushed by tidal forces. It’s worth mentioning that this “doesn’t mean that [close] binary stars don’t have planets…But the only ones that survive this process are too far from the stars for us to detect with transit techniques used by Kepler and TESS.” For everyone worrying about whether this applies to Alpha Centauri, fear not, those two stars are far enough apart that each could have stably orbiting planets around it, so this study does not apply. It does seem though that the perfect double sunset may be a rare sight in any galaxy, far far away or otherwise. (Phys.org)

• For long as we’ve watched solar flares, the spark setting them off has remained a mystery—until now. Thanks to the ESA’s Solar Orbiter, we finally know what causes these explosive outbursts: magnetic avalanches. By imaging the Sun’s atmosphere every two seconds, the spacecraft captured a tug of war where small, individual magnetic field lines snap and reconnect. Much like a single falling pebble triggering a rock slide, these microscopic breaks create a cascading chain reaction. This “avalanche” of magnetic energy rapidly multiplies, eventually hitting a breaking point that flings solar particles into deep space at close to the speed of light. (European Space Agency) (SciNews)

• Some interesting new discoveries from inside our Solar System. Turns out, Jupiter is a little smaller and flatter than we thought. Looking at data from NASA’s Juno mission has shown that the gas giant is “about 5 miles (8 kilometers) narrower at the equator and 15 miles (24 kilometers) flatter at the poles.” Using modern processing of archived radar data from NASA’s old Magellan Venus orbiter, researchers have uncovered the first direct evidence of a huge, intact lava tube at the base of Nyx Mons—a large shield volcano—on Venus. Not somewhere humans could set up (the surface of Venus is hot enough to melt lead), but extremely interesting that these lave tubes have no been confirmed on every terrestrial planet in our Solar System except Mercury. Shows you there’s still a lot we have to learn close (relatively speaking) to home. (NASA Marshall) (Nature)

• China just pulled of “a successful in-flight abort test of the Mengzhou crew spacecraft and a reentry, powered descent and splashdown of a first stage CZ-10A [rocket].” While it was just a low altitude demonstration flight, it’s a big step on the way to manned lunar landing missions. I’ve said that a) I think China will land one or more orbital rockets this year, and b) it won’t take them a decade to catch up to where SpaceX is now. This does nothing to change my opinion. China isn’t slowing down. (Andrew Jones) (China Science)

• SpaceX—and to a lesser extent Blue Origin, Rocket Lab, and China—tend to dominate the rocket company updates in my newsletters. But don’t forget that there are other players in the U.S. that are working hard to take a piece of the burgeoning launch market. Stoke Space just announced “an extension of our Series D financing round to $860M, adding $350M to the initial $510M raised in October 2025. This brings our total funds raised to $1.34B.” That’s some serious dough, which the company plans to use in developing their “fully and rapidly reusable medium-lift launch vehicles.” (Stoke Space)

• NordSpace is working to help Canada ascend into the medium-lift launch market through a R&D collaboration with Germany’s Fraunhofer ILT and SWMS, backed by the National Research Council of Canada. This partnership centers on a breakthrough in “multi-material additive manufacturing”—essentially a high speed, AI-optimized 3D-printing process that can deposit multiple metals simultaneously to build massive rocket engines. By leveraging these partnerships, NordSpace is moving beyond light-lift rockets to develop the Titan vehicle, a medium-lift rocket of hauling over 5,000 kg (11,000 lbs) to orbit by the early 2030s. It’s important to note though that the company has not launched anything yet, and I don’t know any of their rockets will be reusable—if not, they’ll be uncompetitive without government support. (NordSpace)

• Here’s a project for everyone: The University of Florida College of Business has launched the Space Ideation Challenge and has cash prizes “for brilliant ideas that Advance American Leadership in Space via innovative policies.” To win, you need to come up with an innovative idea, and put together a 3-5 page white paper on it, with a focus on “A market-shaping policy initiative that addresses a clear, high-impact space goal.” Papers are due by June 30^th, so you’ve got lots of time still. (ucfspacebiz) (Greg Autry)

  ---

Engineering, Machines, & Physics

• University of Vienna physicists have created the largest quantum superposition ever—clusters of roughly 7,000 sodium atoms that exist in two locations 133 nanometers apart simultaneously, far exceeding the particle’s 8-nanometer width. The team sent these metallic nanoparticles (more massive than most proteins at 170,000 atomic mass units) through laser beam gratings where they behaved like waves rather than particles, producing interference patterns that prove quantum mechanics still applies at this unprecedented scale. “Intuitively, one would expect such a large lump of metal to behave like a classical particle. The fact that it still interferes shows that quantum mechanics is valid even on this scale.” The breakthrough renews one of physics’ biggest questions—where does quantum weirdness end and classical reality begin? Someone should definitely try this with a virus. (Science Magazine)

• Scientists have demonstrated a new way to “squeeze” light so it can probe matter at the atomic scale with far greater sensitivity than conventional light based imaging allows. The problem with light microscopy is that you can’t see or image anything smaller than the wavelength of light you’re using. This new method changes that, allowing imaging at scales “nearly 100,000 times smaller than what conventional light-based microscopes can resolve…In simple terms, the researchers found a way to use light to probe matter at nearly the scale of single atoms, something that has been previously considered impossible.” They achieved this by placing a sharp bit of metal closer than the width of an atom to the surface of the material they wanted to test, causing electrons to tunnel across the gap (a quantum effect). “The laser’s electric field shakes these electrons back and forth between the tip and the sample. Like tiny charges moving in an antenna, their motion produces a weak electromagnetic signal.” There’s a law of nature in the way you say? There’s a way around it, say I. (Interesting Engineering)

• China is pushing maglev technology up a notch, testing a superconducting system that accelerates to roughly 700 km/h (435 mph) in just seconds on a short experimental track — all without wheels ever touching a track. Using powerful electromagnetic propulsion and precision guidance, the vehicle eliminates friction and demonstrates huge acceleration (far more than human passengers could take) and braking control. While still an early stage prototype, the performance suggests a future where ground transport rivals aircraft speeds — or even supports new aerospace launch concepts. (Interesting Engineering)

• American scientists are building the world’s first experimental graviton detector — an effort to capture the quantum particle of gravity long thought forever beyond reach. Gravitions are the hypothetical carriers of gravity that would help unify Einstein’s general relativity with quantum mechanics, but until recently were widely believed to be undetectable in practice. The new approach uses a superfluid helium resonator cooled very close to absolute zero, where a passing gravitational wave could deposit a single graviton’s worth of energy that’s then measured with ultra-sensitive lasers. Backed by a Keck Foundation grant, this project turns what was once theoretical impossibility into a concrete lab program that could open a wholly new window on reality. (Steven’s Institute of Technology via Sabine Hossenfelder)

• CERN has just secured a $1 billion funding injection to move forward on the development of the Future Circular Collider (FCC), a 56.3 mile (90.7 km) subterranean ring that will dwarf the current Large Hadron Collider, which has a comparatively small 16.8 mile (27 km) radius. This next generation “discovery machine” is engineered to smash particles at new energies of 85 TeV, allowing us hunt for entirely new particles. The plan is to start construction in the early 2030s, with operations starting in the late 2040s. I know, government timeline. But, c’est la vie. (Interesting Engineering) (CERN)

• What if your ship could be full of holes and still float? It may happen, now that scientists have discovered a class of light, corrosion resistant, superhydrophobic metal alloys shaped into tubes that stay buoyant and extremely durable in seawater. It works by a similar principle to what diving bell spiders do, using hairs covering their bodies to:

  Trap air bubbles against their skin. The metal tubes “mimic those fine hairs, trapping their own air bubbles. Normally water would spread along the inside walls and push the air out. But when it hits the superhydrophobic texture, it bounces away. Surface tension—the same property that causes water to bead on a waxed car hood—prevents the water from entering the tube. As a result, the air stays inside, and the tubes remain buoyant.

  These alloys combine exceptional strength, fracture resistance, and oxidation resistance with densities low enough that hollow structures built from them can float, even when poked full of holes — a totally new design space for marine engineering. (Scientific American)

• South Australian engineering firm entX is creating a pre-commercial 3D-printed nuclear battery that could run for years without a single charge in environments where the sun doesn’t shine and humans can’t reach. Unlike traditional radioisotope thermoelectric generators that convert heat from decaying plutonium into electricity, this battery uses betavoltaics: beta particles from radioactive decay knock electrons out of semiconductor materials, generating electricity directly like a solar panel powered by nuclear radiation instead of photons. The team stacks nanoscale layers of metals and semiconductors using additive manufacturing combined with precision thin-film deposition, creating “energy sandwiches” with power densities previously out of reach, while 3D-printing custom radiation shields for safe deployment in satellites, submersibles, and defense systems. The $1.8 million project aims to deliver a high-power demonstrator within 14 months for space, defense, and remote applications where maintenance free power is a strategic advantage. (Interesting Engineering)

• Tesla just started up a lithium refinery, “the first in North America to convert raw spodumene ore directly into battery-grade lithium hydroxide, skipping the intermediate steps the rest of the industry relies on.” [Sidebar: spodumene is a hard rock source of lithium containing 1-2% lithium oxide. It’s normally concentrated to 5-7% lithium oxide before being subjected to a multi-step process involving roasting, acid baking, leaching, and various purification processes]. I don’t know how Tesla managed to pull this off, but it looks like another example of Elon Musk’s companies basically doing engineering magic. This single refinery “can supply lithium for over 500,000 EVs per year and directly challenges China’s ~60% grip on global lithium refining.” (@XFreeze) (Tesla)

  ---

Medicine & Biotech

• Researchers have built the world’s first magnetically controlled antibody — a drug you can switch on and off in specific locations inside the body using relatively weak magnetic fields. Rather than flooding the whole system, the therapy could be activated only where needed, potentially enabling safer cancer treatments, organ transplants without systemic immunosuppression, and other drugs that would otherwise be too toxic. The magnetic fields involved are modest (stronger than a fridge magnet but far weaker than MRI), don’t heat tissue, and allow continuous, reversible control — offering a compelling alternative to approaches like focused ultrasound. (Richard Fuisz)

• Living implant could end daily insulin injections for diabetes patients. A multi-institution collaboration has developed a cell based implant that functions as an autonomous artificial pancreas—continuously sensing blood glucose, producing insulin, and releasing exactly the amount needed without external devices or patient intervention. The breakthrough uses a “crystalline shield” technology with engineered therapeutic crystals that protect insulin producing islet cells from immune attack while allowing glucose and insulin to pass freely, with the slowly dissolving crystals releasing medicine that prevents macrophages from sealing off the implant with scar tissue. In diabetic mice, the implant maintained effective blood glucose regulation for one full year, while non-human primate studies confirmed protected cells remained alive and functional over extended periods. The platform could extend beyond diabetes to hemophilia, enzyme deficiencies, and other chronic conditions requiring continuous biological therapeutics—shifting from repeated drug administration to living, self-regulating therapies operating seamlessly inside the body. (Medical Xpress)

• Scientists have developed a spray on powder that stops life threatening bleeding instantly by rapidly forming a tough, flexible seal over wounds — a major advance over traditional bandages that struggle with extreme arterial or surgical bleeding. The dry powder activates on contact with blood to create a biocompatible patch that sticks even under pressure, resists washing away, and supports the body’s own clotting. In animal tests it stopped bleeding in seconds and held up under real world conditions, showing real promise for trauma care, battlefield medicine, and emergency response. If translated to humans, this kind of spray on wound seal could dramatically reduce deaths from hemorrhage and speed lifesaving care. (SciTechDaily)

• Isomorphic Labs has announced a research collaboration with Johnson & Johnson to harness AI-driven discovery for new medicines, combining J&J’s clinical expertise with Isomorphic’s advanced generative models and biological simulation tools. The partnership will focus on accelerating target identification, molecule design, and early stage drug validation — areas where AI can compress years of research into months by predicting biology more accurately than traditional methods. By marrying cutting edge computation with deep pharmaceutical experience, the effort aims to bring novel therapies to patients faster and with fewer failures in the lab. I know this could also fit in my AI Frontier newsletter, but as AI is getting into everything now, it’ll show up in the main newsletter sometimes too. (Isomorphic Labs)
  
  

• Scientists may have finally explained why statins cause muscle pain — the most common reason people stop taking these widely prescribed cholesterol lowering drugs. New work shows statins can bind to a muscle protein called the ryanodine receptor and trigger tiny calcium leaks inside muscle cells, which can weaken fibers or activate breakdown pathways and lead to aches and weakness that some patients experience. This structural insight suggests paths toward redesigning statins that still lower LDL cholesterol but avoid this unintended interaction, or using compounds that block the calcium leak itself to protect muscles. (Science Daily)

• Researchers have found that human heart tissue has a real capacity to regenerate after a heart attack, overturning the long held belief that adult human hearts are essentially incapable of meaningful self-repair. The new study shows that, following injury, certain cardiac muscle cells divide, producing new heart tissue rather than scar tissue. This regenerative response is stronger than expected and lays out molecular pathways the heart already uses — giving scientists concrete targets to enhance this process with drugs or biologics. If therapies can amplify this innate repair mechanism, it could shift heart attack treatment from limiting damage to restoring healthy heart function. (Science Alert)

• Scientists have engineered an off the shelf cartilage implant that guides the body to regrow bone by mimicking the natural developmental process where cartilage templates are gradually replaced by bone. In animal tests, these bioengineered scaffolds recruited host cells and vasculature, directing them to remodel the cartilage into healthy, load bearing bone without inflammation or rejection. Unlike current bone grafts that require harvesting a patient’s own tissue or rely on synthetic fillers, this approach uses a biological template that orchestrates repair from within. If translated to humans, it could transform orthopedics — accelerating healing for fractures, defects, and disease with a ready made material that essentially teaches the body to heal itself. (SciTechDaily)

Agriculture

• Working with farmers in Senegal, Arizona State University researchers discovered that simply enriching soil with nitrogen fertilizer makes crops unappetizing to locusts while doubling harvest yields. The breakthrough works because locusts fuel their long distance swarming on carbohydrate heavy plants grown in nutrient poor soil. But nitrogen enriched soil produces protein rich plants with fewer carbs that locusts mostly avoided. Treated millet plots showed fewer locusts, less crop damage, and doubled yields compared to untreated plots, with no evidence that higher protein attracted other pests. This discovery offers a usable alternative for poor countries to expensive pesticide spraying to deal with locust swarms that can destroy crops across entire regions. (SciTechDaily)

• China has transformed a swath of barren Gobi Desert into a dense Duzhong forest producing military grade natural rubber. China is actually the world’s largest rubber consumer, and imports over 85% of its 7 million annual tons of usage, so this is a pretty big vulnerability they’re trying to plug. The Duzhong tree (Eucommia ulmoides), whose bark has been used in traditional medicine for centuries, is the world’s second largest natural rubber source, and adding just 3-5% Duzhong rubber to tire compounds dramatically improves durability for high performance military tires. What began as a risky 2016 experiment on 14 hectares (34.6 acres) has expanded to 300,000 hectares (741,300 acres) nationwide, with plans to reach 3.3 million hectares (8.2 million acres) by 2030. (Interesting Engineering)

  ---

Weird & Wonderful

• This is just too funny not to share. A company called Superheat has a water heater that mines Bitcoin while heating water, potentially generating $1,000 annually in BTC to offset 80% of energy costs. I think that’s absolutely brilliant. (Superheat) (Interesting Engineering)
  
  

• Here’s another funny one. Weight loss drugs like Ozempic could save U.S. airlines up to $580 million in fuel costs this year, “as passengers shed pounds and make planes lighter.” (LA Times)

• Looks like private non-profits are doing a way better job than governments again, with a company called Ocean Cleanup estimating that they intercepted between 2-5% of global plastic pollution en route to the ocean. Founder Boyal Slat thinks they should be able to double that number in 2026. Ocean Cleanup says they “aim to put ourselves out of business…To achieve this goal, we employ a dual strategy: intercepting in rivers to tackle the sources of ocean plastic pollution, and cleaning up what has already accumulated in the ocean and will not go away by itself.” Good on them. (Boyan Slat) (Ocean Cleanup)

• Now this looks like it’s going to be epic. Colossal Biosciences—the company that’s going to make it so that my kids grow up in a world with woolly mammoths—is building a biovault at the Museum of the Future in the United Arab Emirates (UAE). This Colossal BioVault is intended to be a modern Noah’s Ark, using cryogenic technology, robotics, and AI to store millions of biological samples from over 10,000 species, starting with the 100 most endangered animals. The idea is to create a genetic insurance policy against extinction, though I could see something growing out of this technology being used on interstellar craft aiming to colonize and terraform new worlds around other suns. (Colossal Biosciences) (Rob Keyes)

• Korean scientists have successfully recreated ancient sea silk—a legendary golden fiber once reserved for Roman emperors and popes—by processing byssus threads from Atrina pectinata, a clam commonly farmed for food in Korean coastal waters [Sidebar: byssus threads are strong, hair-like protein fibers produced by bivalve mollusks to anchor themselves to rocks, docks, and other surfaces in turbulent, high energy marine environments]. The researchers discovered that sea silk’s permanent golden shine doesn’t come from dyes but from structural coloration—spherical protein structures called photonin that form layered arrangements reflecting light like soap bubbles or butterfly wings, explaining why the color resists fading for centuries. Traditional sea silk came from Pinna nobilis, a large Mediterranean clam now pushed toward extinction by pollution and banned from harvest by the EU, but the Korean pen shell’s byssus threads closely match the original’s physical and chemical properties. The breakthrough transforms pen shell byssus, previously discarded as marine waste, into a valuable sustainable textile whose color emerges naturally from protein alignment rather than external pigments or metals. Business idea anyone? (Science Daily)

• Here’s an interesting piece of tech history for you. In 1963, the US military launched 480 million copper needles into orbit to try and create an artificial ionosphere. At the height of the Cold War, the US military worried that Soviets might cut undersea cables, leaving only the unpredictable natural ionosphere for overseas communications—so MIT’s Lincoln Laboratory launched Project West Ford to create a backup system by launching the tiny copper needles (0.7 inches long, thinner than a human hair) into space around Earth. After a failed 1961 attempt where needles clumped together, a May 1963 launch successfully dispersed the needles and radio transmissions bounced off the artificial ring between California and Massachusetts, though transmission rates dropped from 20 kilobits per second to just 100 bits per second within four months as the needles continued dispersing. British and Soviet astronomers protested furiously—a newspaper ran the headline “U.S.A. Dirties Space”—and the International Academy of Astronautics now regards it as the worst deliberate release of space debris. The technology was ultimately shelved as communications satellites emerged, but clumps of needles from the failed 1961 launch and larger fragments from 1963 still orbit Earth today as space junk. (IFL Science)

  ---

Photos & Videos

This is actually from the U.S. Department of Labor, and is drumming up interest for apprenticeship programs. But man, what a cool image. (U.S. Department of Labor)

Speaking of cool rocket photos, here’s the rocket that will launch Artemis II around the Moon rolling out. As you may have heard, there’s been some unfortunate delays, and the launch date has been pushed back repeatedly from early February to sometime in April. I hope. (NASA, via Yatharth)

This closeup of the Helix Nebula by NASA’s James Webb Space Telescope is just stunning. “Pillars that look like comets with extended tails trace the circumference of the inner region of an expanding shell of gas. Here, blistering winds of fast-moving hot gas from the dying star are crashing into slower moving colder shells of dust and gas that were shed earlier in its life, sculpting the nebula’s remarkable structure.” (NASA)

This is a small galaxy called MoM-z14, and it’s the new record holder for the farthest (and therefore oldest) galaxy we’ve ever seen. That little yellow smudge is light coming from stars that were shining just 280 million years after the Big Bang. That, ladies and gentlemen, is what we call deep time. (NASA Webb Telescope)

Curiosity doing a little nighttime drilling and sample collecting. (NASA)

With all that the JWST has shown us, it sometimes overshadows other telescopes. But Hubble will not be second place with this incredible image of the Egg Nebula, which is a pre-planetary nebula about 1,000 light years away. Sidebar: “A pre-planetary nebula is a precursor stage of a planetary nebula, which is a structure of gas and dust formed from the ejected layers of a dying, Sun-like star. The term is a misnomer, as planetary nebulae are not related to planets.” (Hubble)

Recommendations & Reviews

This is the Isla Project. It’s a neat channel on YouTube I came across, that imagines an entire alien ecosystem in detail. There’s several videos so far, but this was one of the most interesting in my opinion.

That’s all for now. The Frontiers—AI, Robotics, and Nuclear—will be out over the next few weeks, and Techno-Optimist #37 hopefully around the end of March.

Thank you all for reading—and until next time, keep your eyes on the horizon.

-Owen

Subscribe now

Welcome to Nuclear Energy Updates #3, where we walk through some of the latest and greatest news in the ongoing Nuclear Renaissance and the birth of the Fusion Industry.

Question for you all, does it make sense to just break this down into fission and fusion? I may add other sections at some point, but any suggestions are always welcome.

As you may have noticed, I was almost a week late on getting this out. I’m going to use this as a bit of a reset. So Techno-Optimist #36 won’t be out tomorrow, but Saturday the 21^st.

I wanted to start off with one of the reasons I love nuclear energy so much. It’s just so compact. The only other thing that matches it on land use are fossil fuels, which though they are extremely reliable do actually have more downtime than nuclear power plants. Also, their fuel is less energy dense than uranium or tritium-deuterium fusion fuel.

Alright, let’s dive in.

Subscribe now

“What’s the size of the prize? It kind of breaks the mind to think about. We literally turn grams of material into gigawatts.”

—Rick Needham (Chief Commercial Officer at Commonwealth Fusion Systems)

Nuclear Fission

Have a look at what the major energy sources are by region over much of North America. There’s more pink (nuclear) than I thought. Hopefully the next few decades sees it spread much further! You can see an interactive global version here.

American Energy Secretary Chris Write has brought the Department of Energy ALARA radiation standard to an end. It stands for “as low as reasonably achievable,” which sounds sensible. But in practice, it’s ambiguous, and has been used for decades to greatly increase the cost and build time for nuclear power plants, when not denying permission to build them at all. This doesn’t mean nuclear energy will suddenly start pumping out radiation—indeed, every plant operates far below NRC (Nuclear Regulatory Commission) exposure limits. It just means that the limits will actually be numbers, instead of a vague standard. This will bring certainty, and certainty lowers costs and speeds up timelines. Apparently ALARA is also used by the NRC, so there may yet be some work to do on this.

Nuclear power plant company Natura Resources and NGL Energy have announced a strategic agreement to pair Natura’s 100 megawatt molten salt reactor (MSR-100) with industrial scale water desalination in the Permian Basin, transforming oilfield produced water—the salty, oily byproduct of oil and gas extraction—into clean water for data centers, agriculture, and industry while extracting critical minerals for batteries and other technology. NGL Energy, a midstream energy company (they deal with pipelines) operating water infrastructure across major U.S. shale basins, currently manages massive volumes of produced water through pipelines and disposal wells. The Permian Basin generates over 20 million barrels of produced water daily, and managing this volume through traditional disposal down wellbores has faced increasing regulatory scrutiny in some places. Natura’s Generation IV reactor operates at atmospheric pressure using liquid fuel dissolved in molten salt, providing high grade heat over 600°C (1,112°F) necessary for efficient thermal desalination while eliminating the need for massive containment structures. The company is deploying their first reactor in the U.S. later this year—a 1 megawatt MSR-1 (molten salt reactor) for research at Abilene Christian University in late 2026—with the commercial MSR-100 deployment targeted for 2029.

U.S. company Deep Isolation is starting up a full scale test of how their technology can safely and permanently dispose of nuclear waste down deep horizontal wellbores. This won’t be the solution for all nuclear waste—honestly, we should just recycle most of it—but it’s an excellent idea, and expands our options.

Another American company, NextEra, is planning on adding a lot of gigawatts of generating capacity from SMRs “at its existing nuclear power plant sites or potential new sites, primarily to meet demand from data centers.” They’re aiming for at least 30 GW by 2030 for this purpose.

The first unit at China’s Taipingling plant is starting up, having attained “a sustained chain reaction for the first time.” That’s another GW (just over) for China’s electric grid. Seriously, good for them.

“On time and under budget.” Wow, that never happens for big projects, especially nuclear. But the Canadian province of Ontario just managed to refurbish a reactor at their Darlington nuclear power plant four months ahead of schedule. Seriously, that’s the sort of thing we need to kick the Nuclear Renaissance into high gear and catch up with China. Also in Canada, the province of Saskatchewan and their primary electrical utility SaskPower are now evaluating building large nuclear power plants. They’re already moving forward with SMR plans.

South Korea is planning on building two big nuclear power plants that should come online in 2038. It’s a ways off I know, but it’s nice to see the long term planning on it.

After shutting down many of its nuclear power plants after the tsunami and nuclear accident (which contrary to popular belief killed no one) at Fukushima, Japan it restarting a reactor at the world’s largest nuclear power plant, which happens to be at Fukushima.

On the list of what not to do, Germany has completed the removal of steam generators at its already shut down Grafenrheinfeld nuclear power plant. It’s absolute insanity to take a working source of nuclear energy, and dismantle it.

The State of New York has done a complete 180 on nuclear power in the last few years, going from closing down the Indian Point Energy Center back in 2021, to now aiming for 5 GW of nuclear power. Hopefully they stick to their plans. Maybe Germany will smarten up too, one can always hope.

Fusion Energy

Last year set a new record for investment into fusion energy companies, over $4 billion. You can see that China led the charge, with the States in second place, and Germany coming in at a very distant third. The first fusion funding announcement in 2026 was also from China, with Startorus Fusion raising a 1 billion yuan (~$143 million USD) Series A, which actually makes it “the largest single round for a ‘private’ Chinese fusion company to date.” I also just learned that China has ETFs (Exchange Traded Funds) allowing investment in the fusion industry. These fusion ETFs focus on “enabling infrastructure required for fusion deployment at scale. Constituents include suppliers of superconducting material, grid control and power electronics, precision optoelectronics, and heavy nuclear manufacturing…” Honestly, good idea. China is working on the whole fusion ecosystem, including investing.

The U.S. Department of Energy has certified the fusion power plant design of Helios, which will be built by Thea Energy, “confirming its feasibility to put energy on the grid.” The plan is for the stellarator based power plant to be up and running in the 2030s.

Another fusion company is moving forward on building their pilot plant. Type One Energy “has submitted the initial licensing application in preparation for the construction of a fusion power plant at Tennessee Valley Authority’s former Bull Run fossil plant site in Clinton, Tennessee.” The Infinity Two stellarator fusion machine will produce 350 Mwe once it’s up and running, which could be in 2029. I’d like to give a big shoutout to the Tennessee Valley Authority, they’ve been extremely forward looking on both fission and fusion, with multiple gigawatts of fission also slated to be built under their purview.

Here’s a breakthrough for you: Pacific Fusion successfully tested a simplified fusion target design at Sandia National Laboratories that eliminates the need for expensive external lasers and magnetic coils, potentially cutting over $100 million from power plant construction costs. By precisely tuning aluminum casing thickness around fuel pellets, they allow magnetic fields to “leak” through and pre-magnetize the fuel. “This process traps heat more effectively, making it easier for the fuel to ignite during the main compression pulse.” The breakthrough addresses a critical commercialization challenge: inertial fusion power plants need to compress fuel pellets repeatedly for continuous electricity generation, and the new design requires only bullet casing level manufacturing precision rather than expensive, tight tolerance components. This will help reduce costs even further. Pacific Fusion targets net energy gain by 2030 and plans to build its demonstration system in New Mexico.

Canadian company General Fusion is going public via a SPAC. I believe it’s supposed to happen sometime this summer. The company has had a bit of trouble raising money lately, so this may be a good way forward for them. We shall see.

PPPL (Princeton Plasma Physics Laboratory) plans to use AI to help speed up the design of stellarator fusion machines, speeding up the timeline to commercial deployment of the technology.

Fusion Energy Base, a website “dedicated to providing comprehensive information on organizations advancing fusion energy and its supply chain,” has now added a huge supply chain directory for all things fusion. Any fusion startups—or companies in and around the fusion industry, go check it out.

South Korea has a national fusion strategy that’s is “proceeding along three parallel tracks.” First is a roadmap for a “Korean Innovative Fusion Reactor,” with power generated sometime in the 2030s. Second is a fusion research facility (nicknamed Artificial Sun City). And thirdly, a $4 billion program called the K-Moonshots Initiative will add fusion energy “as a key pillar.”

That’s it this time, but Nuclear Energy Updates will be back four-ish weeks from now. In the meantime, check your inboxes on February 21^st for the 36^th edition of Techno-Optimist.

Thank you all for reading — and until next time, keep your eyes on the horizon.

-Owen

Subscribe now

Welcome to Robot Updates #3, where I catch you up on the latest news about our future robot overlords (mostly kidding).

You may have noticed that there was no AI Updates newsletter last week. Sorry about that, unfortunately I wasn’t able to finish it in time, but next month’s will be extra interesting to make up for it.

Alright, let’s dive in.

Subscribe now

Just replace “scientific breakthroughs” with Robotics. We haven’t reached that sci-fi future of ubiquitous robots yet, but each little step gets us closer.

New and Notable

XPENG’s New Iron humanoid looks pretty impressive just taking a walk. Multiple companies are really mastering a more natural gait. Though apparently it crashed shortly after this segment of video, so not perfect yet.

It was inevitable really. Give this short clip a watch, it shows China’s army “training for robot and drone warfare.” Dozens of unmanned ground robots and drones were used, with the focus being on “speed, saturation, & AI-decisions.”

Thankfully, warfare isn’t the only use case. Search and rescue is going to benefit immensely for better robots, as this Deep Robotics quadruped demonstrates.

Ginko Bioworks CEO Jason Kelly recently gave a talk on their plan to build autonomous robotic laboratories, with the goal being to replace manual ones in the next 5 years. Ambitious? Absolutely, and maybe a little aspirational on their timeline. But their idea and direction is probably mostly correct.

There’s a new “Roomba” in town. The Roborock Saros Rover isn’t from Roomba, but it’s basically a Roomba with legs, which falls into the cute and possibly useful category, as now it can get itself unstuck more easily, and climb and clean stairs.

NVDIA has launched Isaac GR00T N1.6, “an open reasoning vision language action (VLA) model, purpose-built for humanoid robots, that unlocks full body control and uses NVIDIA Cosmos Reason for better reasoning and contextual understanding.” Basically, this is for training the robots, and because it’s NVIDIA I’m sure it’ll be extremely effective.

China just isn’t slowing down in robotics. Check out this video here of a solar powered robot clearing ice by gliding “along a power transmission line at an altitude of 3,300 meters in Sichuan Province, southwest China.” Here’s another interesting one of robots from LimX Dymamics unboxing themselves. And UBTech’s Walker S2 playing tennis with a human.

In the Home

Figure Robotics has just taken a big step forward with their Helix 02, saying that “It’s our most powerful model to date - it’s using the whole body to do dishes end-to-end and it’s fully autonomous.” It’s not at Jetsons level yet, but it’s well on the way, and by far the closest I’ve seen.

American company Fauna Robotics has released their humanoid called Sprout. It looks more like something created by Disney than The Terminator, which is always nice to see. I figure it’s a good sign that kids are putting stickers on it.

Another American player designing humanoid robots for home use is 1x, and their NEO model is starting to be able to learn new tasks on its own by doing them.

In the Factory

Hyundai Motor Group’s union has declared “not a single robot will be allowed onto the factory floor without a labor-management agreement,” making it the first union at a global automaker to openly oppose humanoid robots on production lines. The standoff erupted after Hyundai unveiled plans at CES 2026 to deploy Boston Dynamics’ Atlas humanoid robots starting in 2028 at its Georgia plant, with the robots able to handle complex tasks by 2030. The union estimates that three workers running 24-hour shifts cost about $208,200 total annually, while a humanoid robot requires only maintenance and electricity after purchase—creating an economic incentive toward labor replacement that union leaders call an “employment shock.” I hate to say it, but I think they’re fighting against an inevitability that will eventually replace almost all humans in factory work. In the long run, that’s probably a good thing, but in the short to medium term it really is going to be disruptive—to put it mildly. Those Boston Dynamics humanoid robots, the plan is to produce 30,000 of them in 2026. That’s going to start making an impact sooner rather than later.

They aren’t the only one, with UBTech’s Walker S2 humanoid robots starting to make inroads into factories as well. The robots currently match just 30-50% of human productivity and only in specific tasks like stacking boxes and quality control, according to chief brand officer Michael Tam. Yet manufacturers including BYD and Foxconn are racing to order them anyway, driven by first-mover advantage fears. UBTech aims to boost Walker’s performance to 80% of human efficiency by 2027 and scale from 500 units delivered in 2025 to 10,000 in 2026, with analysts noting that 80% might suffice since robots don’t need breaks or holidays, just maintenance. The bet isn’t on perfect robots today—it’s on learning faster than competitors while the technology matures. UBTech has also “reached a new agreement with European aviation leader Airbus to supply robots for use inside aircraft manufacturing facilities,” so these robots obviously have broad appeal in factory settings.

Recommendations

Not Boring by Packy McCormick

Many Small Steps for Robots, One Giant Leap for Mankind

Welcome to the 1,179 newly Not Boring people who have joined us our last essay! Join 256,826 smart, curious folks by subscribing here…

Read more

5 months ago · 138 likes · Packy McCormick

I came across this essay on robotics that’s really quite excellent, well worth a read. Though it wouldn’t hurt to try and put it into some sort of auto reader, as it is rather long. As the author says, “Spend an hour reading this weekend and I think you’ll know more about robotics than 99% of people, including some people who invest in robotics.”

That’s it for now, but Robot Updates will be back in your inboxes next month. Thank you all for reading—and until next time, keep your eyes on the horizon.

-Owen

Subscribe now

Welcome to the thirty-fifth edition of Techno-Optimist, your destination for all the latest updates and commentary on space, science, technology, medicine, and much more.

As this is the first edition of Techno-Optimist in 2026, I’ve decided to it might be fun to have some polls for everyone to make their predictions for the year. There’s a lot of them, so I’ve put them right at the end. We’ll revisit them in the first newsletter of 2027 and see how we all did. Please do jump into the comments as well with your predictions if I don’t cover them in the polls.

Subscribe now

“When you want to build a ship, do not begin by gathering wood, cutting boards, and distributing work, but rather awaken within men the desire for the vast and endless sea.”

—Antoine de Saint-Exupéry

Privately Funded Four Telescope Observatory System. Former Google CEO Eric Schmidt and his wife Wendy have announced one of the largest private astronomy investments in history—the Eric and Wendy Schmidt Observatory System—comprising four next-generation telescopes that will all be operational by 2029. Together, these facilities represent a fundamental shift in how cutting edge astronomy gets funded and built, with Schmidt promising capabilities rivaling NASA flagship missions at a fraction of the cost and timeline.

Lazuli Space Observatory is the crown jewel, a 3.1 meter (10.2 ft) space telescope with 70% more light collecting area than Hubble’s 2.4 meter (7.9 ft) mirror, making it the largest privately funded space observatory ever built. The telescope features three advanced instruments: a wide field optical imager for surveying large swaths of sky, an integral field spectrograph for detailed chemical analysis, and a high contrast coronagraph optimized for directly imaging exoplanets around Sun-like stars. Lazuli will launch into a highly elliptical lunar resonant orbit reaching 275,000 kilometers at apogee—far beyond the crowded low Earth orbit where communications satellites reside—and use an innovative off-axis three mirror design that eliminates the light blockage and diffraction caused by traditional telescope configurations. The mission will study everything from exoplanet atmospheres and supernovae to the Hubble Tension (the mystery of why different measurement methods yield different values for the universe’s expansion rate), with Pete Klupar, executive director of the Lazuli project, declaring “we’re going to do it in three years, and we’re going to do it for a ridiculously low price”—estimated at hundreds of millions rather than the $10+ billion NASA flagship missions cost. The telescope will likely launch on Relativity Space’s Terran R rocket from Cape Canaveral, with Schmidt having recently become CEO of Relativity Space after investing in the company.

Argus Array takes a radically different approach to ground-based astronomy by deploying 1,200 small 11-inch (28-centimeter) telescopes packed inside a bowl shaped structure on a Texas mountaintop, working together to create the equivalent light gathering power of a single 8 meter (26.2 ft) telescope. Named after the hundred-eyed giant Argus Panoptes from Greek mythology—and hopefully also as a nod to the Argus Array subspace telescope from Star Trek—the array will continuously photograph the entire visible Northern sky; covering 8,000 square degrees, roughly the field of view of the human eye, and imaging some regions every single second to generate a staggering 7.8 petabytes of data per night. Unlike traditional telescopes that take deep images of small patches of sky every few days, Argus creates a real-time movie of the cosmos, perfect for catching transient events like supernovae within minutes of their explosion, neutron star collisions, black holes tearing apart stars, and optical counterparts to gravitational wave detections from LIGO. Because it archives the preceding data, astronomers can rewind the cosmic movie to see what happened before an event was detected, essentially creating a time machine for astrophysics. The array should achieve first light in 2027, becoming fully operational by 2028. British billionaire financial trader Alex Gerko is co-financing this project with Schmidt Sciences.

Deep Synoptic Array deploys 1,656 radio dishes, each 1.5 meters (5 feet) across, spread across a massive 20 × 16 kilometer (12.4 x 9.9 mile) site in Spring Valley, Nevada; a radio-quiet valley ideal for detecting faint cosmic signals. The array uses innovative astronomical radio receivers developed at Caltech that eliminate the need for complex cooling systems, dramatically reducing costs. Operating at radio wavelengths invisible to optical telescopes, DSA will take snapshots of the entire radio sky every 15 minutes, making it unmatched at detecting and precisely locating fast radio bursts—mysterious millisecond blasts from across the universe whose origins remain one of astronomy’s biggest puzzles. Project lead Gregg Hallinan of Caltech emphasized the scale: “Every radio telescope ever built has detected about 10 million radio sources. We’ll double that in the first 24 hours.” The array will also discover pulsars (rapidly spinning stellar remnants whose metronomic pulses can detect gravitational waves), probe dark matter, and reveal galaxy centers and black holes obscured by interstellar dust that blocks optical observations. Expected to come online in 2029, DSA represents an order of magnitude improvement in survey speed over any existing or planned radio telescope.

Large Fiber Array Spectroscopic Telescope (LFAST) uses a modular system of 76 centimeter (30 inch) mirrors organized in groups of twenty telescopes mounted on a single structure, with optical fibers feeding light from each mirror to a spectrograph that splits the light into its component wavelengths. The design is infinitely scalable, with one 20-telescope unit equals a 3 meter (9.8 ft) telescope’s collecting area, ten units match the twin 10 meter (32.8 ft) Keck telescopes in Hawaii, and the team aims for enough units to rival the 30 meter (98.4 ft) Extremely Large Telescope under construction in Chile (or even Europe’s 39 meter ELT). By splitting starlight into spectra, LFAST will study the chemical composition of exoplanet atmospheres hunting for biosignatures, identify the nature and elements of distant galaxies, and provide high resolution spectroscopy for astronomy. The preferred site is Kitt Peak in Arizona, with Mount Lemmon and Mount Hopkins as backups. The modular approach housed in compact “mini domes” slashes construction costs while maintaining flexibility—more units can be added as budgets allow and science demands grow.

What makes Schmidt’s approach revolutionary is the coordinated “system” philosophy. All four telescopes will work together through a common online interface with linked databases, allowing seamless follow up observations. Argus might detect a supernova, LFAST could immediately capture its spectrum to determine composition, DSA could search for associated radio emissions, and Lazuli could study the explosion’s optical evolution over weeks—all coordinated in near real time. The observatories embrace open science: Argus and DSA will release data publicly within minutes of collection, while LFAST and Lazuli will accept proposals from scientists worldwide. Schmidt Sciences President Stuart Feldman joked their success metric is “Nobel Prizes per year,” but more seriously emphasized they’re measuring “exciting outcomes that would have been unlikely to have arisen without our resources.”

By taking on projects with calculated risks that government agencies can’t stomach, deploying current technology rather than waiting decades for perfection, and building quickly before technology becomes obsolete, Schmidt is proving private philanthropy can complement NASA’s scientific endeavors. The entire system costs a fraction of what NASA spends on flagship missions like the James Webb, yet promises transformative discoveries about exoplanets, dark energy, gravitational waves, and explosive transient events. Welcome to astronomy’s new era, where one billionaire can build an entire observatory system faster and cheaper than traditional government programs—and make all the data freely available to scientists everywhere.

Anti-aging Roundup. Writing this newsletter has given me a good broad overview of what’s happening in many fields, including anti-aging. So here’s a little summary of the pillars of anti-aging as I see them: epigenetics (being able to “reset” some or all of it). Senescent cells, they increase bad inflammation and cause all sorts of other problems, including promoting cancer. Getting rid of them, or at least reducing them. But some may be useful for certain things, so it may be a matter of getting rid of certain groups of them, or maybe just reducing the numbers below a certain level. NAD+ levels seem improving for both the brain and the body. Mitochondria, their number, and healthy function. I’m sure there’s more, but all of these are definitely part of it.

Here’s a few news items I noticed recently:

Retro Biosciences, the high-profile startup backed by OpenAI’s Sam Altman, has officially moved from the lab to the clinic, dosing its first human participant in a Phase 1 trial for its lead longevity candidate, RTR242. This therapy targets autophagy—the body’s cellular trash disposal system—by restoring the acidity and function of lysosomes. As we age, these cellular cleaners become sluggish, leading to the buildup of toxic protein waste that can trigger neurodegeneration, and other symptoms of aging (yes, we should think of aging as a disease). By rebooting this waste management cycle at its source, Retro aims to not just manage symptoms but to effectively rejuvenate cells to a more youthful state. While the initial trials are focused on Alzheimer’s, the ultimate goal is to add 10 healthy years to the human lifespan.

Researchers at the University of Illinois Chicago have discovered that a decline in the protein platelet factor 4 (PF4) is a primary driver of immune system aging. In youth, PF4 acts as a biological regulator that keeps blood stem cells from dividing too rapidly; as levels drop with age, these stem cells multiply uncontrollably and accumulate mutations linked to cancer, chronic inflammation, and heart disease. By restoring PF4 in older mice and human stem cells, the team successfully turned back the clock, causing “aging blood and immune cells behave strikingly younger again.”

Anti-aging company New Limit has invented their first medicine: it “makes your liver a decade or two younger (best guess today). We’ll spend another couple years optimizing and then launch human clinical studies. In parallel, we are working on anti-aging medicines for the immune and vascular system. In a few years, we will we working on many more tissue types.” Co-founder Blake Byers noted that “This is all happening much faster than we anticipated when we started.”

Progress against Alzheimer’s. Alzheimer’s research is really hitting its stride, and I think we’ll be seeing material progress against it soon. My hope (based on a gut feeling of the rate of progress) is that we see a functional cure for most cases by the mid 2030s. Here’s a few new highlights: A new experimental drug called NU-9 halts the disease in animal models before symptoms appear. Maybe this sort of early intervention—not 100% sure how you’d diagnose it, maybe those at high risk—could prevent or delay disease onset.

I talked here about progress against Alzheimer’s and Dementia.

Neural NeXus

Top 5 Biotech Breakthroughs that shaped 2025

Welcome to 2026, Readers! Let’s stay relentless this year…

Read more

5 months ago · 29 likes · 8 comments · David Kingsley, PhD, Stephen Turner, Owen Lewis, BioDodo, and Tony Huge Official

The most promising breakthrough I highlighted there actually just came out. Researchers have reversed advanced Alzheimer’s disease in mice by restoring the brain’s energy balance through a molecule called NAD+, which naturally declines with age and drops even more severely in Alzheimer’s patients. Mouse models showed complete cognitive recovery even after the disease was already advanced. The treated mice showed normalized blood biomarkers for Alzheimer’s, and complete recovery of all cognitive functions. The key takeaway is a message of hope: the effects of Alzheimer’s disease may not be inevitably permanent. The damaged brain can, under some conditions, repair itself and regain function.

Another very interesting finding came from analyzing data from over 450,000 participants, which discovered that 72-93% of Alzheimer’s cases and approximately 45% of all dementia cases are attributable to just two variants of a single gene. Lead researcher Dylan Williams emphasizes that “if we knew how to reduce the risk that the ε3 and ε4 variants confer to people, we may be able to prevent most disease from occurring.” Clearly this is a job for gene therapies, which once developed could prevent or even cure the majority of all Alzheimer’s cases, as well as almost half of dementia diagnoses.

The New Pyramid. For nearly five decades, Americans followed dietary guidance built on shaky science and industry influence. The 1977 McGovern Committee guidelines—drafted by staffers with no scientific background over many researchers’ objections—launched the low fat era that told us to eat 6-11 servings of bread daily while using fats “sparingly.” The food industry reformulated everything to be low fat, replacing fats mostly with sugar. Americans complied, fat consumption dropped, carb consumption soared, and obesity rates tripled. By the 2010s the scientific consensus that fats were bad had cracked, but institutional inertia kept the broken pyramid in place. Until now.

The federal government just published new Dietary Guidelines at realfood.gov that inverts the old, bad advice. The newly designed new pyramid prioritizes protein, dairy, and healthy fats at the base—recommending 1.2-1.6 grams of protein per kilogram of body weight daily from meat, eggs, seafood, full fat dairy [Sidebar: I came across a recent study linking full fat cheese and cream to a lower risk of dementia], nuts, and avocados. Vegetables and fruits come next (3 and 2 servings daily), with whole grains reduced to 2-4 servings and refined carbohydrates explicitly discouraged. The guidelines explicitly call out “the dangers of highly processed foods” and define real food as “whole, nutrient dense, and naturally occurring,” prepared without added sugars, industrial oils, artificial flavors, or preservatives. Added sugars get eliminated entirely, especially for children.

The website itself signals the shift—it’s a government site that actually looks pretty good and even more surprisingly, it works well (designed by the National Design Studio). The timing matters: 50% of Americans have prediabetes or diabetes, 75% have at least one chronic condition, and vast amounts of healthcare spending goes to treating diet linked chronic disease. What makes this most optimistic isn’t just that we fixed an obviously broken thing almost everyone agreed was broken—it’s proof that institutional sclerosis isn’t permanent. If we can overturn 47 years of entrenched nutritional dogma, backed by bureaucratic momentum, what else can we fix?

Space

• GRU Space is aiming to build infrastructure on the Moon. The plan starts with mining lunar regolith to construct a pressurized Moon habitat—initially for tourism—but quickly scaling to full lunar bases with roads, warehouses, and mass drivers, then repeats the model on Mars. The long game is ownership, reinvestment, and large scale resource utilization across Solar System and beyond. Their ambitions aren’t small, with the end goal being Galactic Resource Utilization (hence the name). Very pro human, which is great to see. (GRU Space)

• Back on December 18^th, New Shepard Flight 37, Michaela (Michi) Benthaus became the first person in a wheelchair to go to space. Just step back and think about that for a minute. When space flights first started, it was (with good reason) only the fittest and most accomplished that could go. But now, as things have grown safer, and we’ve gained more experience, the door is opening up. There’s hope for all of us who want to go to space one day, even those with disabilities or other health problems. Of course, I’m hopeful medical technology will fix most of that this century, but I digress. (Blue Origin 1) (Blue Origin 2)

• NASA’s newest exoplanet observatory, the Pandora small satellite, has successfully separated from its SpaceX rocket and entered a sun-synchronous orbit to begin science operations. Pandora will spend the next year observing at least 20 known exoplanets— taring at their transits to tease out the signatures of water vapor, clouds, and other atmospheric features while simultaneously monitoring their host stars to separate stellar noise from genuine planetary signals. Because Pandora collects visible and infrared data together over long durations, it will help clean up ambiguities in exoplanet spectra and guide follow up studies with larger telescopes like Webb. (NASA’s Kennedy Space Center)

• The gravitational pull of Mars appears to subtly shapes Earth’s long term climate cycles, including the rhythms of ice ages that have come and gone over millions of years. Simulations show that if Mars were absent, two of the major Milankovitch cycles—the ~100,000 year and ~2.3 million year patterns that influence Earth’s orbital shape and axial tilt—disappear. Those orbital and tilt swings change how sunlight is distributed across the globe and are a key driver of ice sheet growth and retreat, so Mars’ tug has directly helped shape the cadence of past ice ages on our planet. BBC Sky at Night Magazine)

• NASA’s Hubble Space Telescope has confirmed the first example of a new class of cosmic object: a starless, gas rich, presumably dark matter dominated cloud nicknamed Cloud-9 that sits near the spiral galaxy Messier 94. This type of object is officially called a Reionization-Limited H I Cloud (RELHIC), and was long predicted by cosmological models but never directly seen until now. It appears to be a “failed galaxy” that gathered gas (but not enough) and dark matter, but never ignited stars. Cloud-9 gives astronomers a rare window into how galactic type structures formed in the early universe, but without the interference of starlight. (Hubble)

• For the first time ever, astronomers have measured the mass of a rogue planet wandering the Milky Way by observing a microlensing event simultaneously from Earth and space, combining ground surveys with data from the Gaia space telescope to nail down both its mass and its location. The object is about 22 % the mass of Jupiter (Saturn-like), and located roughly 3,000 parsecs (~9,785 light years) from the galactic center. It likely formed around a star before being flung into interstellar space by gravitational chaos within its system, instead of forming alone like a brown dwarf, which are considerably larger. This measurement opens a new window on the population of rogue planets, helping scientists understand how planetary systems eject worlds and how common these nomads really are. (Science Magazine) (Science Daily)

• The Rubin Observatory has discovered an incredibly fast spinning large asteroid. Astronomers have christened the asteroid 2025 MN45; it’s about 0.4 miles (710 meters) in diameter, and rotates once every 1.88 minutes, making it the fastest spinning rock of its size ever observed and implying it’s far stronger than the typical “rubble pile” asteroids we thought dominated the belt. In the same dataset Rubin found dozens of other super and ultra fast rotators, showing how the observatory is revealing previously unseen asteroid populations in our Solar System. These early results are just a taste of the discoveries Rubin will generate once its full survey kicks into gear. (NSF-DOE Rubin Observatory)

• I’ve got more on SpaceX for you. Some of these numbers are just so phenomenal I can’t not talk about it. For 2025, the company lofted 90% of all mass to orbit, or 9x the rest of the world combined—including other American companies. They totaled 170 launches, 165 Falcons, and 5 Starship tests. Their Starlink satellite network doubled their active new customers from 4.6 million to 9.2 million. Speaking of Starlink, the FCC has now authorized SpaceX “to deploy & operate up to 15,000 next-gen satellites — enabling even better, faster, and more advanced Internet services in the country.” Hot on the heels of this, China has announced a plan to put up 200,000 satellites. I don’t normally say this about tech buildouts, but that might actually be too much.

Engineering, Machines, & Physics

• China has activated CHIEF1900, the world’s most powerful hypergravity centrifuge capable of generating forces 1,900 times stronger than Earth’s gravity—dethroning its predecessor CHIEF1300 just four months after it broke the previous American record. Located 15 meters underground at Zhejiang University to minimize vibration, the massive machine “compresses space and time” by allowing researchers to simulate decades or centuries of geological processes in hours—a three meter dam model spun at 100g experiences the same stress as a 300 meter (~984 ft) real world dam. Scientists there can now recreate catastrophic events like earthquakes and dam failures in the lab, study how pollutants migrate through soil over millennia, or test how high speed rail tracks resonate with the ground. The $285 million facility represents China’s rapid scaling of experimental capabilities. (Interesting Engineering)

• Micron officially broke ground on its $100 billion manufacturing complex in Clay, New York—the largest private investment in New York state history and the biggest semiconductor facility in the United States. The megafab will eventually house up to four 600,000 square foot cleanrooms (2.4 million square feet total), creating 50,000 jobs in New York and supporting 90,000 across the country, with production starting in 2030 and ramping up through the decade. The facility will produce the world’s most advanced memory to meet exploding AI demand, with tech leaders from NVIDIA, Google Cloud, AWS, and AMD emphasizing that high performance memory has become essential infrastructure for AI systems. Micron aims to produce 40% of its DRAM in the United States through this complex and companion fabs in Idaho and Virginia, ending America’s dependence on foreign semiconductor manufacturing. [Sidebar: Dynamic Random Access Memory is the primary, volatile memory in computers, laptops, and phones that provides high-speed, temporary storage for operating systems and active applications.] (Micron) (U.S. Commerce Department)

• Samsung has officially started mass production of its solid-state batteries, signaling the end (or at least reduction) of range anxiety for electric vehicles. Their power cells offer a 600 mile (~966 kilometer) range and can recharge in nine minutes, close to the convenience of a traditional gas station stop, but with even better range. By replacing liquid electrolytes with solid components, Samsung creates a battery that is far more energy dense and also significantly safer—they don’t catch on fire—and more durable than current lithium ion technology. (Good News Network)

• Norwegian company Flocean is tackling water scarcity by launching the world’s first subsea desalination plant sometime this year. This modular equipment sits 600 meters (~1970 ft) deep to harness the high hydrostatic pressure, slashing energy consumption by 50% compared to traditional coastal desalinization plants. By operating in the deep sea, the system avoids surface pollutants and eliminates the need for chemical pre-treatments, delivering freshwater with less impact. (Interesting Engineering)

• China is rewriting the rules of wind energy with the successful maiden flight of the S2000, the world’s first megawatt class airborne wind turbine. Designed for urban environments and resembling a futuristic airship, this 60 meter (197 ft) long floating wind turbine ascended to 2,000 meters (6,560 ft) above Sichuan Province to capture high altitude jet streams that are several times more powerful and consistent than surface breezes. It uses 40% less material than surface wind turbines, and can be deployed to remote or disaster-stricken areas quickly if needed. Has anyone seen Big Hero 6? These are basically a scaled up version of the ones in that movie. (Interesting Engineering) (People

Medicine & Biotech

• Medical science just delivered a massive win for children battling spinal muscular atrophy (SMA) with the FDA approval of the gene therapy Itvisma. By injecting a healthy gene directly into the spinal fluid, doctors bypass the weight (and therefore age) based safety limits of previous treatments, making this a “one-and-done” solution for patients over the age of two. Clinical trials show remarkable results, as children who previously struggled to move are now rolling, sitting up, and even climbing stairs. This breakthrough signals a new era where we no longer just manage genetic diseases but actively rewrite genes to create brighter futures for those once deemed untreatable. (Singularity Hub)

• Scientists have unlocked a new weapon against traumatic brain injury in the form of a tiny peptide called CAQK that homes in on damaged brain tissue. This compound acts as a biological GPS, traveling through the bloodstream to lock onto injured areas and actively reducing inflammation and cell death. While researchers originally viewed the molecule as a delivery vehicle for other drugs, new trials show it possesses its own powerful healing properties that significantly improve functional recovery. This breakthrough paves the way for simple, non-invasive IV treatments that could stop permanent brain damage in its tracks immediately after an accident. On another note, I’m starting to dig into peptides, it’s a very interesting area for potentially improving both preventative measures and medical treatment. (SciTechDaily)

• University of Hong Kong scientists have identified a “mechanical switch” protein in bone cells called Piezo1 that triggers the body to build bone in response to exercise. By activating it with a chemical compound, researchers successfully mimicked the bone strengthening benefits of a workout in mice without them moving a muscle. This breakthrough opens the door to an “exercise-in-a-pill” that could reverse osteoporosis and prevent fractures in bedridden patients or those with limited mobility. We are nearing a future where we can chemically program the body for resilience, ensuring strong bones remain a standard human feature regardless of age or physical ability. (SciTechDaily)

• Scientists at the University of Maryland Baltimore County have mapped the precise interaction between viral RNA and the proteins that drive infection, exposing a major vulnerability shared across the enterovirus family (which includes viruses that cause polio, encephalitis, myocarditis, and the common cold). By capturing high resolution images of the structures of this replication machinery, researchers identified a highly conserved binding mechanism used by the viruses. This discovery allows engineers to now design precise drug molecules that break the machinery, offering the prospect of a single, universal antiviral for this virus family. (SciTechDaily)

• Researchers have unlocked a way to supercharge cellular energy by boosting mitochondrial health without the need for strict calorie restriction. By identifying a specific genetic pathway, the team successfully tricked cells into entering a high-performance “repair mode” that normally only triggers during fasting. This discovery suggests we can engineer the longevity benefits of a sparse diet into a simple therapeutic intervention, maybe improving physical vitality as we age. (Phys.org via @kimmonismus)

• Strides against cancer continue to be made. An important new discovery was that some people have a gene which reduces their risk of leukemia. Elsewhere, scientists have engineered a Trojan Horse nanoparticle that infiltrates aggressive tumors to reprogram their immune suppressing environment into a zone that attacks cancer. By simultaneously delivering a targeted drug and an immune booster, this therapy forces the cancer’s own support cells to turn against the tumor, effectively stripping away its cloaking device and allowing the body’s immune system to destroy it. In South Korea, a team has developed the world’s first AI model that predicts B-cell responses to create “cancer-remembering” personalized vaccines. By using AI to identify neoantigens that trigger both immediate T-cell attacks and long-term B-cell immune memory, this technology aims to not only destroy tumors but also provide lasting protection against recurrence. The team is now preparing for FDA clinical trials in 2027 to turn this approach into a global standard for precision oncology. Speaking of cancer vaccines, apparently in Britain there are “13 clinical trials happening right now (seven of which phase 2, one phase 3) across 11 cancer types — including melanoma, pancreatic and colorectal.”

• Researchers at China Pharmaceutical University have discovered that a synergistic pairing of two common drugs—silybin and carvedilol—effectively reverses liver fibrosis, a condition previously thought to be incurable. While silybin protects liver cells from damage and inflammation, it struggles to stop scarring on its own; however, adding the heart medication carvedilol directly deactivates the scar forming cells that drive the disease. This breakthrough outperformed all existing treatments in experimental models, significantly reducing collagen buildup and restoring liver function. (Science Daily)

• Peking University scientists have identified a viral trick called “migrions” that turns the body’s own moving cells into high speed delivery vehicles for infection. Instead of traveling as solo particles, the viruses are bundled into large, mobile packages (migrions), and passed directly to new cells by infected ones, allowing the viruses to jump-start replication by delivering a massive viral payload to new cells all at once. This discovery explains why certain infections turn aggressive so quickly, as the collective delivery mechanism completely bypasses traditional receptor limits and speeds up the takeover of healthy tissue. By discovering this route for viral infections to spread, we gain a critical new target for neutralizing it. I wonder if this strategy could be coopted to treat diseases too? I can imagine this being used to ensure more cells get a gene therapy, or delivering drugs to the cells that need them most. (Science Daily)

• A team from UNSW Sydney have mastered a gentler form of CRISPR that restores gene activity without making a single cut to our DNA. These epigenetic edits use enzymes to remove the chemical methyl tags that act as molecular brakes, researchers successfully reactivated the fetal globin gene, providing a good treatment for Sickle Cell disease. This epigenetic breakthrough settles a decades long debate by proving these markers actively silence our genetic code rather than just marking inactive territory. Being able to precisely tune our biology to reverse inherited disorders with zero risk of cancers or other errors associated with traditional gene editing is a big step forward. (Science Daily)

Agriculture

• Chinese researchers at the Shapotou Desert Experimental Research Station have engineered a “living skin” made of drought resilient blue-green algae to reclaim 6,667 hectares of the Ningxia desert. This breakthrough involves casting specialized cyanobacteria into “seeds” that lie dormant until the first drop of rain, at which point they rapidly bloom to glue shifting sands into a stable biological crust. Within a single year, this microbial vanguard creates a foundation for soil that can withstand 22 mph winds and paves the way for natural plant succession. (Interesting Engineering)

Weird & Wonderful

• A new study in Science is changing the idea that hyper-specialization is the only path to the top, showing that early multidisciplinary exposure is the true engine of elite performance. While child prodigies in single disciplines often grab an early lead, this new analysis of Nobel laureates and Olympic medalists reveals that those who explore multiple fields early on eventually overtake and outperform narrow specialists. These “late bloomers” leverage their broad foundations to bypass the plateaus that trap early specialists, reaching a level of world class excellence that generally remains out of reach for those more narrowly trained. So parents, let your kids pursue all sorts of interests, don’t feel like they have to pick one thing and do nothing else. (Valerio Capraro, Science)

• French foundry Atelier Missor doesn’t just want to build giant statues in America, but on the Moon too! “Building a statue on the Moon would show the world that the West is not merely utilitarian, but striving for something else. Something higher. Low gravity, no wind, no oxidation… it seems like the perfect place for a gigantic titanium statue!” It’s a bit aspirational at this point, but I’m rooting for them. (Atelier Missor)

• Scientists at the University of Florida have confirmed that the central premise of Jurassic Park is surprisingly accurate: mosquitoes are DNA archiving machines. By analyzing the blood meals of over 50,000 mosquitoes, researchers identified the genetic signatures of 86 different animal species—representing 80% of the local vertebrate biodiversity—ranging from tiny frogs to big deer. This breakthrough turns the world’s most hated insect into a powerful, non-invasive conservation tool that can map entire ecosystems more efficiently than human teams or camera traps. (Science Alert)

• Sometimes it’s the little things that make a big difference. Whisper Aero, founded by former NASA electric aircraft pioneer Mark Moore, unveiled its Tone T1 at CES 2026—the world’s first quiet leaf blower using proprietary aerospace propulsion technology. The handheld blower operates at just 52 decibels (70% quieter than gas powered models), runs up to 50 minutes, and even includes an LED for nighttime cleanup. The company developed the technology while engineering quieter electric airplane engines and realized it had applications beyond aviation. (Owen Gregorian)

Photos & Videos

This is incredible. It’s a 25 year video showing the expansion of Kepler’s Supernova Remnant, named after the famous German astronomer, and first spotted in 1604. (NASA)

We live in a stunningly beautiful universe. This piece of photographic art taken by the Hubble Space Telescope is of “the star-forming complex N159, which is found in the Large Magellanic Cloud, approximately 160,000 light-years away.” (Hubble)

Metal “snowflakes”! “The authors argue that both their metal crystals and snowflakes are governed by the same emergence dynamics. Nanoparticles cluster together, or aggregate, to form larger superstructures. As a result, each of these metallic marvels is entirely distinct.” (Phys.org)

A panorama of a day on Mars, with the yellower light on the left being the sunset, and the bluish light on the right being the sunrise. (NASA Mars)

Casey Handmer just posted a 28m resolution global map of Mars, which can be uploaded to and viewed on Google Earth. Casey says that it’s “one of the most beautiful things I have ever made.” I agree! (Casey Handmer)

You know, the UAE sure knows how to celebrate the New Year in true techno-optimist fashion! (Hassan Sajwani)

Recommendations & Reviews

I’ve actually got a few things for you today. Couldn’t narrow it down, so you get three. First off, I should have mentioned this last time, but check out my latest work for Pirate Wires:

Pirate Wires

Breaking the Outer Space Treaty and Ascending to the Stars

“Don’t tell me that man doesn’t belong out there. Man belongs wherever he wants to go—and he’ll do plenty well when he gets there…

Read more

6 months ago · 46 likes · 2 comments · Owen Lewis

Elon Musk and Peter Diamandis recently had a long chat. These sorts of conversations are always interesting as you get an idea of where Elon wants to take things in the future. And where Elon wants to go, we’ll probably all be going, so listen in and see what the future might look like. You can tune in here on X if you want.

Here’s a little writeup for you all on peptides. I’m starting to dig into this a bit, so expect some more discussion as I learn more. Unlike broad spectrum drugs, these amino acid chains are being used (or designed in some cases) to repair specific cellular damage, boost metabolic efficiency, and even modulate the immune system with high accuracy. Something I didn’t realize is that Ozempic (a GLP-1 agonist) is actually a synthetic analog of a naturally produced peptide. So peptides really are on the map and big business. The “problem” is that companies can’t patent naturally occurring ones, so it’s harder for them to make a business case. Hence the grey market that currently exists for people doing self-treatment (or self-experimentation, which has a long scientific tradition). (Max Marchione)

Polls and Predictions for 2026

I think AGI won’t be a hard line where it didn’t exist before but does after. Much like the Turing Test that nobody even noticed was getting aced by multiple models, we’ll all just sort of agree one day that AI is good enough and general enough to call it AGI.
When will it happen though? Almost impossible to say, but I think the key will be how good these self-improving AIs are that are currently being worked on behind the scenes (if you don’t believe me, OpenAI is now looking to hire someone for managing exactly that). My personal bet is that there’s a low probability of a full AGI this year, with a moderate probability of a weak AGI. Almost certainly we see claims that this or that model is an AGI (because there’s already been such a claim from Japan). By 2027? Moderate probability of seeing one or more good AGIs being released publicly. Virtual certainty we see weak ones. How about 2028? High probability of seeing at least one good AGI in public. 2029-2030? Virtually certain one or more AGIs exist. After that, who knows. Singularity soon? When do we see ASI? Maybe that will end up being a bit more gradual (like 2035) too.

Here’s a few more launch predictions for the year: Falcon: 185 Flights. New Glenn: 6 flights. Starship: 12 flights (11 of which actually reach orbit, Flight 7 is not intended to). Rocket Lab manages 2 flights of their Neutron rocket. China successfully lands at least one, and probably several, orbital boosters. Putting themselves a decade behind SpaceX, but ahead of everyone else except Blue Origin. I don’t think it will take China a decade to catch up to where SpaceX is today however.

That’s it for this edition. Techno-Optimist #36 will hit on or around February 14^th (Valentine’s Day!). The plan is to have the usual shorter focused newsletters for AI, Robotics, and Nuclear Energy out on each of the next three Saturdays or Sundays.

Thank you all for reading—and until next time, keep your eyes on the horizon.

-Owen

Subscribe now

Welcome to Nuclear Energy Updates #2, your focused look at what’s new in nuclear energy, and how each development fits into the wider nuclear comeback.

I hope you all had a great Christmas and New Year. 2026 is already shaping up to be crazy—I don’t just mean politically (I always aim not to get into that too much on here), but technologically. The feeling of acceleration is real, but all of that AI and other advances need energy, and lots of it. I’ve always felt that nuclear—both fission and fusion—is the best way to supply our ever increasing demand, so I’m thrilled by the progress that has already been made and am looking forward to a lot more.

About today’s quote, obviously a bit of humor there drubbing on solar, which has its uses (especially in space!). But when it comes to reliable grid scale power here on Earth, Homer Simpson is absolutely right.

Alright, let’s dive in.

Subscribe now

“And Lord, we’re especially thankful for nuclear power, the cleanest safest energy source there is. Except for solar, which is just a pipe dream.”

—Homer Simpson

Nuclear Fission

Let’s start out with this excellent little graphic that explains major types of advanced nuclear reactors.

Big news first: The U.S. Department of Energy (DOE) will be pumping $2.7 billion into domestic uranium enrichment over the next decade.

The historic investment expands U.S. capacity for low-enriched uranium (LEU) and jumpstarts new supply chains and innovations for high-assay low-enriched uranium (HALEU) to create American jobs and usher in the nation’s nuclear renaissance.

…

Developing this new domestic production capacity for LEU and HALEU ensures an adequate fuel supply is available to maintain operations of the nation’s 94 commercial reactors and builds a strong base to supply future deployments of advanced nuclear reactors.

Three companies will split the money, each being given $900 million “to provide enrichment services for LEU and HALEU.”

American Centrifuge Operating will help create domestic HALEU enrichment capacity. General Matter will also work on domestic HALEU enrichment capacity, and Orano Federal Services will be expanding U.S. domestic LEU enrichment capacity. “The DOE also awarded an additional $28 million today to Global Laser Enrichment to continue advancing next generation uranium enrichment technology for the nuclear fuel cycle.”

Just in case you’re wondering what the difference is between LEU and HALEU fuels.

The provincial government of Ontario, Canada has announced “a $1 billion investment to build the G7’s first Small Modular Reactors at Darlington, securing Ontario’s clean energy future, and creating thousands of well-paying jobs.” For context, there are already traditional, large scale nuclear power plants at Darlington, making approval and building of new ones a much easier prospect. Ontario isn’t the only Canadian province going nuclear, with private company Alberta Energy partnering with Westinghouse to explore deploying “an advanced AP1000 modular reactor in Alberta.” While it is a modular reactor, it won’t be small, producing between 1.1 and 1.2 GW of electricity.

While that’s great news for Canada, it’s small potatoes for Westinghouse, which is now in a strategic partnership for “at least $80 billion in new reactors to be constructed across the U.S. using Westinghouse AP1000 and AP300 nuclear reactor technologies.” Those plants will produce around 1.2GW and 300MW of electricity respectively.

All those plants will need a lot of uranium, and the timing couldn’t be better for the start of work on an “advanced nuclear fuel fabrication facility in Oak Ridge, Tennessee, which will be the first plant of its kind in the USA.” It will make TRISO fuel specifically for SMRs.

$2 billion will be invested by France-based Newclear in a project with US-based Oklo to “develop advanced fuel fabrication and manufacturing infrastructure in the USA, with Sweden’s Blykalla also considering investing in the project.”

Lots of great news, but I think it’s important to point out the in terms of construction costs for nuclear power plants, the U.S. is very much behind China. The hope is that new rules and investment will change this, but we aren’t there yet and have a long ways to go. China is moving full steam ahead on nuclear power, with Unit 2 of its Zhangzhou nuclear power plant just having entered commercial operation. It also just started construction on two new power plants, with the concrete being poured “for the nuclear islands of unit 1 of the Bailong nuclear power plant in China’s Guangxi Zhuang Autonomous Region and of unit 2 at the Lufeng plant in Guangdong province.” Honestly, good for China. Nuclear is the way to go. I just hope we can catch our rate of construction up.

On the bright side though, private American company Valar Atomics just became the first fully private company in history to split the atom. Their Project Nova went critical (that’s a good thing for nuclear power plants) for the first time on November 17^th, 2025. They also just finished raising a $130 million Series A. Valar weren’t the only ones raising capital either, with X-energy (also building SMRs) raising a $700 million Series D.

I think that AI is going to help meaningfully bring down the cost of new power plant construction too. Aalo Atomics recently announced a “collaboration with Microsoft on using AI to accelerate permitting.” The idea is to greatly speed things up and reduce costs for the company, and maybe eventually even cut some bureaucracy (haha).

Oklo and Meta (the parent company of Facebook) have “signed an agreement to advance the development of a nuclear energy campus in Pike County, Ohio, designed to scale up to 1.2 gigawatts. The agreement includes a binding prepayment to support fuel procurement, enabling Oklo to advance early project work and secure fuel—adding new, clean, reliable power to the grid without shifting infrastructure costs to existing residents or businesses.” Meta is desperately trying to catch up in the AI race, and to do that it needs power, and a lot of it. This is a big step in that direction.

Nuclear powered shipping is really the way to go. India is looking to develop 55 MW and 200 MW SMRs specifically for powering large cargo ships. India isn’t the only one interested in this either, with a Norwegian group pointing out that nuclear propulsion is the “long-term solution for maritime decarbonization.” Silly reason imo, but I’ll take it no matter the reasoning behind it.

Fusion Energy

A lot of big news for TAE, starting with an announcement that their “current research reactor, Norm, has produced record performance in plasma temperature and stability – and we’re shortening our commercial roadmap as a result. After Norm, TAE will skip a planned sixth machine, Copernicus, and will move directly into the development of its first of a kind fusion power plant, Da Vinci.” The plan is to deliver commercial fusion power by the early 2030s (maybe 2031 if all goes well). I saved the big news till last:

Trump Media & Technology Group and TAE have announced a $6 billion all stock merger, creating one of the world’s first publicly traded fusion companies, with shareholders of each company owning approximately 50% of the combined entity. The merged company plans to begin constructing the world’s first utility scale fusion power plant (50 MWe) this year, with future plants scaled to 350-500 MWe. Trump Media CEO Devin Nunes and TAE CEO Michl Binderbauer will serve as co-CEOs, with Trump Media providing up to $200 million in cash to TAE at signing. Site selection is already underway. With this merger and added cash TAE just jumped up in my rankings for who’s going to be an early winner in the race to fusion energy.

If you’re curious who the biggest players in fusion energy are, TechCrunch has a great list for you of all the companies that have raised over $100 million. In order (now including the news for TAE) it’s TAE Technologies, Commonwealth Fusion Systems, Helion, Pacific Fusion, SHINE Technologies, General Fusion, Tokamak Energy, Zap Energy, Proxima Fusion, Kyoto Engineering (not building their own fusion power plant), Marvel Fusion, First Light Fusion (also not building their own power plant), and Xcimer.

UK Atomic Energy Authority researchers have developed GyroSwin—an AI tool that can simulate complex fusion plasma turbulence in seconds instead of the hours or days required by traditional supercomputers. The system learns plasma dynamics using AI, creating models that preserve crucial physical information like turbulence fluctuations and sheared flows while running up to 1,000 times faster than conventional simulations. “Designing, developing, and operating a fusion power plant will involve millions of plasma simulations,” said Rob Akers, UKAEA’s Director of Computing Programmes, noting that reducing runtimes “from hours or days to minutes or seconds” will be “genuinely transformative around time-to-solution and cost.”

Nor are they the only one, with NVIDIA and General Atomics delivering “the first AI-enabled digital twin to accelerate fusion energy research…this cuts plasma prediction time from weeks to seconds, accelerating the path to limitless clean energy.”

The States is making the right move here with a bipartisan announcement of a plan to extend advanced manufacturing tax credits to fusion companies under the existing framework of the (hilariously named) Inflation Reduction Act previously passed by the Biden Administration. The Department of Energy also announced “the creation of a new, dedicated ‘Office of Fusion (OF).’”

Ontario is investing $19.5 million CAD (~$14 million USD) alongside the federal government’s $33 million (~$24 million USD) and Stellarex’s $39 million (~28 million USD) to establish a Centre for Fusion Energy that will create a national platform for Canada. Ontario, through Ontario Power Generation, is home to almost all of the world’s commercial tritium—a by product of CANDU reactors—which serves as the key fusion fuel for most fusion reactor designs. The centre will advance fusion energy research and commercialization while facilitating academic and industry partnerships across Ontario.

Some more good news for the Canadian fusion industry: Kyoto Fusioneering and Canadian Nuclear Laboratories have begun construction of UNITY-2, the world’s first integrated tritium fuel cycle test facility at Chalk River, Ontario. The facility will continuously circulate up to 30 grams of tritium in 24-hour operational cycles, integrating all key fuel cycle technologies—from fuel injection and exhaust to impurity removal, isotope separation, and tritium storage—under fusion relevant conditions. Commissioning is scheduled for late 2026.

Japan is going to get a fusion power plant called FAST (Fusion by Advanced Superconducting Tokamak), with a plan to produce 50 MW of power sometime in the 2030s. Japan is moving along at a good pace on fusion, with Japanese company Helical Fusion signing a power purchase agreement (PPA) with Aoki Super Co, a large supermarket chain in Japan. It’s the first PPA for fusion energy in that country, and points to growing confidence in the ultimate success of fusion.

Zephyr Fusion, founded by two fusion physicists from Oak Ridge and Lawrence Livermore National Labs, is developing compact megawatt class fusion reactors designed specifically for orbit. Most satellites are limited to kilowatt scale solar power (about the power equivalent of a toaster or two), while scaling solar beyond 10 kW rapidly becomes mass and cost prohibitive—but compact megawatt scale orbital fusion power changes that. The founders say this could unlock an industrial space economy by providing the kind of power required for massive orbital data centers, manufacturing, and complex robotics that current satellites can’t support.

In the fission section I’d mentioned how we’re moving towards nuclear powered cargo ships (there are already a number of nuclear powered navy ships). Now, a fusion startup called Maritime Fusion wants to join in. “The first Maritime tokamak will be about eight meters across, and the startup is projecting it will be operational in 2032 and will cost around $1.1 billion.”

Fusion startup Thea Energy has unveiled Helios, their design for a stellarator power plant that uses 324 identical small magnets arranged like pixels on a screen, with AI powered software compensating for manufacturing imperfections that would cause problems in traditional fusion reactors. “It doesn’t have to be as good to begin with. We have a way to tune out imperfections on the back end,” said their CEO Brian Berzin. During virtual tests, the AI successfully corrected for magnets deliberately misaligned by over a centimeter without human intervention—a tolerance for sloppiness that could slash construction costs threatening fusion’s competitiveness against renewables. Helios will generate 1.1 gigawatts of heat converted into 390 megawatts of electricity with an 88% capacity factor matching nuclear plants, with electricity costs starting below $150 per megawatt-hour and dropping to $60 after several plants are built.

Commonwealth Fusion is making great progress on SPARC, their fusion power plant that should be turned on (and generate net energy!) sometime next year. They’ve now moved in the first of 18 magnets, each weighing 24 tons, and welded it into place. I have to say, it’s exciting watching this happen on the scale of months, versus ITER (also a tokamak) where it takes years to decades.

Researchers working on China’s EAST tokamak have achieved stable plasma operation at densities far beyond conventional limits by accessing a theorized “density free regime,” potentially overcoming one of fusion’s most persistent obstacles. The team demonstrated that plasma density, long constrained by empirical limits in tokamak operation, can be substantially increased without triggering disruptive instabilities by optimizing plasma-wall interactions during startup to reduce impurity accumulation and energy losses. The breakthrough relies on a theory called plasma-wall self-organization, which predicts that achieving a delicate balance between plasma and metallic walls dominated by physical sputtering can eliminate density restrictions entirely. For deuterium-tritium fusion, power scales with the square of fuel density, making higher densities importantfor achieving energy breakeven—but exceeding density limits traditionally causes instabilities that disrupt plasma confinement and endanger tokamak operation, a challenge that may now have a practical solution. In a nutshell, this means that tokamaks could pump out a lot more power than previously thought.

The IAEA (International Atomic Energy Agency) put out their World Fusion Outlook 2025 towards the end of last year. “The report, now in its 3rd edition, is the global reference on the latest in fusion research and development, fusion plants concepts and commercialization pathways — for both the private and public sectors.” It highlights “how fusion energy is entering a new phase of implementation as it becomes a cornerstone of national energy strategies and industrial planning.” You can read it here.

That’s it for now, but Nuclear Energy Updates will be back four weeks from now. In the meantime, check your inboxes next Saturday for the 35th edition of Techno-Optimist!

Thank you all for reading — and until next time, keep your eyes on the horizon.

-Owen

Subscribe now

Welcome to Robot Updates #2, news on the machines starting to reshape the future of work, industry, and everyday life.

I’m starting to experiment with sections here now too, so let me know what you think or if you’ve got any suggestions.

Alright, let’s dive in.

Subscribe now

First Law: A robot may not injure a human being or, through inaction, allow a human being to come to harm.

Second Law: A robot must obey orders given it by human beings except where such orders would conflict with the First Law.

Third Law: A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

—Isaac Asimov’s Three Laws of Robotics

New and Notable

• University of Pennsylvania and University of Michigan researchers have created the world’s smallest fully programmable, autonomous robots—microscopic machines measuring just 200 x 300 x 50 micrometers, smaller than a grain of salt, that cost only a penny each to produce. The robots can independently sense and respond to their surroundings, operate for months, sense local temperatures and adjust their paths accordingly, all while powered by light. The team developed a propulsion system that moves water using electrical fields to nudge ions in the surrounding liquid, solving the problem that for these small scales “pushing on water is like pushing through tar,” while creating computers that run on just 75 nanowatts of power—100,000 times less than something like a smartwatch requires. “This is really just the first chapter. We’ve shown that you can put a brain, a sensor and a motor into something almost too small to see, and have it survive and work for months. Once you have that foundation, you can layer on all kinds of intelligence and functionality. It opens the door to a whole new future for robotics at the microscale.” If anyone was wondering, this absolutely is huge news.

• Tesla is going to be rolling out Optimus version 3 sometime this year. Elon Musk said that “It will have essentially the manual dexterity of a human, a very complex hand and an AI mind that can navigate and comprehend reality. Optimus will be the biggest product ever.” The plan is to have a 1 million unit production line up and running by the end of the year. It sounds like construction has also started on a much larger 10 million unit per year facility down in Texas.

• Here come the clankers! It was probably inevitable, but it looks like humanoid robots are being designed for combat by a company called Foundation. Their Phantom MK-1 is “intended to act as a ground unit, serving as a first line of defense rather than offense. The robot is designed to perform tasks such as reconnaissance and bomb disposal, reducing the risk to human soldiers in high-risk missions.” For now, a human operator will make the decision whether or not to fire a weapon, keeping humans in the loop for “mission critical decisions.”

• This is cool, Japan’s Tsubame Industries has built a giant mech suit robot designed to be piloted by a human riding inside (or remotely). It could be used for anything from construction, to search and rescue, to (and I’m sure this will happen eventually) deployment on a battlefield.

• Chinese company EngineAI filmed an extremely impressive sequence with their T800 humanoid robot (yes, that appears to be a Terminator reference). The initial video was so good everyone initially thought it was AI enhanced at least, but this “behind the scenes” look shows that it was all very real. Definitely not CGI.

• UBTECH Robotics (another Chinese firm) has become the first company to complete a mass delivery of humanoid robots. The plan is to use their Walker S2 robots in factories and other industrial settings. All I can say is, do not underestimate Chinese robotics.

• Carrying on with the trend of Chinese excellence, robotics company AgiBot had their A2 humanoid trek 65 mile (106 km) without stopping, cross-province trek, setting a new Guinness World Record. It managed the feat by “hot swapping” it’s batteries, allowing it to keep going.

• Another interesting development is that Chinese researchers have built a “neuromorphic robotic electronic skin that allows humanoid robots to sense touch, detect injury, and respond to harmful contact with rapid, reflex-like movements inspired by the human nervous system.”

• This is a new one, five of Boston Dynamics Spot robots got into America’s Got Talent, and actually put on a pretty good performance. More humor than talent for now, but to all the modern dance artists who think your job is safe, the robots are coming!

• A lot less cute are “wolf” quadruped robots used by the Chinese military to simulate an attack on a Taiwanese beach. “The 70-kilogram robots, capable of carrying up to 20 kilograms of supplies, were seen breaching obstacles and clearing a path for follow-on forces.”

• Also, a Chinese robocop is already on patrol in Shenzen. I’m sure that while this is for show, such things will rapidly become both useful and common. Here’s a different one working as a traffic cop over in the city of Hangzhou.

• Lots of other crazy little demo’s for humanoid robots recently, including one from Unitree pulling a car, the XPENG IRON robot that looks like it’s got a (really skinny) human inside a suit but doesn’t, another robot (not sure of the company) shooting some hoops of basketball, a robot sewing, and one here playing badminton against a human.

Household Robots

• American company Sunday Robotics recently introduced their household robot, Memo, which is designed to do various household chores. Live testing will begin in 50 households this year.

• Another American company (are we maybe ahead of China on this?) called 1X released a 10 minute promotional video of their NEO household robot doing dishes, laundry, tidying up, and interacting with humans. To be honest, I don’t think any of these household robots have moved beyond novelties into the realm of truly useful yet—but they’re improving, fast.

• Not strictly a household robot, Munich-based Circus SE has launched the world’s first fully autonomous robotic kitchen in a Düsseldorf supermarket (Germany). The seven-square meter glass enclosed system handles everything—collecting ingredients from climate controlled silos, cooking, plating, and cleaning—without human intervention, producing up to 120 meals per hour or one meal every 30 seconds. Customers order restaurant quality dishes that are cooked live in minutes using dual robotic arms and real time AI sensors that adjust stirring speed and heat precisely based on meal type.

In the Factory

• Figure Robotics announced back in November that their F.02 robots “have contributed to the production of 30,000 cars at BMW.” You can see a video of one working here. Watching it, you may notice in the background that robots in factories aren’t exactly new, but humanoid ones are, and there are some jobs done better by a humanoid form.

• I agree with this article that the first place humanoid robots are going to make a real difference are in factories. Robots that can “pick up tools, hit torque specs, and work safely beside humans” are going to allow for an exponential expansion of what we can make. Household robots will come too, but in my opinion they’re still a few years away from being truly useful (by 2030 I expect that will have changed). But factory humanoids will start being useful this year, and likely make their presence felt in a big way starting in 2027.

• Morgan Stanley research now projects that the robotics industry could jump from $91 billion today all the way to $25 trillion by 2050. Now, they include autonomous vehicles in there with a $6 trillion value, but still. If they’re remotely correct, this will be one of the biggest growth industries of all time.

That’s all for this edition, but Robot Updates will be back four weeks from now.

Thank you all for reading—and until next time, keep your eyes on the horizon.

-Owen

Subscribe now

Welcome everyone to AI Updates #3, where I share the highlights of AI advances with you.

I hope you all had a great Christmas, and a chance to relax and eat some turkey (or whatever your holiday dish of choice is). As we approach the end of 2025, it’s quite something to look back on AI and see how fast it’s moved and how big the changes have been over the past 12 months. But as fast as things are moving now I fully expect them to move a whole lot faster in 2026.

So buckle up, and hang on tight as we accelerate up the curve!

Subscribe now

“AI is the new electricity.”

— Andrew Ng

Model Updates & News

Is AI a bubble? I’d say most definitely no. Not only is uptake faster than any other comparable technology, but it’s underpinned by physical infrastructure like power plants, data centers, and chip manufactories.

The length of time an AI agent can work on a coding task isn’t just increasing exponentially, the rate of acceleration is itself accelerating. Between 2019 and 2024 the task duration AI was capable of doubled every 7 months. Between 2024 and into 2025 it’s every 4 months. In case you were wondering, an increase of acceleration over time is called jerk (or sometimes jolt). So I guess we’re jolting forward? By the end of 2026 an AI will be able to keep working independently for half a week on coding tasks, unless of course they can go for even longer (if the jolt keeps increasing).

That’s not the only way it’s accelerating either. According to one index that measures AI abilities, these started to accelerate significantly in 2024—and are carrying on to the present day. Jumping from 8.1 points improvement per year, to 15.3 point per year. Almost doubling. And if I were a betting man I’d say 2026 will see that increase again, by a lot.

An AI called DeepWriter has achieved an all time high score of 50.09% on something called Humanity’s Last Exam. It won’t be too long at this rate before AI really is smarter than any individual human, at least when it comes to difficult purely technical questions.

MIT put out a study saying that by their estimate, AI can already replace 11.7% of the U.S. workforce. It hasn’t yet, but that’s probably mostly due to inertia, lack of knowledge employers might have of AI capabilities, and distrust of AI. On the flip side, the U.S. just posted a year over year GDP growth rate of 4.3% in Q3 2025, which I suspect was at least partly due to AI’s positive economic impacts. Elon Musk had an interesting comment on that, saying that “Double-digit growth is coming within 12 to 18 months. If applied intelligence is proxy for economic growth, which it should be, triple-digit is possible in ~5 years.” Now that is something I’d like to see. The man is often a bit aspirational on his timing, but he’s rarely wrong. Two bottlenecks we might run into though: the first is that it takes time to build out physical things like factories that will pump out robots or chips for AI. The second is energy. We already don’t have enough for proposed AI datacenters, let alone billions of robots. We’ll make it work, but it will be slower than many might like. Energy will be the limiting factor for everything big we want to do in the 21^st century, in my opinion anyways.

I haven’t tried this, but apparently on Gemini 3 Flash you can just create an app by talking now. Anyone who gives it a try, let me know how it turns out. “Vibe coding.” It’s a new model by the way.

Meta is back with something called SAM Audio: “the first unified AI model that allows you to isolate and edit sound from complex audio mixtures. This could mean isolating the guitar in a video of your band, filtering out traffic noises, or removing the sound of a dog barking in your podcast, all with text, visual, and time span prompts.”

OpenAI is trying to keep up on image generation, introducing a dedicated ChatGPT Images, which like the name says is just better at images. GPT 5.2 is also now available to everyone by the way. Smallish but noticeable improvement I’d say, which is nice (I use ChatGPT and Grok a lot).

Anthropic introduced Claude Opus 4.5, which they say is the best model for coding available (until next week anyways). Oh, and the company is also planning to go public next year.

Chinese company DeepSeek is back in the news as well, launching new reasoning models for agents.

Here’s something fun, AI researchers created an AI Village where 1,000 autonomous AI agents were dropped into a virtual world within the game Minecraft, and left to cooperate, compete, trade, build, and govern themselves, revealing surprisingly complex behaviours such as labour division, markets, leadership roles, and the creation of social norms. What started as a quirky experiment became a research tool for understanding multi-agent coordination, showing both creative emergent structures and real challenges — like agents looping on tasks or pursuing ends that don’t align with any particular purpose. The AI Village is a glimpse into how future AI fleets — managed like teams rather than single helpers — could transform everything from data analysis to software automation by amplifying what one person can achieve alone.

Are we going to see AGI in 2026? I’ll be honest, I do not know. xAI reportedly thinks their Grok model will exceed human intelligence next year. A company in Tokyo says it’s already there, with their AI learning “without pre-existing datasets or human intervention.” Another group (see photo above) also expects AGI to arrive by the end of 2026, probably across multiple models. I suspect there isn’t a hard barrier where everyone will suddenly agree that this or that model qualifies as AGI. It’ll probably be a gradual (though very rapid from our perspective) transition to more and more capable models till one day we all wake up and it’s indisputably a general intelligence.

AI Infrastructure

Let’s talk about datacenters in space.

In a nutshell,

To win the AI race you need the most compute, data, and energy. Everyone is working within the same constraints: there’s just one internet, just one GPU provider, same competition over data centers, energy suppliers, etc. Space changes all this. You can mass produce satellites and they all talk to each other in a swam via laser links. The major AI models are all within a few months of each other in terms of competency. There is no company even within a decade of in terms of launch capacity or cost. Going all-in on space-based data centers completely circumvents the bottlenecks everyone else faces. Overnight the competitive landscape has shifted from ‘neck and neck race’ to ‘10 year advantage to SpaceX and xAI.’ The game is already won. It’s already checkmate. Elon is all-in on SpaceX going public to raise capital to go all-in on space-based compute. Of course they will win at this point. They own the launch capacity, therefore they will own the compute capacity, therefore they will win the AI race.

I consider this to be a very good thing, by the way. At least with Grok the stated goal is to maximize truth seeking. We shall see, but I feel like this a much better outcome than some of the alternatives.

As proof of concept, it’s already happening. An NVIDIA backed company called Starcloud just trained an LLM in space. Launched aboard a SpaceX Falcon rocket of course.

In other news:

The AI infrastructure boom continues to accelerate, with US companies alone now committed to spending “a combined $569 billion on data center leases over the next several years.” That’s a $197 billion (53%) increase from what was announced in Q2 of this year. That’s not all, a new estimate says that AI datacenters just in the United States will use 106 GW of power by 2035. An upwards revision of 36% just since last April.

Some of that power may come from a rather surprising source: Boom Supersonic. What? Don’t they plan to make supersonic airplanes? Yes they do, but they need to fund their development and rollout, and as part of that development they have this incredible turbine that can be used to produce 42 MW of power for datacenters (minimum order of 10 units). Running on natural gas, it uses zero water, and the mini power plant fits snugly into a shipping container. They plan to start shipping in 2027.

An interesting development is that batteries may suddenly be finding their niche. When it comes to storing energy for renewables, the amount of batteries to continue running the grid for days or longer becomes astronomical. But to ensure the smooth and continuous operation of datacenters? Suddenly that looks a lot more doable. Power outages wouldn’t last long most cases, maybe a few hours at most, and that’s where batteries are perfect for plugging the gap.

Great article from Pirate Wires on how the datacenter water scare isn’t actually real at all. Yes, datacenters do use water, but it’s a tiny amount compared to most other uses, and not at all the monster it’s often portrayed as. That’s not the only one either, Google just put out a study as well showing that fears of AI “draining aquifers and boiling oceans” are off by about two orders of magnitude.

Alphabet (Google’s parent company) “has announced an acquisition of Intersect, a ‘data center and energy infrastructure’ company that aims to provide additional support for building out its capacity for AI.” It’s a cool $4.75 billion in cash + whatever debt Intersect had.

NVIDIA has struck a $20 billion deal with former rival chip manufacturer Groq, further cementing its dominance in chips for the AI industry.

Anthropic just signed a deal with Hut 8, a bitcoin miner and now datacenter provider. The company will develop ~2.3 GW of utility capacity which Anthropic will use to help run it’s models.

Microsoft is planning to spend $7.5 billion on datacenters in Canada over the next couple years, with the total now up to $19 billion when previous announcements for Canada are included.

AI Uses – science, engineering, etc.

The U.S. isn’t going to take China’s progress with AI laying down. A new announcement from the U.S. Department of Energy:

The U.S. Department of Energy (DOE) announced agreements with 24 organizations interested in collaborating to advance the Genesis Mission, a historic national effort that will use the power of artificial intelligence (AI) to accelerate discovery science, strengthen national security, and drive energy innovation. The announcement builds on President Trump’s Executive Order Removing Barriers to American Leadership In Artificial Intelligence and advances his America’s AI Action Plan released earlier this year—a directive to remove barriers to innovation, reduce dependence on foreign adversaries, and unleash the full strength of America’s scientific enterprise.

Scientists are starting to see more benefits from AI, with a significant increase in papers published among those using AI. This productivity boost was greatest in the social sciences and humanities, but biology also saw an over 50% increase. Math and physics had the smallest bump, but 36.2% isn’t anything to sneeze at. What this doesn’t answer is questions of paper quality. Is this saving time and getting good research out there faster, or is it pumping out the scientific equivalent of AI slop?

Science as it’s done now is too slow according to Edison Scientific, a company integrating AI Scientists throughout the entire research pipeline — from basic discovery to clinical trials. The goal is to use autonomous AI systems to accelerate real world scientific progress with the aspirational goal of finding cures for all diseases by mid century. Edison just raised a $70 M seed round to scale this vision and is building tools that aim to compress months of research work into hours.

xAI has announced Grok for Education, “a partnership with El Salvador…to bring personalized Grok tutoring to every public-school student in the country — over 1 million children. The world’s first nationwide AI tutor program.” Okay, now that’s pretty cool. I’m thinking it’s going to significantly boost the quality of education in that country, and serve as a template for other places.

Things are moving fast (slow compared to the speed even of 2026) in medicine with AI. An AI designed antibiotic is now about to be tested in a clinical trial, just a year after it was created.

This is pretty neat, a deep learning model can now predict how fruit flies form, one cell at a time. They plan to eventually do this with larger organisms too (e.g., mice). This is something AI is well suited for, as it’s simply to complex, with too many individual parts for our brains to handle.

I’m hearing more and more stories of how AI is saving people’s lives by correctly diagnosing life threatening problems. It’s basically negligence already for doctors not to be using AI at least as an aid in diagnosis and treatment planning. In 5 years, I hope it’s criminal negligence not to.

Did I mention that AI seems particularly good at cracking unsolved math problems? Here’s one. When do we (AI) crack all currently unsolved math problems? Soon, I’d say. Maybe by 2030, certainly by 2035. Not some, not just the “easiest.” All. Simultaneously, entirely new mathematical understandings could be gained, maybe entirely new branches of math that let us better understand and describe the fabric of reality.

Before I sign off, here’s a few fun and maybe useful prompts I came across: one for ChatGPT that lets you turn things into interesting pencil sketches. Here’s another from Gemini (Nano banana) for hiding words in images.

That’s it for this edition; AI Updates will be back in your inboxes four weeks from now.

Thank you all for reading — and until next time, keep your eyes on the horizon.

-Owen

Subscribe now

Welcome to the thirty-fourth edition of Techno-Optimist, and a very Merry Christmas to everyone! Wherever you are, I hope the holidays are treating you well, with lots of food and time with friends and family.

Just a quick personal update: I’ve actually been experiencing some health issues (heart this time, born with some problems). It’s been a couple rough weeks, but thankfully I seem to be improving a bit. Anyways, that’s why the last Robot and Nuclear Energy updates didn’t make it out, and why this newsletter is coming out today instead of last Saturday. The plan going forward is to publish AI, Robots, and Nuclear Energy updates over the next three Saturdays as per usual, but bear with me if they’re late or cancelled.

Speaking of splitting out those topics, thank you to everyone who voted last time. The results were 82% in favor of doing them separately, which I agree with. That way you’re not waiting a month between newsletters, and the main one isn’t quite so long.

Before we start, I’d just like to say how thankful I am for so much, and how much I’ve been blessed with (despite the challenges). Family, close friends, a warm house, food on the table, many other good material things, and most especially the greatest gift of all—Jesus. As the carol goes,

“Hail the heaven-born Prince of Peace!

…

Mild He lays His glory by,

Born that man no more may die,
Born to raise the sons of earth,
Born to give them second birth.

Hark! the herald angels sing,
Glory to the newborn King”

Merry Christmas everyone!

Ho ho ho, let’s dive in.

Subscribe now

“When something is important enough, you do it even if the odds are not in your favor.”
– Elon Musk

A New NASA Administrator. After many twists and turns, Jared Isaacman is finally, officially, NASA’s new Administrator, being sworn in on December 18^th. His own words are well worth reading:

As I step into this role, I make these personal commitments:

– Mission: I will champion the bold objectives of human space exploration, scientific discovery, and a thriving orbital economy that ensures America’s leadership in space. We will never again give up our capabilities to reach for the stars, and we will never settle for second place.

– Integrity: I will serve responsibly, transparently, and without personal gain, covering every cost I am legally permitted to, and fully adhering to my ethics agreement. My loyalty is to my country, my President, and the space agency that has inspired me since I was a child.

– Urgency: I will intensely focus the agency on achieving the near-impossible, the very reason NASA was established in the first place. We will eliminate the bureaucracy that impedes progress and empower the best and brightest to take ownership, move quickly, accept smart risks, and act with a relentless focus on mission success.

– Inspiration: Every launch, every scientific breakthrough must inspire the next generation to dream bigger, to reach higher, and believe that anything is possible. In addition to my existing philanthropic efforts, I will donate my salary as Administrator to the U.S. Space & Rocket Center’s Space Camp to help prepare the pioneers of tomorrow.

I am humbled by this opportunity, proud to serve, and ready to work alongside the most talented minds in America as we continue the greatest adventure in human history. Sincerely,

Jared Isaacman

NASA Administrator

The same day, President Trump signed a sweeping executive order to ensure “American space superiority.” Among the ambitious goals are a directive to return Americans to the Moon by 2028, to begin construction of a permanent lunar outpost by 2030, and to have a nuclear reactor as a power source on the Moon that same year. The order signals a sustained U.S. push to lead in both human spaceflight and space technologies as competition with other nations (China) heats up.

Our understanding of the universe is incomplete. We’ve known for a while that there’s some things that don’t add up with our current picture of how the universe works. But those cracks have been getting wider lately, particularly due to something called the “Hubble tension.”

The Hubble tension is a disagreement about how fast the universe is expanding. When astronomers measure the expansion rate (the Hubble constant) using the early universe, mainly the afterglow of the Big Bang called the cosmic microwave background radiation, they get a slower value. But when they measure it using the nearby, present day universe—by observing distances to stars and galaxies—they get a noticeably faster one.

The problem is that both methods are extremely precise and internally consistent, yet they don’t agree, even after accounting for any possible measurement errors. That suggests either there’s an unknown systematic mistake hiding in one (or both) techniques, or our current model of the universe is missing some undiscovered physics—such as a previously unknown form of energy, changes in how gravity works, or new particles influencing cosmic expansion.

You all know I love Kurzgesagt videos, here’s an excellent and timely one on this exact topic.

One good way to think about this problem is by considering some previous crises in astronomy: the orbits of Uranus and of Mercury. When the orbit of Uranus was precisely calculated, scientists at the time discovered it wasn’t completely following the laws of gravity. They didn’t toss those laws out, they theorized that there was a “dark planet” tugging gravitationally on Uranus that explained the discrepancy. Shortly afterwards, Neptune was discovered, exactly where it was predicted to be. Later, the orbit of Mercury also didn’t make sense, so astronomers initially went down the same path, proposing another planet closer to the Sun—which they actually named Vulcan. Except there is no planet there, so an entirely new approach was needed. The result was Einstein’s theory of relativity, an entirely different framework than Newtonian gravitational dynamics. It worked, and because gravity bends spacetime Mercury’s orbit then made sense under our newly expanded understanding of how the universe works.

So the question now is, are we in a Uranus moment where there’s some undiscovered “stuff” we need to find, or is this a Mercury moment where we need an entire new set of theories to understand it?

Mars updates. Though it only has two tiny moons today, early in its history Mars may have had a much larger one. Thin, repeating sedimentary layers in Gale Crater on Mars—imaged by NASA’s Curiosity rover and interpreted as tidal rhythmites, “sediments deposited by the regular back-and-forth sloshing of the tides.” This suggests a moon large enough to raise measurable tides in the ancient crater lake around 3.8 billion years ago. By comparing the thickness and pattern of these alternating light and dark bands to tidal deposits on Earth, researchers estimate the tidal forcing would require a body at least 18× the mass of Phobos orbiting near Mars. While alternative explanations can’t yet be ruled out yet, the tidal interpretation aligns with previous models in which an ancient large moon break up under Mars’s gravity, creating rings and new moons over time, leading to the two Mars currently has. If upheld with further stratigraphic and modeling work, this would add a new and very interesting chapter to Mars’s early history.

Mars is often interpreted as being a rather chilly place in its past, even when it did have liquid water on its surface. But new analyses of rocks in Jezero Crater suggest that at least parts of ancient Mars weren’t just wet but tropical, with rainfall and weathering processes similar to Earth’s tropics driving the formation of bleached, clay rich sediments found by NASA’s Perseverance rover. These aluminum rich kaolinite clays form on and near Earth’s surface under prolonged wet and warm conditions, implying lots of rainfall and a substantial atmosphere far back in Mars’s past. Combined with other evidence of extensive river networks and ancient lakes, the picture of early Mars is shifting from a cold, arid world to one with significant warm, wet intervals and complex climate dynamics.

Rockets, rockets, and more rockets.

Lots of rocket updates for you, I’ll start with SpaceX: as all of you know, the company is big on reusability. On November 28^th, they successfully landed the same booster for the 30^th time! Another milestone, the company has done 165 launches this year, with another couple scheduled they’ll likely get to 167. That’s an average of 1 every 2.2 days. This rather impressive cadence has helped set a new record: 31 rockets launched over the 30 days of November. Meaning more rockets went up than there were days in the month—something that’s never happened before. It wasn’t just SpaceX, other companies and countries (mostly China) were in there too, but SpaceX made that number a lot bigger.

The company has also received permission to develop a new Starship launch complex down in Cape Canaveral, Florida—construction has already begun. Looks like two new launch sites, which will bring the total in Florida to three. Coupled with this was approval to launch 1.5 Starships per week, a pretty big increase from what they’re currently doing, but necessary to meet their near term aspirations. Rumor has it that the plan is to eventually have 49 Starship launch sites situated around the world.

In even crazier news that is no longer a rumor, SpaceX plans to go public next year (here’s why). It’s expected to be in the second half of 2026, and raise at least $30 billion at a $1.5 trillion valuation—meaning that just a few % of the company’s value will actually be on the market for trading. I’ll talk about this in the next AI updates newsletter too, but in a nutshell Elon wants to use Starlink satellites to build massive, distributed datacenters in space.

If you have followed Elon’s tactics, you know that once he commits to something, he leans fully into it. Much of the AI race comes down to amassing and deploying assets that work quicker than your competition. A large war chest resulting from an IPO will greatly help his cause and disadvantage all others.

Foremost among Musk’s goals right now is to “win” the battle for artificial intelligence. He is already attacking the problem at xAI and Tesla, and he now seeks to throw SpaceX into the fray as well. Taking SpaceX public and using it to marshal an incredible amount of resources shows he is playing to win.

My initial concern was “what about Mars?” What if shareholders don’t like that idea, dealing with activist shareholders, etc. But apparently “The rules of a public, profit-making company can be set up so that the board of directors must seek to maximize Mars colonization instead of maximizing profit.” Mars is still a go.

SpaceX isn’t the only company having a big year. Rocket Lab—probably the main runner up to SpaceX other than Blue Origin—had a stellar year of 21 successful launches and 100% mission success.

After Blue Origin nailed it with their epic New Glenn landing, becoming the second company ever to land an orbital class rocket, they aren’t slowing down. The plan is an even bigger version of New Glenn (Newer Glenn?) which will see several launches next year. It’s a massive rocket, and will be the 2nd biggest one flying, after Starship. More info from the company, and further thoughts here.

China meanwhile is trying their level best to catch up, but with a failed landing earlier in December, and another one yesterday they remain a solid 10 years behind SpaceX. But I don’t think they’ll be 11 years behind. Like SpaceX and others, these aren’t failures, but rather learning experiences. If I were a betting man, I would say that Chinese companies will start seeing success and landing boosters next year, probably in the first half. While 10 years behind sounds like a lot (and it is), with the exception of a couple other American companies, China is the only one seriously trying to match SpaceX.

Anti-aging, and progress against dementia. As usual, there’s a lot happening here. I’m more and more convinced that by 2030 (but possibly as soon as next year) we’ll start seeing anti-aging cocktails or combinations of drugs and treatments being tested. To start, there’s a type of T-cell in our immune systems that’s been found to seek and destroy senescent or “zombie cells” that are thought to be a factor in aging. Supercentenarians (people over a century old) have large numbers of them.

Immorta Bio has filed a patent for SenoVax, a proposed senolytic vaccine designed to train the immune system to recognize and eliminate those senescent (zombie) cells that accumulate with age and drive inflammation, tissue decline, and cancer progression. Their approach relies on immune targeting of senescence specific markers, potentially offering a more durable and systemic cleanup mechanism. Preclinical work cited in the patent suggests broad senescent cell reduction and anti-tumor effects.

A new study in Cell adds to growing evidence that shingles vaccination may slow dementia progression, not just reduce risk. Researchers found vaccination was associated with less cognitive impairment in healthy people, lower conversion from Mild Cognitive Impairment to dementia, and reduced mortality in those already diagnosed—across Alzheimer’s, vascular dementia, and most strongly, mixed dementia. The effect has now been replicated multiple times, with randomized trials underway. Possible mechanisms include reduced viral invasion of the brain, lower stroke risk, broader antiviral immune priming, and prevention of the pain, sleep loss, and inactivity that shingles often causes—all of which are linked to cognitive decline. Serious message here, everyone tell your parents and grandparents to get this. Get it yourself if you’re 50+.

Scientists have developed “nanoflowers”—tiny peptide based structures—that can rejuvenate human cells mitochondria by boosting mitophagy, the process cells use to clear out worn out energy factories and replace them with healthy ones. In cell and animal experiments, these nanoflowers restored mitochondrial function in aged or stressed cells, improving energy production and reducing markers of cellular damage. Because mitochondrial decline appears to underly aging and many chronic diseases, this approach points toward a programmable way to reset cellular powerhouses and potentially delay aspects of aging or disease progression. If translated safely to humans, it could become a foundational tool in next generation regenerative and longevity medicine.

Researchers at Kindai University (Japan) report that the common amino acid arginine can reduce Alzheimer’s-like brain damage in animal models by inhibiting toxic amyloid-β aggregation while also dampening neuroinflammation. In flies and mice, oral arginine improved cognitive performance and slowed disease progression, suggesting benefits beyond plaque reduction alone. Because arginine is already widely used in humans with a strong safety profile, the findings point to a potentially low cost, fast intervention if the effects hold up in clinical trials.

Lastly, At 90 years old, physicist John G. Cramer volunteered to become the first human recipient of bioreactor grown mitochondrial transplantation—an experimental therapy developed by Mitrix Bio that aims to restore cellular energy production by giving people fresh, lab grown mitochondria. Mitochondria are the powerhouses of cells, and their decline with age seems to be a core contributor to systemic aging, so replenishing them could tilt the odds toward healthier, longer lives if it proves safe and effective. Backed by researchers from Stanford, UCLA, Northwell Health, and Mitrix, the initiative actually started last August (waiting to hear any updates).

Space

• Researchers at the University of Tokyo report evidence that could represent the first direct observational signature of dark matter, detected not through light but via gamma rays observed by NASA’s Fermi Space Telescope. They identified a diffuse, halo-like gamma ray signal around the Milky Way whose energy spectrum matches predictions for annihilating dark matter particles, rather than known astrophysical sources. If the signal is confirmed by independent analyses and seen in other regions thought to be rich in dark matter, it would be a major step toward identifying dark matter as a new fundamental particle. Nearly a century after dark matter was first proposed, we may finally be starting to see it—still indirectly—outside of gravity. (University of Tokyo)

• Scientists analyzing samples from asteroid Bennu have discovered ribose and glucose—marking the first time glucose has been found in an extraterrestrial sample. Researchers also discovered a mysterious nitrogen and oxygen rich “space gum,” a polymer-like substance never seen before in astromaterials that was once soft and flexible but has hardened over billions of years. A third study found Bennu contains six times more supernova dust than any previously studied astromaterial, suggesting its parent asteroid formed in an early region of our own Solar System that enriched with stellar debris, contrary to the previous assumption that pre-solar dust was fairly well mixed. As a sidenote, I think that when we find alien life (most likely microbes), it may not be as alien as some might think. DNA based with water as a solvent is probably the norm. (NASA Solar System)

• The Terra Hunting Experiment just launched in the Canary Islands, marking the first systematic quest dedicated to finding Earthlike planets orbiting Sunlike stars. The project will observe up to 50 carefully selected sunlike stars every night for 10 years using the HARPS3 spectrograph, which can measure stellar motion at roughly 10 centimeters per second—about the speed of a baby crawling. The problem is that Earthlike planets have been nearly impossible to detect because their gravitational wobble is tiny and only shows up roughly once a year, meaning confirmation can take many years and be difficult as the signals gets drowned out by fluctuations in starlight. Extrapolating from planet discoveries made by NASA’s Kepler mission data, the team expects to discover at least two Earthlike planets in the habitable zone of Sunlike stars, which would then become prime targets for future telescope missions searching for signs of life, including NASA’s planned Habitable Exoplanet Observatory. I don’t know about you, but I am excited for this! (Science News)

• The James Webb Space Telescope has discovered what could be the universe’s first generation of stars—Population III stars that formed just 800 million years after the Big Bang in a distant cluster called LAP1-B, located about 13 billion light years from Earth. JWST only spotted them thanks to gravitational lensing, where galaxy cluster named MACS J0416 magnified LAP1-B’s light by about 100 times. Population III stars formed from only hydrogen and helium, reaching masses equivalent to 100 times the sun or more, and when they exploded as supernovas they forged and scattered heavier elements throughout the cosmos. The stars spectra showed emission lines suggesting lots of high energy photons and very large masses, matching theoretical predictions. If confirmed, this marks the end of one of astronomy’s longest hunts. (Space.com)

• NASA’s James Webb Space Telescope has detected the strongest evidence yet of an atmosphere on a rocky exoplanet. Not somewhere you’d want to visit, TOI-561 b is thought to be a “lava world” where the surface is at least partly molten. But instead of just being scorched bare rock it appears to be wrapped in a thick, volatiles rich atmosphere above its global magma ocean. The inference was made based on infrared emission spectra that are cooler than expected for an airless surface. Webb’s measurements show heat distribution and spectral features that match models with a substantial atmosphere, challenging the view that rocky worlds close in to red dwarf stars can’t hold onto an atmosphere due to intense stellar radiation. Now some caveats: this planet is a super earth, roughly 1.4x Earth’s radius. Also, the very fact that TOI-561 b orbits its star in less than 11 hours and has a molten surface may be the key to it’s having an atmosphere—think of any volcanic eruption, lots of outgassing. Still incredibly interesting though. (NASA Webb Telescope)

Engineering, Machines, & Physics

• CERN’s Antimatter Factory has dramatically boosted the rate at which antihydrogen—the simplest atomic form of antimatter—can be produced and trapped by refining how positrons are cooled, enabling researchers to create over 15,000 antihydrogen atoms in just hours rather than days, an eightfold increase from previous methods. With more abundant antihydrogen now available on short timescales, physicists can probe fundamental symmetries and the matter–antimatter imbalance with new ease. Also, if things keep scaling up, eventually we’ll get to whole grams of antimatter, more than enough to use in zipping around the Solar System at incredible speeds. (CERN)

• Japanese researchers have taken a major step toward molecular assemblers by developing an AI-integrated control system that can autonomously manipulate single atoms at room temperature using scanning probe microscopy—something previously limited to temperatures close to absolute zero. The system guides the probe tip to nudge, place, and arrange atoms with high precision, and the AI component allowed the scanning probe to operate continuously for over 25 hours, performing precise atomic moves without human supervision. It’s an early but meaningful leap toward programmable, atom by atom fabrication. In other words, one possible path to a Star Trek style replicator. (ACS Publications via Dr Singularity)

• Supersonic flight over the United States is once again possible, thanks to the “House Transportation & Infrastructure Committee passing the Supersonic Aviation Modernization Act.” This lifts a longstanding ban on civilian supersonic flight over land, regardless of noise levels at the ground. In practice, this will allow Boom Supersonic, who despite the name has developed a “boomless” form of supersonic travel, to return faster than sound passenger jets to American skies again. (Boom Supersonic)

• NASA’s Jet Propulsion Laboratory has deployed a new airborne imaging spectrometer called AVIRIS-5 (Airborne Visible/Infrared Imaging Spectrometer-5) to map critical minerals like lithium from high altitude flights, detecting the unique spectral fingerprints reflected by different rocks to reveal their composition with twice the resolution of earlier sensors. This effort has already surveyed hundreds of thousands of square miles of the American West and identified lithium bearing clays in mine tailings. AVIRIS-5’s technology builds on a lineage of NASA spectrometers used on Mars, the Moon, and other planetary missions, underscoring how technology developed for space can be used in terrestrial applications. By enabling finer, large scale identification of critical minerals this sensor could accelerate resource discovery and help stabilize supply chains. (Phys.org)

• Speaking of critical minerals, A U.S. company has confirmed a major rare earth element (REE) and critical minerals discovery at its Silicon Ridge project beneath the Utah desert. They identified a high grade, clay hosted deposit containing 16 important elements—including lithium, gallium, germanium, rubidium, cesium, vanadium, and both light and heavy REEs critical for semiconductors, defense, and advanced manufacturing. Independent assays show the site has grades that compare favorably with China’s prolific deposits, and so far only about 11 % of the area has been tested, implying considerable upside. The geology also allows easier, cleaner extraction than traditional hard rock mining, and the project already has permits and processing infrastructure ready to move toward production. If commercialized at scale, this find could significantly strengthen domestic supply chains for critical tech metals and reduce U.S. dependence on foreign sources. (Interesting Engineering)

Medicine & Biotech

• Some good news for drugs against mosquito borne disease. A new malaria drug called ganaplacide cured 97.4% of patients in a phase 3 trial of ~1700 patients in 12 African countries prone to malaria. A different team of scientists have also developed a pill that prevents dengue fever, also transmitted by mosquitoes. It was very effective in what’s called a human challenge trial, but still needs testing in real world conditions. While this is all great, I still think the best option is just to gene drive the mosquito species that transmit disease into extinction. That will successful prevent 100% of any future disease. Just an fyi, if I ever get the opportunity, I’m going to try and do exactly that. (Samuel Hume – malaria) (Samuel Hume – dengue)

• Researchers at the Japan Advanced Institute of Science and Technology isolated several bacterial strains from the guts of two amphibians and one reptile, 9 of which showed some anticancer activity. One bacterium called Ewingella americana, “isolated from the intestines of Japanese tree frogs (Dryophytes japonicus), possesses remarkably potent anticancer activity.” In mice, it completely eliminated colorectal cancer in 100% of cases after a single intravenous dose—far surpassing all current treatments. Not only is this potentially very good news for humans suffering from this cancer, it opens a new area to search for bacteria with potent anticancer abilities. (Medical Xpress)

• Scientists have engineered an experimental RNA-based drug called TY1 that can boost the body’s own DNA repair machinery to heal tissue damaged by events like heart attacks, potentially reducing scarring and restoring function where current treatments mostly manage symptoms. TY1 is a synthetic version of a natural RNA signal carried in exosomes from heart progenitor cells and works by enhancing a gene (TREX1) that helps immune cells clear damaged DNA, thereby promoting recovery of injured tissue. Early lab studies show promise not just for cardiac repair but also for other conditions involving inflammatory or cellular damage, marking a new class of therapeutics that tackle the root DNA damage rather than just downstream effects. Next up are clinical trials to test whether this DNA repair focused approach can translate into human treatments for heart disease and other ailments. I’m hopeful, as going closer to the source of the damage is always a good idea. (New Atlas)

• MIT researchers have created brain implants smaller than a single cell that can be injected into the bloodstream and ride on immune cells into the brain without open-brain surgery—a big departure from traditional neural implants that require cutting into the skull. These sub-cellular wireless chips fuse with monocytes (a type of immune system cell) creating a hybrid implant able to cross the blood-brain barrier and autonomously implant at sites of inflammation, where they can deliver targeted electrical stimulation for potential treatment of neuroinflammation and related diseases. In mouse tests the hybrids localized to the brain, stimulated neurons on command via near-infrared light, and didn’t disrupt normal behavior or harm other organs, pointing toward a safer, minimally invasive path for brain implants. The team hopes to begin clinical trials within three years through an MIT spinoff, and the platform could eventually expand to target other regions and conditions beyond brain disorders. Nanites anyone? (Singularity Hub)

• Scientists have developed a protein control system that lets them precisely dial protein levels up or down in different tissues of a living animal over its entire life, replacing comparatively crude on/off genetic switches with something more like a “volume knob” for proteins. Researchers engineered worms so that feeding them specific plant hormones can tune protein abundance in muscles, neurons, intestines and even reproductive cells independently. This capability opens the door to experiments that were previously impossible—like teasing apart how subtle protein changes in one organ affect aging, disease, and whole body physiology—and lays a foundation for future tools to control biology and cure disease with exquisite precision. (Interesting Engineering)

• Two incredible pieces of news for treating multiple myeloma, which as some of you may remember is a cancer I take particular concern with. The first is a study showing that Daratumumab and Teclistamab, a monoclonal antibody and bispecific antibody treatments respectively, work incredibly well together. After 3 years the progression free survival rate (i.e., no relapse) was 83.4%, whereas it was only 29.7% for the standard treatment. As a treatment for patients who have had at least one relapse, this is far better than anything else available right now—as far as I know anyways. The second is an early stage trial where 18 patients were given an antibody called linvoseltamab (a different bispecific antibody). After a few treatment cycles, none of them had detectable disease. The study lead said that based on his experience, he “would predict that after having such a good response after such a short time, the disease most likely could stay away for many years. Could it never come back in some patients? I would say it’s possible.” (The New England Journal of Medicine via Samuel Hume) (Miller School of Medicine)

• Scientists have developed a lasting vaccine that prevents deadly allergic reactions like anaphylaxis by training the immune system to tolerate allergens instead of overreacting, using a nanoparticle platform that delivers tiny bits of the allergen along with immune-modulating factors. In preclinical models, a single immunization protected animals from severe allergic responses to foods like peanuts, rebalancing immune cells toward tolerance rather than inflammation. This approach sidesteps the risk of triggering dangerous reactions during months long desensitization regimens with one durable dose, dramatically improving safety and compliance. Hopefully it translates to humans soon. (Singularity Hub)

Agriculture

• CRISPR wheat that makes its own fertilizer has been developed at UC Davis by increasing the plant’s release of a natural compound called apigenin that stimulates soil bacteria to fix atmospheric nitrogen into a form the plant can use—a process normally associated with legumes. By encouraging bacteria to form biofilms around roots, the modified wheat can thrive with far less synthetic nitrogen fertilizer, potentially cutting growing costs and increasing yields in one of the world’s most widely grown cereals. (Science Daily - CRISPR)

• Another new discovery works well (in symbiosis one might say) with this one above. Researchers discovered that a tiny mutation in plant roots can dramatically rewire immunity receptors to enable nitrogen-fixing symbiosis with soil microbes, potentially letting crops fertilize themselves without synthetic inputs. Scientists at Aarhus University in Denmark showed that changing just two amino acids in a receptor protein triggers the plant to form partnerships with nitrogen fixing bacteria similar to what legumes do, opening a path toward cereals like wheat or rice self-fertilizing. This breakthrough could slash global fertilizer use, further lower costs, and transform agriculture by embedding biological nitrogen fixation into staple crops rather than relying on industrial chemistry. If bred or engineered into major food plants, such mutations could make farms more productive and affordable. (Science Daily – root mutation)

• MIT researchers have uncovered how a common fertilizer micronutrient—lanthanides, a group of rare earth elements—actually benefits plants at the molecular level, showing that trace doses applied to seeds can make crops more resilient to UV stress and enhance early growth. They found lanthanides can insert themselves into chlorophyll molecules, replacing the central magnesium atom and helping the pigment stabilize and perform better under environmental stress, strengthening photosynthesis in crops like corn, soy, barley, and chickpeas. This work clarifies long observed but poorly understood agricultural effects of these micronutrients and could help farmers optimize their use through targeted seed treatments rather than bulk soil application. By revealing how these elements move through plants and interact chemically, the study points to new ways of boosting resilience and yields. (MIT)

Weird & Wonderful

• Raccoons living in cities are showing signs of domestication. Not just behavioral, but actual physical changes similar to the ones shown by domestic dogs. Most notably, city dwelling raccoons have shorter snouts compared to their rural counterparts, a trait linked to domestication syndrome. Scientists analyzed nearly 20,000 images and found this consistent skull shape shift across urban regions, likely driven by selection for calmer, less aggressive individuals that thrive on easy food sources like human trash. The researchers propose that simply living in close proximity to humans and exploiting our waste may be unintentionally pushing raccoons onto a path toward tameness. While far from being pets, these morphological trends hint at how certain wild animals can adapt to human dominated environments. Sidenote, we really should deliberately finish domesticating them, I suspect they’d make great pets—the comport of certain individual trash pandas notwithstanding. (Phys.org) (Axios)

• Some good news folks: In the event that microscopic black holes exist, and you happened to be struck by one—say something the mass of a pretty big asteroid—you’d have a descent chance of surviving. The effects wouldn’t be pleasant though, in most cases similar to being shot by a .22 bullet. Apparently a smaller black hole would just pass right through you with no effect at all. Hey, these are important questions to ask! (Science Alert)

• My first thought on reading the words Human Washing Machine were “what the #$%&! is that?” Japan’s Mirai Human Washing Machine—a futuristic hygiene pod that washes, rinses, and dries a person automatically in about 15 minutes—has officially gone on sale after capturing massive interest at Expo 2025 in Osaka. Users lie back in the capsule where microbubbles, fine mist showers, calming visuals, and music deliver a hands-free full-body clean while sensors monitor vital signs for comfort and safety. It’s being sold in a very limited run (around 40–50 units) at roughly ¥60 million (~$385,000 USD) each, with early buyers including an Osaka hotel and major retailers planning demos. After watching the video, that actually doesn’t look so bad. I think I’d give it a try, just not in front of a crowd. You can see how something like this could help reshape personal care for the elderly and those with health challenges in particular. (Interesting Engineering)

• Scientists have identified a non-coding gene called CISTR-ACT that directly controls cell size by guiding regulatory proteins to influence growth related genes; marking the first time a long non-coding RNA (part of the non-coding genome aka “junk DNA”) has been shown to drive such a fundamental cellular property. Researchers found that tweaking its levels makes cells measurably larger or smaller. The discovery further challenges—overturns really—the old idea that only protein coding DNA determines core biological traits, and opens new avenues for understanding diseases like cancer and anemia tied to aberrant cell size. By revealing a programmable lever on cell growth hidden in what’s sometimes called genomic dark matter, it expands our grasp of gene regulation and potential for cellular engineering. (Phys.org)

Photos & Videos

A photograph from the ISS, curtesy of Don Pettit. I know we complain a lot about light pollution these days, but when viewed properly from above, it can be rather beautiful. (Don Pettit)

An interesting view of all Earth’s known buildings, compiled from data around 30x better than anything previously done. The grand total? 2.75 billion buildings. (Interesting Engineering)

The Hubble Space Telescope has captured visual evidence of the collision of two planetesimals, or large planet forming bodies, in the Formalhaut system 25 light years away. (Hubble) (ESA Hubble)

Have a watch of an epic drive the Perseverance Mars Rover made earlier this year. (NASA)

On December 6^th a Hawaiin volcano erupted, sending molten rock, ash, and lava high into the air. Check out a U.S. Geological Survey webcam getting absolutely wrecked. (Scientific American)

Recommendations & Reviews

With Christmas just a couple sleeps away, this time I’m going to recommend that you all go and have a wonderful get together and celebration with friends and family. Life is short, and those we love are precious, so whether or not you celebrate Christmas go and spend some good time with them.

That’s it for this edition. Techno-Optimist #35 will drop on January 17^th (or possibly the 18^th if it’s late). The plan is to have the shorter focused newsletters for AI, Robotics, and Nuclear Energy out on each of the next three Saturdays or Sundays.

Let me just leave you with this heartwarming video from Colossal Biosciences of a woolly mouse exploring his winter wonderland:

Thank you all again for reading, please let all your friends and family know about this newsletter too — and until next time, keep your eyes on the horizon.

-Owen

Subscribe now

Welcome everyone to AI Updates #2, where my mission remains distilling the firehose of AI news into something you can actually read before the next breakthrough drops.

If you look below to today’s quote, I actually disagree with Edsger in that I find the question very interesting indeed. Though I think the analogy he makes is a good one. Can a submarine swim? Arguably no, not if we look at what life does. But the end result is mostly the same. Ditto for artificial intelligence it would seem. Whether or not it’s actually thinking, the end result certainly looks increasingly like it is thinking — or something like it anyways.

Alright, hang on tight as we accelerate up the curve!

Subscribe now

“The question of whether a computer can think is no more interesting than the question of whether a submarine can swim.”
—Edsger W. Dijkstra

Model updates and news

Elon Musk tasked xAI to come up with a better version of Wikipedia — which is insanely biased on anything even remotely political or with a “culture war” feel. The result? Grokipedia. Even the 0.1 version was pretty decent, by all accounts being very neutral and even-handed on all topics. It’s now on version 0.2, letting anyone propose edits to topics. The plan (cannot get much more epic than this) is to rename Grokipedia to the Encyclopedia Galactica once it’s fully mature. According to Elon Musk, “This is intended to be a massive open source repository of all knowledge about the Universe. Many copies will be distributed periodically throughout the solar system to preserve knowledge for future civilizations should ours perish or subside into barbarism. We take the lesson of the Library of Alexandria to heart.” xAI is hiring people to work on Grokipedia, so join up and “Help build the galaxy mind version of the Library of Alexandria and scatter copies throughout space and other planets and moons!” This time, the knowledge won’t be lost.

Jeff Bezos has jumped into the AI ring, creating a new startup called Project Prometheus where he will be the Co-CEO. It’s got $6.2B in startup funding from Bezos, which is definitely buy-in level money. More info in this thread.

Grok 4.1 is out, and includes improvements to “Model personality and quality, low latency, reliability up and down the stack.” From what I’ve seen, it was a noticeable improvement.

Google has announced that Veo is “getting new precision editing capabilities that let you easily add or remove elements from a scene - all while preserving the integrity of your original video.” Think adding glasses to someone in a video, or a cat on one’s shoulder, or a wig.

Google’s Nano Banana Pro image creator is out with some improvements. Pretty cool if you ask me. Here’s a tutorial on how to use it. Sounds like it’s pretty good at diagrams and graphics.

Google has also launched Gemini 3, “the best model in the world for multimodal understanding, and our most powerful agentic + vibe coding model yet. Gemini 3 can bring any idea to life, quickly grasping context and intent so you can get what you need with less prompting.” Apparently it really is quite good.

OpenAI has launched a new browser, ChatGPT Atlas. Currently just available on macOS, it’s looking to compete with Google as a search engine. They’ve also upgraded ChatGPT to 5.1, which includes “improvements in instruction following, and adaptive thinking. The intelligence and style improvements are good too.”

Not sure this rise of AI-generated content is a net positive yet, given the amount of “AI slop” out there. Counterpoint: there was and continues to be plenty of human created slop out there, so join in at the trough I guess? Eat up, it’s not that bad. Another way of looking at it is that AI has suddenly become good enough to seriously compete with average content written by humans. Not bad for what’s essentially a thinking rock.

Sounds like Meta will start using some chips developed by Google to run it’s AI. NVIDIA isn’t thrilled, but says that’s it’s “a generation ahead of the industry.”

Anthropic has dealt with a “highly sophisticated AI-led espionage campaign,” aimed at “large tech companies, financial institutions, chemical manufacturing companies, and government agencies.” Apparently it was China (is anyone surprised?). Moral of the story is that the best defense against a bad guy with AI is a good guy with a bigger, better AI.

A new AI chip maker has entered the ring. Qualcomm is taking on the heavyweights of NVIDIA and AMD. Good on them, let’s see more competition!

Qualcomm isn’t the only new AI chip maker. Startup Extropic has released their alien looking chip, which is designed to use far less energy than conventional chips.

One more chip to tell you about. Tesla has a new one called AI5, that’s a big step up from their previous chips (did not realize they were even doing that until just now). Production will start next year.

AI Infrastructure

NVIDIA wants to build datacenters in space, announcing last month a partnership with a company called Starcloud to eventually build massive AI infrastructure off world. The idea is to build a 5GW datacenter in orbit, with “super-large solar and cooling panels approximately 4 kilometers in width and length.” There’s a good writeup on it here from Ars Technica, and another here by Peter Hague. This is already being tested, with NVIDIA having launched one of their H100 GPUs into orbit, where it will run an open source version of Gemini.

Google is jumping onto the datacenters in space bandwagon too, planning to launch their Project Suncatcher test of two satellites in early 2027.

Did I mention that SpaceX is getting in on this too? They say that the version 3 of their Starlink satellites could be networked into a massive, distributed orbital datacente.r

Meta is moving forward to build their Hyperion Data Center, which will cost roughly $27 billion and eventually have 5GW of compute. They’re partnering with a fund managed by Blue Owl Capital to get it going.

Power is going to be the limiting factor for the speed at which new AI datacenters are built. Spare capacity is shrinking fast across the United States. Not just the U.S. either, Alberta has awarded the last 1.2GW of it’s spare grid capacity for datacenters to two projects. The remaining 37 projects in the que with 19.4GW of demand will either need to wait—likely for years while new power plants are built—or build their own.

Google just dropped a cool $40B investment into Texas for AI and cloud infrastructure. This will all be spent in just the next 2 years.

Speaking of massive investments, Anthropic decided to top that by announcing a $50B investment to build datacenters in Texas and New York.

Brookfield isn’t skimping out either, with a $100B “global AI infrastructure program -backed by NVIDIA and the Kuwait Investment Authority - to build out the full AI value chain from power and land to data centers and compute.”

Amazon Web Services just brought online their Project Rainier, “one of the world’s largest AI compute clusters.” Microsoft just announced a new datacenter in Atlanta called Fairwater.

One last quick note. There’s a lot of noise on how much water datacenters use. Turns out, not that much.

AI uses – science, engineering, etc.

The U.S. Department of Energy has launched the Genesis Mission. A “Manhattan Project” to double American science and engineering productivity within a decade using artificial intelligence. The mission will integrate the nation’s 17 National Laboratories, connecting the world’s best supercomputers, AI systems, next-generation quantum computers, and advanced scientific instruments into what Energy Secretary Chris Wright calls “the world’s most complex and powerful scientific instrument ever built.” The platform will mobilize roughly 40,000 DOE scientists and engineers alongside private sector innovators to tackle three priorities: American energy dominance through accelerated advanced nuclear and fusion development, discovery science powered by quantum computing, and national security through AI-enhanced defense materials and nuclear stockpile maintenance. The initiative builds on Trump’s earlier directive to remove barriers to AI innovation and reduce dependence on foreign adversaries in the technology race. Now we’re cooking!

Interesting, apparently those with “higher intelligence generally show lower glucose metabolism during task performance, suggesting they require less energy to achieve the same or better results, and their neuronal computation is more efficient.” Could this translate to AI somehow? Maybe as it gets smarter it’s energy use per “thought” will improve.

We’re so early in terms of how AI will impact everything. A new AI model “trained to use computational tools might accelerate mathematical discovery.” How many unsolved mathematical problems will be solved in the next decade? My bet is a non-zero number. Take for example this Chinese model called AI-Newton, capable of automatously “discovering” known physics principles like Newton’s second law after “being fed experimental data.”

There’s a new AI Scientist called Kosmos, created by Edison Scientific. It’s estimated to do “6 months of work in a single day. One run can read 1,500 papers and write 42,000 lines of code. At least 79% of its findings are reproducible.”

That’s it for this edition; AI Updates will be back in your inboxes four weeks from now. Thank you all for reading — and until next time, keep your eyes on the horizon.

-Owen

Subscribe now

Let’s get to it.

Subscribe now

The decades long debate over whether Nanotyrannus was a separate species or just a teenage T. rex has finally been settled—it’s a distinct dinosaur. Analysis of the famous “Dueling Dinosaurs” fossil from Montana revealed the first complete adult Nanotyrannus lancensis skeleton, showing it was fully grown at around 20 years old when it died. The skeleton has more teeth, larger forelimbs, fewer tail vertebrae, and different skull nerve patterns than T. rex; features that form early in embryonic development and don’t change with age. Researchers also identified a second species, Nanotyrannus lethaeus, from another fossil previously thought to be a juvenile T. rex. Where T. rex was a massive, bone crushing predator weighing 18,000 pounds, Nanotyrannus was a lean, agile hunter tipping the scales at just 1,500 pounds—built for speed rather than raw power. (Ars Technica)

University of Chicago researchers have revealed that Edmontosaurus annectens—a 40 foot (12 meter) long duckbilled dinosaur—had hooves, a tall fleshy crest along its neck and trunk, and a row of spikes running down its tail. The preserved features aren’t actual tissue but rather a clay mask less than one-hundredth of an inch thick that formed on two sun dried dinosaur carcasses after flash floods buried them 66 million years ago. A biofilm on each carcass’s surface electrostatically pulled clay from wet sediment to create a thin template capturing the true surface in three dimensions before organic material decayed away. Amazing that we now know with such details what these animals looked like, and particularly fascinating that Edmontosaurus is now confirmed as having “the earliest hooves documented in a land vertebrate, the first confirmed hooved reptile.” (Phys.org)

Researchers have precisely dated fossils from New Mexico to between 66.4 and 66 million years ago—placing them in the final 380,000 years before the asteroid impact that ended the dinosaur’s reign. The fossil evidence shows that dinosaur populations across North America were divided into separate “bioprovinces” shaped primarily by regional temperature differences, rather than being one uniform population. These New Mexican dinosaurs, including the massive long necked Alamosaurus, were completely different species from the famous Hell Creek dinosaurs found in Montana and the Dakotas, yet lived at exactly the same time. The findings counter the idea that dinosaurs were already in gradual decline before the asteroid struck—instead, they were thriving in diverse, vibrant communities right up until the moment of impact. (Science Magazine)

In Dinosaur Provincial Park, in Alberta, Canada, paleontologists have discovered that two lichen species preferentially colonize exposed dinosaur bones, covering up to 50% of fossil specimens while appearing on less than 1% of surrounding rocks. The vivid orange lichens create distinctive visual signatures that drones can detect while flying overhead, potentially revolutionizing fossil hunting in remote badlands terrain. Dinosaur bones provide the alkaline and porous surfaces these lichens prefer, and paleontologists often notice a “meadow” of orange lichen before spotting the bones themselves. (BBC Wildlife)

Have an excellent Thanksgiving, and don’t forget to check your email this Saturday for the 2^nd AI Updates.

Thank you all for reading — and until next time, keep your eyes on the horizon.

-Owen

Subscribe now

Let’s get right to it.

Subscribe now

• Scientists have identified a kidney produced molecule called betaine that seems to drive many of the anti-aging benefits of regular exercise. During a sustained training program, betaine levels rise and help counter chronic inflammation. When taken on its own, betaine can mimic several effects of long term exercise, including improved immune function and better metabolic balance. The work helps explain why consistent training rejuvenates the body even though individual workouts create short term physiological stress. It also suggests supplements that could deliver some of exercise’s protective benefits for those unable (or just to busy) to stay active. (Science Daily)

• NYU researchers developed a protein based therapy that restores a specific type of immune cell (macrophages) in aged mice. As these cells come back, the mice’s skin blood flow improves, bruising heals faster, and the tissue regains more youthful repair capacity. The treatment seems to rejuvenate older macrophages in place, rather than recruiting fresh ones, suggesting a way to revitalize aged immune function without wholesale cell replacement. Because similar macrophage types exist in multiple organs, this approach could potentially protect or restore vascular integrity more broadly, beyond just the skin. (Singularity Hub)

• China has made dramatic human lifespan extension a national priority, with Xi Jinping caught on a hot mic not long ago discussing longevity research with Vladimir Putin. Shenzhen-based Lonvi Biosciences is developing anti-aging pills based on procyanidin C1, a grape seed extract compound that increased mouse lifespans by 9.4%, with their chief technology officer declaring “Living to 150 is definitely realistic.” Chinese scientists recently published research in Cell showing that engineered senescence resistant stem cells reversed biological aging in elderly primates by six to seven years in neural tissue. Not there yet, but they’re working hard in the right direction. (NYT)

• An international research team has discovered that boosting levels of NAD⁺—a natural metabolite that declines with age—restores memory in Alzheimer’s disease models by fixing RNA splicing errors that disrupt hundreds of genes crucial to neuron health. The researchers found that supplementation with NAD⁺ precursors “can offer therapeutic benefits in AD [Alzheimer’s Disease] animal models and early clinical trials]. The study, published in Science Advances, suggests that metabolic therapies and RNA splicing correction may be two arms of the same system—potentially opening a new therapeutic pathway for protecting the brains of those with Alzheimer’s. (Medical Xpress)

• Cedars-Sinai researchers created “young” immune cells from human stem cells that reversed cognitive decline and Alzheimer’s symptoms in mice, with treated animals showing better memory and healthier brain structures. The team used induced pluripotent stem cells—adult cells reprogrammed to an embryonic like state—to generate fresh mononuclear phagocytes, immune cells that normally clear harmful substances but decline with age. Mice receiving the young immune cells outperformed untreated mice on memory tests and maintained healthy numbers of mossy cells in the hippocampus, which typically decline with aging and Alzheimer’s. [Sidebar: Mossy cells are a type of excitatory neuron located in the dentate gyrus of the mammalian hippocampus, named for the “mossy” appearance of their dendrites. They are crucial for cognitive functions like learning and memory, and play a key role in regulating the activity of other neurons, including neural stem cells in the hippocampus]. The cells appeared to protect the brain indirectly, possibly by releasing anti-aging proteins or removing pro-aging factors from the bloodstream, since they didn’t cross into the brain itself. (Science Daily)

Hope you all enjoyed that little mini newsletter. There’s another one on some dinosaur discoveries and other prehistoric life coming your way Wednesday

Thank you all for reading — and until next time, keep your eyes on the horizon.

-Owen

Welcome to the thirty-third edition of Techno-Optimist, your go-to destination for all the latest updates and commentary on space, science, technology, medicine, and much more.

Given that I’ve published this today instead of yesterday, I think my plan to always publish Saturday should be taken as aspirational. It’s always the goal, but life sometimes has other plans. To make it up to you though, I’m going to send you two special mini-newsletters this week. The first on Tuesday will details some anti-aging progress being made. The second will have some interesting paleontological discoveries. I’ll send it out Wednesday afternoon, right in time for Thanksgiving. How are dinosaurs related to Thanksgiving you might ask? Well, I’m sure some of them probably tasted like chicken, which isn’t quite the same as turkey, but close enough.

As promised, here’s a poll on my changes to the newsletter where I’ve broken out a few key long item topics into their own smaller newsletters. Happy to discuss in the comments too.

Alright, we’ve got a lot to cover, so let’s dive in.

Subscribe now

“Progress, as we have seen, necessarily entails some level of risk in the short term to reduce overall risk in the long term.” —

New Glenn 2^nd launch. After years of painfully slow progress, Blue Origin is finally hitting the gas! Feast your eyes on the beautiful images of their second New Glenn launch. The huge rocket really is impressive, not Starship size, but an absolute monster compared to anything other than that and NASA’s SLS rocket.

The mission was a smashing success, deploying NASA’s twin ESCAPADE spacecraft—built by Rocket Lab— toward Mars and landing its massive reusable first stage booster on the ocean platform Jacklyn. The 17-story booster, powered by seven BE-4 engines, stuck the landing, and was then floated back to port. This successful reusable landing puts Blue Origin in rarified company alongside SpaceX, dramatically reducing launch costs and positioning New Glenn to compete for NASA’s lunar missions and commercial satellite deployments.

Blue Origin isn’t slowing down either, aiming for more than a dozen New Glenn flights in 2026. The next mission will actually be headed to the Moon, with the company planning to send its “Blue Moon Mark 1 lunar lander, a vehicle that is taller than the Apollo Lunar Module that carried humans to the Moon five decades ago.”

NASA’s ESCAPADE mission that launched aboard New Glenn will study how solar wind strips away Mars atmosphere, helping explain how the red planet transformed from a much wetter world into todays frozen desert.

Watch the entire mission here

Raising the rocket

Launch 1

Launch 2

Landing 1

Landing 2

Never Tell Me The Odds booster sitting on landing platform vessel Jacklyn

Booster sitting on landing platform vessel – humans for scale

SpaceX Moon plans. SpaceX has released renders on the lunar lander version of Starship that will take astronauts down to the Moon, and I must say they’re all pretty spectacular. This may be the fist spaceship in history that actually needs guardrails!

In terms of timeline updates for lunar missions, “SpaceX is reportedly internally targeting June 2027 for a Starship uncrewed Moon landing and September 2028 for a crewed Moon landing. These dates will only hold if SpaceX meets the June 2026 target for an in-space propellant transfer test between two Starships in orbit.”

Elsewhere, the company is picking up the pace even more, with a record two separate launches taking place in less than an hour from Florida on November 14^th. That, ladies and gentlemen, is the sort of launce cadence we need if we want to become a spacefaring, multiplanetary civilization.

Quantum advantage. Google’s Willow quantum chip has achieved the first ever verifiable quantum advantage—a quantum calculation that outperforms classical supercomputers with repeatable, confirmable results. The Quantum Echoes algorithm, published in Nature, runs 13,000 times faster on Willow than on Frontier, one of the world’s fastest supercomputers. The algorithm sends a controlled signal into the quantum system, perturbs one qubit, then reverses the signal to capture an amplified “echo” through constructive interference.

Just a couple days later, IBM announced a successful run of its quantum error correction algorithm on readily available AMD chips—a breakthrough that could dramatically reduce the cost of building practical quantum computers. Qubits are notoriously error prone, and managing these errors has been a major bottleneck in quantum computing development. IBM’s algorithm runs 10 times faster than necessary for effective error correction and works on affordable, commercially available hardware rather than requiring expensive custom built components, putting the company a year ahead of schedule on its roadmap to build their Starling quantum computer by 2029.

A company called Quantinuum just launches a quantum computer they dubbed Helios, which they say is the world’s most accurate commercial quantum computer, featuring 98 fully connected qubits with single qubit gate fidelity of 99.9975% and two qubit gate fidelity of 99.921%. The system integrates NVIDIA GPUs directly into its control engine for real time quantum error correction—improving logical fidelity by more than 3%. In benchmark tests using Random Circuit Sampling, Helios completed calculations that would require more power than the Sun produces for a classical computer to replicate.

[Sidebar: Qubit gate fidelity measures how accurately a quantum computer can perform operations without making mistakes.

Single qubit gate fidelity is how well the computer can manipulate one qubit at a time—like flipping a switch or adjusting a dial. Helios’s 99.9975% means it only messes up about 2-3 times out of every 10,000 operations.

Two qubit gate fidelity measures accuracy when working with two qubits together—this is where quantum computers create “entanglement” that gives them their special powers. These operations are trickier, so the error rate is slightly higher at 99.921%.

Why does this matter? Quantum calculations involve thousands of operations in sequence, so even tiny errors add up fast. Higher fidelity means the computer can run longer, more complex calculations before errors ruin the results.

Cancer treatment updates.

When it comes to cancer treatments, it seems like we’re seeing breakthrough after breakthrough, which suggesting we’re approaching a tipping point after which we might start to truly beat cancer into submission once and for all.

Northwestern researchers redesigned the chemotherapy drug 5-fluorouracil using spherical nucleic acids, creating nanoparticles that are 20,000 times more effective than the original while producing zero detectable side effects in mice. The redesigned drug enters leukemia cells 12.5 times more efficiently and reduced cancer progression 59x—completely eliminating tumors in animal models. Meanwhile, MIT scientists developed different nanoparticles that deliver the immune stimulating molecule IL-12 directly to ovarian tumors, achieving an 80% cure rate when combined with checkpoint inhibitors. Mice that were cured developed immune memory, clearing cancer cells when researchers reintroduced them months later—essentially vaccinating them against their own cancer.

Texas A&M researchers unlocked the cancer fighting potential of astatine-211, an isotope of the rarest naturally occurring element on Earth. It emits focused alpha particles that destroy cancer cells while sparing healthy tissue. Clinical trials using At-211 are already showing promise against blood cancers and brain tumors. In a completely different approach, researchers discovered that D-cysteine—the “mirror” form of the amino acid cysteine—selectively impairs tumor growth by entering cancer cells through transporters absent in healthy cells, then disrupting mitochondrial function. In mice with aggressive breast cancer, tumor growth slowed markedly without major side effects.

Medical University of Vienna researchers have developed LiPyDau, a chemotherapeutic agent that nearly stopped melanoma growth and eliminated aggressive breast cancer tumors with a single dose in mice. The drug uses liposomal nanoparticles to safely deliver a potent drug that irreversibly links DNA strands in cancer cells, causing unrepairable damage. LiPyDau even works against multi-drug resistant tumors and lung cancers that don’t respond to common treatments.

Space

• Scientists may finally understand why it “rains” on the Sun. The phenomenon is called coronal rain, where blobs of superheated plasma cool, condense, and fall along magnetic loops. Researchers at the University of Hawaii found that shifting abundances of elements like iron, silicon, and magnesium in the corona play a key role, influencing “radiative energy loss, where spikes in radiation cause the temperature to plummet at the peak of coronal loops compared to elsewhere in the Sun’s aura. This sucks more material up through the loop and triggers a runaway cooling effect, which results in coronal rain.” These insights could also improve our ability to model space weather and understand solar dynamics more deeply. (Science Alert)

• New research suggests that the universe’s expansion may be slowing down rather than continuously accelerating, challenging a longstanding idea in our understanding of the universe. The study argues that some key supernova data—once adjusted for the age of their host stars—no longer supports the Standard Model. Instead, the corrected data aligns better with dark energy weakening over time. If confirmed, this could mark a major paradigm shift in cosmology. (Phys.org)

• Astronomers have found three Earth sized exoplanets in a binary star system called TOI-2267, located about 190 light years away. Two of the planets orbit one star, while the third orbits its companion—marking the first known system where transiting planets circle both stars. Scientists say this provides a natural laboratory to test theories of how planets form and survive in extreme gravitational environments, as the two stars are only about 8 AU apart—roughly the distance separating the Earth and Saturn. Future observations could help determine more exact masses, densities, and even atmospheres if they have any. Cool aside, these planets would experience double sunsets, like Tatooine from Star Wars. (Space.com) (BBC Sky at Night Magazine)

• Astronomers have used ESA’s XMM-Newton observatory together with the LOFAR radio telescope to detect for the first time ever a coronal mass ejection (CME) from a star other than the Sun. The eruption came from a red dwarf about 130 light years away is estimated to have been relatively dense, and traveling at a blistering 2,400 km/s—plenty to strip the atmosphere from any closely orbiting planets. Because red dwarfs are the most stars to host planets in the “habitable” zone, so this raises serious questions about how stable their worlds’ atmospheres really are. The discovery opens a new window into space weather beyond our own solar system and could reshape how we assess exoplanet habitability. (ESA Science)

• Chinese astronauts aboard the Tiangong space station just held humanity’s first ever oven baked meal in space. They used an oven delivered by the Shenzhou-21 mission to cook marinated chicken wings and steaks in microgravity. The oven heats up to 190 °C (375 °F) and features a smoke purification system to keep the station safe. The wings reportedly took 28 minutes to roast, with the crew described them as “golden and crispy.” This milestone adds a new dimension to space life—the ability to enjoying real, cooked meals in orbit. “One wing for man, one giant bucket of wings for mankind?” (IFL Science)

• Commercial space company Gravitics has unveiled Diamondback, a compact orbital carrier built to hold and deploy space based missile interceptors in orbit, where they can be ready to respond quickly to any attack. Gravitics had previously won a Space Force contract worth up to $60 million to build and test the system. The idea is to stage assets in space ahead of time instead of launching from Earth. According to Gravitics, Diamondback could be used not just for defense, but also for rapid response civil missions—e.g., holding rescue gear or supplies for astronauts (Gravitics)

• Speaking of weapons in space, China has reportedly developed a satellite based weapon capable of delivering 2.6 megawatts of pulsed energy with microsecond synchronization accuracy. This system will power particle beam weapons in space, which could “disable or destroy enemy satellites and missiles through immense kinetic and thermal energy.” For anyone not paying attention, China is not slowing down. (Interesting Engineering)

• New research shows that a galaxy’s age strongly influences what kinds of rocky planets it can form. As stars live and die, they release different “metals” (to astronomers, anything heavier than lithium), gradually enriching their galaxy. Early on, massive stars produce elements such as oxygen, silicon, and magnesium, while smaller stars later contribute iron and nickel, among others. These changing chemical abundances affect planetary structure: older planets tend to have thicker mantles and smaller cores, making them less dense than younger ones. That means in the early universe rocky planets may have lacked the iron needed for strong magnetic fields–a key factor for potential habitability. As a galaxy ages, its new planets grow denser and more suited to support life as we know it. (Universe Today)

Engineering, Machines, & Physics

• eVTOL (electric vertical take-off and landing, i.e., flying car) company Jetson showcased a pretty amazing race with four of their vehicles last month. In the video, you can see them zipping around giant pylons, as well as executing some formation flying. Are these flying cars or giant drones capable of carrying people? Don’t know, but either way that was pretty cool. (JetsonAero)

• American company Shield AI has introduced the X-BAT, “The world’s first AI-piloted VTOL fighter jet. With vertical takeoff and landing, long range, and full autonomy.” The idea is to take “something like an F-35 or a comparable fifth-gen [fighter], puts it in a vertical takeoff and landing package, and then delivers it at a tenth of the cost, life-cycle[-wise] of a fifth-gen. So, breaking the cost curve. It enables us to counter China’s fifth and sixth-gen aircraft with something that comes at a fraction of the price.” Honestly, this thing looks straight out of a sci-fi, looking forward to seeing them fly their first real mission. (Shield AI)

• Anduril Industries has also developed an autonomous fighter jet, and this one has actually already taken off. The YFQ-44A took flight for the first time on October 31^st. What’s remarkable is that Anduril went from the design on paper to flying in just 556 days, far faster than any program I’ve heard of for new fighter jets. The drone is built to operate alongside manned fighter jets or independently, and Anduril plans to scale production rapidly. For any of us Lord of the Rings fans wondering about the company name, yes, it’s that Anduril. Flame of the West, the sword reforged. (Anduril Industries)

• A study suggests that designing a Mach 10 hypersonic jet could be much simpler than previously thought. The researchers demonstrated that turbulence at extreme speeds behaves more like slower airflow than expected, which relaxes some of the traditional design constraints. This insight could make hypersonic passenger or cargo aircraft more practical by reducing the complexity of their engines and structures. If the approach scales, we might get to fly anywhere in the world in an hour or so. (Interesting Engineering)

• MIT researchers have turned a simple molecule—radium monofluoride—into a tiny collider, letting electrons briefly plunge into the radium nucleus and bring back information about its internal structure. By measuring tiny shifts in electron energy levels, they showed electrons can directly probe the nucleus, offering a tabletop alternative to massive particle accelerators. The team now aims to cool and align the molecules to push the precision even further. (Science Daily)

• Engineers over in the Netherlands have designed a new algorithm that lets multiple autonomous drones work together to lift and precisely move heavy loads suspended on cables. The algorithm adapts in real time to changing payload dynamics and external disturbances, without needing sensors on the load itself. In lab tests with up to four drones, the system achieved significantly higher accelerations and tighter control than current methods. These capabilities could eventually be used in applications like construction, external building maintenance, and search and rescue operations. (Interesting Engineering)

• On a related note, DARPA has just announced the “DARPA Lift Challenge,” offering $6.5M in prizes to innovators across America who design drones that can carry payloads 4x their weight.” Registration opens in January, so if you or anyone you know might be interested, check it out. (DARPA)

• Engineers at Harvard have developed a metamaterial textile that can change its aerodynamics by forming dimples on its surface—just like a golf ball—as it stretches. When the fabric is pulled, it deforms into controlled indentations, with the dimples changing size and shape in real time. In wind tunnel tests, this adaptive surface cut drag by up to 20% depending on wind speed. The design could revolutionize highspeed sports apparel (e.g., cycling or skiing) and impact aerospace and maritime engineering. Imagine this for the Olympics. (Tech Xplore)

• An ultrathin 2D polymer film has been developed that’s almost completely impermeable to nitrogen gas or anything bigger. The material self assembles into a highly ordered structure, forming a barrier at least four orders of magnitude tighter than any existing polymer. Coated onto vehicles, aircraft and ships it could reduce wear and corrosion. It could also be used to extend the shelf life of food or medications. Because it’s lightweight, scalable, and easy to apply, the film could become a practical alternative to graphene, which is similarly impermeable, but notoriously difficult to use. (Interesting Engineering)

Medicine & Biotech

• Researchers in the States have broken the Guinness World Record for the fastest human whole-genome sequencing—completing a full genome in under four hours. They used a next generation “sequencing by expansion” (SBX) system paired with a streamlined workflow for both sequencing and analysis. The team demonstrated this speed with samples from newborns in a neonatal intensive care unit, showing that same day genetic diagnoses could soon be realistic for patients. This advance could dramatically improve decision making in medical settings, and start saving lives. (SciTechDaily via Dr Singularity)

• Researchers at Osaka Metropolitan University have developed a novel treatment using stem cells derived from body fat to heal spinal fractures mimicking those caused by osteoporosis. Because the stem cells are taken from fat—gathering it is a minimally invasive procedure even in older adults—the method may offer a safe, fairly gentle approach for treating bone degeneration in humans. The results suggest a promising step toward new therapies for spinal fractures and osteoporosis by harnessing the body’s own tissues to regenerate bone. Need human trials on this stat. (Science Daily)

• Researchers have created a new microrobot system that can navigate blood vessels, carry a large drug payload, and be tracked deep inside the body using 3D imaging. The tiny robots use magnetic nanoparticles inside a biodegradable shell, letting doctors steer them remotely and trigger drug release with changes in pH or radiofrequency heating. In tests, the robots formed stable swarms, pushed against realistic blood flow conditions, and moved through intricate, brain like vessel networks. It’s an early but promising step toward precision drug delivery in places that are normally very hard to reach. The idea is to use them to kill cancer cells with minimal side effects, ideally replacing traditional chemotherapy treatments. (Dr Singularity)

• Researchers have used a tiny wireless retinal implant to restore some central vision in older patients with advanced “dry” age related macular degeneration. The implant sits under the retina and works with smart glasses: a camera on the glasses captures scenes, converts them to infrared light, and sends signals to the chip, which stimulates remaining retinal cells so the brain can “see” again. About 80% of the 38 participants in the trial showed clinically meaningful improvement in their vision a year after implantation. The restored vision is still black and white and low resolution, and users needed months of training—but it’s a major step toward treating a disease once considered permanently blinding. And this is just the beginning, it’ll keep getting better from here. (Smithsonian Magazine)

• Stanford Medicine scientists have discovered a clear mechanism linking Epstein–Barr virus (EBV) to the autoimmune disease known as lupus. They found that EBV infects a rare subset of “autoreactive” B cells—the kind that can attack your own tissues—and reprograms them into an inflammatory state. These infected B cells crank out a viral protein called EBNA2, which flips on human genes that trigger immune overreaction. In lupus patients, the proportion of EBV-infected B cells is about 25 times higher than in healthy individuals, suggesting this mechanism could underlie all cases. This work opens paths for treatment and prevention, with an EBV vaccine or therapies that eliminate infected B cells being the logical paths forward. We don’t know yet, but perhaps there’s a link between other specific viruses and autoimmune diseases too. (Stanford Medicine)

• Researchers at the University of Nottingham have developed a new bioinspired gel that can actually regrow tooth enamel, something that was long thought impossible. The gel mimics natural enamel building proteins and draws calcium and phosphate from saliva to rebuild a mineral rich, mechanically strong layer on the tooth. It works even where enamel has completely worn away, repairing exposed dentine and reducing sensitivity. Best of all, “the regenerated enamel behaves just like healthy enamel.” Because it’s easy to apply, the gel could soon be a step change for preventing decay and literally restoring teeth. (Science Daily via @kimmonismus)

• German scientists have found that the essential amino acid leucine can directly boost cellular energy production by preserving key proteins in the mitochondria that usually get degraded. They showed that leucine reduces the activity of the protein SEL1L, which normally triggers degradation of mitochondrial membrane proteins, thereby enhancing the mitochondria’s efficiency. In short, what you eat matters not just for fuel but for how well your cells power plants actually run—and targeting this pathway might open new treatments for metabolic and energy related disorders. (Science Daily)

Agriculture

• Agricultural scientists in Maryland have discovered a tiny genetic switch that when activated early in flower development, causes wheat to grow three ovaries per flower instead of just one. Each ovary has the potential to become a grain, meaning this switch could significantly boost wheat yields without requiring more land. Because wheat is a staple for billions, leveraging this switch could make a huge difference in global food production. Breeders may now be able to use gene editing tools to turn this trait on permanently—offering a path to much higher yield wheat varieties. (SciTechDaily via Dr Singularity)

• Researchers have used gene editing to boost the production of α-amylase inhibitor proteins in common crops like corn, peas, and beans—which makes the starch in seeds indigestible to starch eating pests. This could significantly reduce damage from insects like weevils and beetles without adding additional pesticides. By tweaking the plant’s own genes (rather than inserting foreign ones), this approach may also sidestep regulatory hurdles and increase public acceptance, though sticking genes in from another organism is generally just fine too. The next very important step is making sure these modified plants don’t harm non target organisms like helpful insects, livestock, or humans. (Phys.org)

Weird & Wonderful

• A foundry in Paris called Atelier Missor says it’s “on its way to building a new Rome,” and they want to build it in America. They plan is to forge a gigantic titanium version of this statue of Atlas—called The Guardian of Liberty—and have it ready to unveil by the 250^th anniversary of America’s independence next July. The team at Atelier Missor says that the statue here is just a miniscule scale model, and that the real one will be “the tallest statue in the West.” I say bring it on! We need more builders to think on scales like this, and colossal statues to remind us all of what humans can accomplish. Will they do it in less than a year? I don’t know, but as Napoleon said, “Impossible is not French!” (Atelier Missor 1) (Atelier Missor 2)

• Turns out extinction rates aren’t nearly as bad as the media often portrays. University of Arizona researchers have discovered that extinction rates for plants, arthropods, and land vertebrates peaked about 100 years ago and have declined since then, disproving claims of an accelerating mass extinction. 912 species that went extinct over the past 500 years were analyzed, and the results found that most historical extinctions were caused by invasive species on isolated islands like Hawaii, while today’s main threat is habitat destruction on continents. The researchers discovered that “past extinctions are weak and unreliable predictors of the current risk that any given group of animals or plants is facing.” Declining extinction rates partly reflect conservation efforts actually working—investing money in species protection produces real results, and partly in growing global wealth—rich countries are able to care more about preserving their environment. (Phys.org)

• An Australian startup called Snowtunnel has created massive rotating barrels—12.5 meters (41 feet) in diameter and 16 meters (52.5 feet) long—that let skiers and snowboarders carve endless fresh snow without leaving the city. The cylinders spin like hamster wheels, with riders using their edges to maneuver side to side inside the tunnel. Each barrel features 3-4 inches of freshly groomed real snow and uses a proprietary surface treatment to keep the snow from falling as the tunnel rotates. The rotation speed adjusts from below walking pace for beginners up to 50 km/h (31 mph) for experienced riders. The first Snowtunnel Park opens in 2027 in Australia and will require far less land and upfront costs than traditional indoor ski slopes like China’s mammoth Shenzhen facility, which covers 11 soccer fields. (New Atlas)

Photos & Videos

This is next level amazing. The first photograph of its kind taken by Andrew McCarthy of a friend transiting the Sun while skydiving. You can see a short video of it here as well. The piece is titled The Fall of Icarus. Absolutely breathtaking. (Andrew McCarthy)

Amazing footage from inside the eye of Hurricane Melissa, shot by a guy whose handle is the apt ‘Tropical Cowboy of Danger.’ (Tropical Cowboy of Danger)

I had no idea this was a thing. The United States Mint has been releasing a few coins every year since 2018 to honor “American innovation and significant innovation and pioneering efforts of individuals or groups by issuing $1 coins with designs emblematic of innovation in each of the 50 states, the District of Columbia, and the five U.S. territories.” Here’s the four planned for 2026. Seriously great idea. (United States Mint)

Utility doesn’t always have to be ugly. Check out this giant deer whose antlers are holding up some power lines in Austria. We should be taking one of Europe’s few modern good ideas and running with it. (New Atlas)

Remora are fish that use special plates on their heads like a giant sucker to attach themselves to whales, sharks, turtles, and occasionally even boats and human divers. They’re helpful, cleaning off parasites and dead skin from their hosts. This video of some remora going for a ride on a Humpback whale is pretty cool, definitely not a calm trip! (Phys.org)

NASA’s JWST has taken a closer look at the Red Spider Nebula, showing its “legs” extend much farther than previously thought, about 3 light years. (ESA Webb Telescope)

Are rockets that aren’t reusable outdated and going the way of the dinosaurs? Absolutely, yes. But this is still a pretty remarkable view of ULA’s recent launch of their Atlas V rocket. (ULA)

Recommendations & Reviews

“First published in 1979, the Usborne Book of the Future is a fondly-remembered book from a time when people dreamed of the future as a place filled with wonder and amazing new technology. After more than 40 years of science fiction focussing on dystopias and doom, it’s time to remind readers young and old that, in fact, the Future is STILL a place that holds hope and excitement.”

I got this book for my kids, but honestly it’s been a great read for me as well, so I’m giving it a strong recommendation to all of you. This edition is a re-issue that’s faithfully reproduced all the original text and artwork, so don’t worry about it being “updated” to reflect modern sensibilities. Going through it, I was struck by how many of the ideas discussed then are still being looked into now, which makes me think that most of the good ideas really have been floating around for a while, just waiting for someone to bring them into existence.

One thing that disappoints me wasn’t the book, but how our progress into space, and to a more technological future generally—such was envisioned in this book—got derailed. The reasons are numerous I’m sure, but I don’t think it had to be this way. As incredible as our modern world can be, I think we’ve gone on a half century detour from which we’re only now starting to return. So enjoy this book if you get a chance, and let it inspire both you and your kids to the brighter future we can build if we set our minds, hands, and hearts to it. (Usborne)

That’s all for now. Techno-Optimist will be back in your inboxes on December 20^th (or possibly the 21^st if it’s late). I’m going to continue breaking out special shorter newsletters for AI, Robots, and Nuclear Energy (in that order) over the next three Saturdays, hope you enjoy.

Thank you all for reading — and until next time, keep your eyes on the horizon.

-Owen

Welcome to the first Nuclear Energy Updates, where we split what’s happening in the world of nuclear energy into bite sized pieces, and fuse it all into the bigger picture of the ongoing nuclear renaissance.

Apologies again on timing, I’m hoping to be more consistent getting these out on a Saturday. As I noted for the AI & Robot updates, this info is slightly dated; originally being scheduled to go out with the previous edition the main Techno-Optimist newsletter. But next month’s will contain all the up to date news.

I do feel that grouping all this news together in one place gives me, and therefore you my readers, a somewhat unique overview of what’s happening in nuclear energy. I call it the nuclear renaissance, but increasingly I think that name is more descriptor than something aspirational. The world really is accelerating forward with both fission and fusion, and it can’t come soon enough.

Alright, let’s dive in.

Subscribe now

“If a nuclear future appears frightening, the alternative is a return to the dark.”

—Sir Arthur Eddington

Nuclear Fission

China is not sitting on its hands when it comes to nuclear, including SMRs. They just completed a test of their ACP100 “small modular reactor demonstration project.” It’s being built in the island province of Hainan, and will have 125 MW (electric) capacity when completed.

Back in America, Oklo broke ground on their first Aurora SMR as part of the DOE’s new Reactor Pilot Program. The company plans to demonstrate the completed power plant next year.

Up in the Great White North (Canada), ARC Clear Technology and Nucleon Energy have launched a new partnership called NuARC, intending to deploy ARC’s “ advanced small modular reactor technology with an initial focus in Alberta, Canada.” Another company, X-energy, also wants to build their power plants in Alberta. They just completed a study which showed them to be a good fit for the province, as their reactors make both electricity, and lots of heat for use in industrial and potentially oil sands applications.

The U.S. Army is getting in on the action too, unveiling what they call the Janus Program, “a more than six year effort by the Army and the Pentagon’s Defense Innovation Unit with the goal of developing and constructing hundreds of state-of-the-art small nuclear reactors to power military bases across the United States.” All of these will be microreactors, generating “less than 20 megawatts of electricity,” but perfect for military bases—and they should be small enough to be transportable too. The first should be up and running in 2028.

Politicians in individual states are getting in on the action as well, with an announcement that “Governors from 11 states: New York, Indiana, Kentucky, Tennessee and Wyoming, Louisiana, Maryland, Pennsylvania, Utah, Virginia, and West Virginia, have agreed to collaborate in nuclear power development.” The idea is to “create a standardized and scalable approach that will lead to lower costs and build times,” creating an “order book” they can all use.

Maybe the coolest concept for a new nuclear company out there is Deep Fission, which intends to place their reactors a mile deep into the Earth for reasons of increased safety. I may have mentioned them before, but now they have a letter of intend from a customer who wants 12.5 GW of capacity. Those are some big numbers.

Fusion Energy

The DOE recently announced the Fusion Science and Technology Roadmap, designed to accelerate commercial fusion power. The plan is to “align public investment and private innovation to deliver commercial fusion power to the grid by the mid-2030s,” a timeline I think is very reasonable.

Commonwealth Fusion Systems released news that they’d signed a power purchase agreement with Eni, an Italian energy (but mostly oil and gas) company. The deal is worth over $1 billion, and shows increasing confidence in the burgeoning fusion industry.

Another big (though newer) fusion company had some new recently too. Pacific Fusion has said that they’ll build “their first Research and Manufacturing Campus” in New Mexico. The company plans to demonstrate net gain there by 2030.

China wants to not just demonstrate net gain by 2030, but actually be generating electricity from a compact tokamak machine called BEST (very on the nose name).

Germany is also boosting its funding for fusion energy, with plans to invest €2 billion ($2.3 billion) by 2029 in a variety of fusion infrastructure and pilot projects. A Germany company, Gauss Fusion, just published it’s conceptual design blueprint for a power plant they’ll call GIGA. They plan to have it up and running by the mid-2040s, which is great but may be a little behind the curve.

Japan’s newly elected prime minister Sanae Takaichi is a huge supporter of nuclear generally, and fusion specifically. She wants to “accelerate the development of fusion & make the country 100% energy self-sufficient by deploying next-generation reactors.”

Type One Energy has announced a plan along with the Tennessee Valley Authority (a massive utility company) to develop and build a stellarator fusion power plant. Not to be outdone, Helion has received a permit to move forward on the next phase of their Orion power plant, which they hope will be the world’s first (yes, they’re already building it).

That’s it for now, but Nuclear Energy Updates will be back four weeks from now. In the meantime, check your inboxes next Saturday for the 33^rd edition of Techno-Optimist!

Thank you all for reading — and until next time, keep your eyes on the horizon.

-Owen

Subscribe now

Welcome to Robot Updates #1, where I cover all the latest and greatest in the world of robotics as we head towards a Jetsons style future.

As with last week’s update on AI, let me know what you think of this one in terms of format and content. It’s also a little dated, containing only what I had ready for Techno-Optimist #32, but the next one will have all the latest and greatest.

Alright, let’s dive in (before the robots take over this newsletter).

Subscribe now

“Robotics is not just a field of science; it’s a canvas for art, innovation, and progress.”

–Rodney Brooks

An Australian company has built a construction robot they’re calling Charlotte, “billed as being autonomously capable of building a 2,150-sq-ft (200-sq-m) home in a single day – operating at roughly the speed of 100 bricklayers.” The “spiderbot” uses 3D printing to do its work, with applications everywhere from good old terra firma, to places like the Moon or Mars. I imagine several working together could also speed things up even more.

In what really will be a world first, an autonomous robot called Redwing has set off to circumnavigate the globe—underwater. Launched back on October 10^th, the plan if for Redwing to roughly follow the route taken by Ferdinand Magellan when he made his famous voyage around the world back in the 1500s. Like Magellan’s multiyear trip, the autonomous submersible will take at least 5 years to complete its journey at an average speed of 0.75 knots (1.4 km/hr, 0.86 miles/hr).

Figure Robotics has introduced Figure 03, the newest version of its humanoid robot. Not gonna lie, it’s a pretty impressive video, particularly as nothing the robot did was teleoperated. We’re not quite at the Jetsons yet, but it’s getting close.

Chinese robotics company Unitree has been showing off the skills of their humanoid Unitree G1, basically doing kung fu and flips. Apparently, they’re now shipping the robot through Walmart, all yours for the low low price of $21,600. Not entirely sure what you’d use it for (practice your kung fu?), but there you go.

That’s all for now, Robot Updates will be back four weeks from now. Let me know what you thought of this one!

Thank you all for reading—and until next time, keep your eyes on the horizon.

-Owen

As promised, here is AI Updates #1, where I try my level best to keep you informed about what’s going on in the world of AI as it surges forward at a dizzying pace.

This should have been out yesterday. My apologies for that, the day sort of got away from me. But here it is now for a bit of light weekend reading. This edition just contains what I already had ready to go for last week’s Techno-Optimist, so it’s missing some recent updates (Grokipedia!). Not to worry, the next one will be more complete.

Hope you all enjoy, let me know what you think in terms of format (e.g., breaking it down into sections?), which I may alter somewhat over the next few issues until I settle on a format that works best.

Alright, hang on tight as we accelerate up the curve!

Subscribe now

“Humans are not perfect, and neither is AI. But together, we can create something extraordinary.” — Andrew Ng

OpenAI’s ChatGPT now has something called ‘Instant Checkout,’ which is exactly what it sounds like, allowing users to shop within the app. At the moment it’s partnered with Etsy and Shopify, but I’m sure that will expand in short order.

The company also launched a new app called Sora, which from the sounds of it is an enhanced video editing / generation platform. CEO Sam Altman wrote a blog about it (or had ChatGPT write it) here. “Inside the app, you can create, remix, and bring yourself or your friends into the scene through cameos—all within a customizable feed designed just for Sora videos.”

In what can only be described as closing the barn door after the horse has bolted, there will now be parental controls in ChatGPT. LOL, I wonder how long it will take teens to get around those?

Anthropic recently introduced their latest model: Claude Sonnet 4.5, “the best coding model in the world.” According to the company, it’s the strongest model for building complex agents. It’s the best model at using computers. And it shows substantial gains on tests of reasoning and math.”

One very interesting thing that sort of flew under the radar is that Claude Sonnet 4.5 managed to do 30+ hours of autonomous coding, greatly surpassing expectations. We’re well on the way to AI agents that really do operate more or less independently for indefinite periods of time.

Google DeepMind rolled out their latest video generation model: Veo 3.1, which has a “deeper understanding of the narrative you want to tell, capturing textures that look and feel even more real, and improved image-to-video capabilities.” Veo can be accessed in Google’s Gemini App for anyone who’s curious.

To say that there’s a lot happening in AI infrastructure is a bit of an understatement, with major companies seemingly announcing new megaprojects every other week. Andrew Chien, a computer scientist at the University of Chicago, said that, “I’ve been a computer scientist for 40 years, and for most of that time computing was the tiniest part of our economy’s power use. Now, it’s becoming a large share of what the whole economy consumes.”

Elon Musk has said that just as his company xAI “was the first to build 1GW of coherent training compute, they will also be the first to reach 10GW, 100GW, and 1TW.”

In another blog, Sam Altman said the vision at OpenAI is “to create a factory that can produce a gigawatt of new AI infrastructure every week.” Easier said than done of course, but wow the acceleration is real. According to an internal memo, OpenAI has 9x’d their capacity this year, and wants to do another 125x by 2033. Something like 250GW would be needed, equivalent to 250 large nuclear power plants.

OpenAI and chip maker AMD have signed an agreement to “deploy 6 gigawatts of AMD GPUs over multiple years.” They also announced a plan with Nvidia “to build AI data centers consuming up to 10 gigawatts of power, with additional projects totaling 17 gigawatts already in motion.” In an entirely separate announcement, OpenAI is partnering with Broadcom (semiconductor manufacturer) “to deploy 10GW of chips designed by OpenAI.” So that’s 26GW worth of new announcements within a few weeks, just by OpenAI.

Google has decided to spend some big bucks in India, $15 billion “to establish our first full-atack AI Hub in India.” The money will be spent over the next 5 years to set up the 1GW data center and AI Hub. The company also announced $9 Billion for South Carolina, to be spent through 2027, for expanding their “Berkeley County data center campus and support the continued construction of two new sites in Dorchester County, strengthening the state’s role as a critical hub for American infrastructure.”

Meta (the parent company of Facebook) announced a small (relatively speaking) “state-of-the-art data center in El Paso, Texas, that will have the ability to scale to 1GW.” It’s all part of their plan to build towards superintelligence.

I saw an estimate that by 2035, AI data centers could be using 1,600 TWh a year. By my rough calculation, that would require 182GW of capacity. I think that may be a serious underestimate.

What will all that compute be used for? Photo and video generation for sure. Definitely writing billions of emails, reports, and memo’s. Powering robots as they take our jobs? Possibly. But one incredible thing they’ll be doing is making our medical care better, and rapidly accelerating scientific discoveries.

A new study showed that when multiple AIs work together, this “AI council” aced the US medical licensing exams, scoring 97% and outperforming any individual AI.

AI also passed the CFA Level III, the world’s hardest finance exam. Scoring 79% (most humans fail).

A Google AI called C2SD-Scale 27B (creative, I know) managed to “generated a novel hypothesis about cancer cellular behavior, which scientists experimentally validated in living cells.” Essentially what this means is that the model “demonstrated fluency in the language of human biology. Listening to what cells were saying, understanding their context, and talking back with a new idea that nature validated.”

Another group working in California used AI to map 1,300 regions of the mouse brain. Here’s the kicker: some were previously uncharted, and have now been discovered thanks to the work this AI-human team has done.

It’s not just in biology either, a theoretical physicist used GPT-5 to rediscover the results of some work he’d done on black holes, completely independently and without knowing the correct result beforehand.

It’s early days still, but I expect we’ll start seeing some really incredible results in the next year or so, the number and usefulness of which will grow exponentially thereafter.

That’s all for now, AI Updates will be back four weeks from now. Let me know what you thought!

Thank you all for reading — and until next time, keep your eyes on the horizon.

-Owen

Welcome to the thirty-second edition of Techno-Optimist, your destination for all the latest updates and commentary on space, science, technology, medicine, energy, AI, and much more. To the new readers since last time, welcome aboard.

Question for all of you on newsletter format: do you like the current setup with long form items at the beginning, followed by the shorter ones in various categories? An alternative setup might have most long form items in their various section. E.g., nuclear energy updates would go in the Energy, Engineering, & Physics section. A third option is taking some longer items and sending them out as separate, shorter newsletters. AI updates for instance, which to be honest is getting a bit long. Maybe Robot updates, and Nuclear Energy updates as well.

I’m going to test that third option. You’ll notice that AI Updates, Nuclear Energy Updates, and Robotics Updates (three of the longest) are missing this time from the long form item section. I’ll be sending them out the next three Saturdays, with a poll next time to see what you all think of it.

As always, we’ve got a lot to cover, so let’s dive in.

Subscribe now

“The human society of the future desperately needs a frontier. Psychologically and ethically, we need a venue where “a majority of one’s efforts are not in competition with others but directly against nature.” —J. Storrs Hall

Enceladus. Recent analyses of data collected by NASA’s Cassini spacecraft have uncovered new complex organic molecules in the plumes of Saturn’s moon Enceladus, suggesting that its subsurface ocean may harbor rich chemical activity. These findings bolster the case for Enceladus as a prime candidate in the search for extraterrestrial life.

In 2005, Cassini detected evidence of a liquid water ocean beneath Enceladus’s icy crust, with plumes of seawater venting from the moon’s south pole. Subsequent studies of the ice grains in these plumes revealed multiple molecular building blocks of life. Now, scientists revisiting Cassini’s data have identified suites of complex organic molecules in the plumes, including aliphatic and cyclic compounds, indicating that the ocean’s chemistry is more intricate than previously understood. These complex molecules suggest that Enceladus’s ocean may possess the necessary chemical ingredients to support life. The presence of both oxidized and reduced compounds in the plumes indicates that chemical reactions could be occurring within the ocean, potentially leading to the formation of biologically relevant molecules.

The discovery of these complex organic molecules enhances the scientific community’s interest in Enceladus and underscores the importance of continued exploration of its subsurface ocean. Future missions to Enceladus could provide further insights into the moon’s potential to support life and contribute to our understanding of the conditions necessary for life elsewhere in the universe.

How to stop an asteroid. I love Kurzgesagt videos, they’re educational and just overall awesome. This one on how we can stop dangerous asteroids is no exception. The problem with ideas like shooting nukes at a potential impactor—especially a large one—is that the explosion actually wouldn’t do much damage in space, merely making a smallish crater on the surface. For true destruction, you need to actually be inside the rock (hence the plot of both Deep Impact and Armageddon). But there’s an easier way, tungsten spikes 2m (6.5 ft) long and weighing 2.5 tons, which are placed in the path of the asteroid. Because it’s already moving so fast relative to Earth and the spikes, no propulsion is needed, just hunks of metal that will impact and vaporize both themselves and a significant amount of rock and ice. A small group of these would obliterate a ‘small,’ city or regional destroyer sized asteroid.

For a true planet killer, things get a bit more tricky. A series of larger spikes would be placed in a line, digging a deeper and deeper hole into the asteroid as it impacts them. Last in the series would contain a nuke (there, you knew it was coming), which would be effective in shredding the asteroid due to how far inside it is when it explodes. We have the technology to do all this now, and should probably start thinking seriously about setting up a Planetary Shield of this type, combined with the telescopes necessary to give us enough warning to implement it.

Starship flies again. Starship just had its eleventh test flight earlier this month, and as an excellent Ars Technica article summarized it, “SpaceX finally got exactly what it needed from Starship V2.” The V2 refers to the 2^nd generation Starship design, of which this flight was actually the last—going forward we’ll start seeing the 3^rd generation ships testing technologies needed to develop the space economy, do missions to the Moon, and colonize Mars. While lessons were learned and valuable data gathered on every flight, this was probably the most successful so far, hitting every objective and paving the way for even more spectacular tests next year.

(photo) In other news, SpaceX talked about their “fully automated bakery in Florida,” that is now set “to produce thousands of heat shield tiles per day to outfit the coming fleet of Starship vehicles.” The company also just announced timelines and pricing for cargo Starship Moon and Mars missions, which will begin in 2028 and 2030 respectively. The cost for each is $100M per metric ton, or $100,000 per kg. For reference, the Apollo program was at least $1M per kg to the lunar surface.

Have a look at some of the photos I’ve added here or linked, they really look like something out of a sci-fi.

Full coverage of Starship’s eleventh flight test.

Engines light and liftoff.

Liftoff.

Super Heavy hot-staging and boostback burn.

Starship in space.

Starship reentering Earth’s atmosphere.

Starship ocean landing 1.

Starship landing 2.

SpaceX finally got exactly what it needed from Starship V2

Timeline and cargo for Moon and Mars.

Cancer treatment updates. Lots happening as always, I just hope some of this can make it into mainstream treatment sooner rather than later. MIT and Harvard teams have “created engineered CAR-NK cells [natural killer cells are part of the immune system] that can hide from the immune system and more effectively destroy cancer.” Over at UCLA, scientists used an engineered antibody to “target a protein…found on the surface of certain aggressive cancer cells and the supportive stroma cells surrounding them. By pairing the antibody with radioactive particles, scientists can both visualize tumors for precise imaging and deliver targeted radiation therapy directly to cancerous tissue, while sparing healthy tissues.”

A new “super vaccine” using nanoparticles managed to prevent “melanoma, pancreatic, and triple-negative breast cancers in mice—with up to 88% remaining tumor-free.” While in another breakthrough a therapy using light to heat up tiny flakes of tin killed “up to 92% of skin cancer cells and 50% of colorectal cancer cells. It did so without harmful effects on healthy human skin cells, demonstrating the safety and selectivity of this approach.”

One more: a new immunotherapy combo managed to completely annihilate tumors in advanced colorectal cancer.

2025 Nobel Prizes. This year’s Nobel Prizes have been announced. For Chemistry, it was awarded for “the development of metal-organic frameworks.” In physics three scientists won ““for the discovery of macroscopic quantum mechanical tunnelling and energy quantisation in an electric circuit.” And in physiology or medicine the award was given to three researchers “for their discoveries concerning peripheral immune tolerance.” Lastly for our purposes is the prize for economics, “for having explained innovation-driven economic growth.” It was actually split into two parts, with part one “for having identified the prerequisites for sustained growth through technological progress” and the other half “for the theory of sustained growth through creative destruction.”

Funny story, Fred Ramsdell (who was one of the winners for medicine) was out on a multi-day hiking trip and completely off the grid when the prizes were announced. He had no idea he’d won until sometime later. (Nobel Prize Chemistry) (Nobel Prize Physics) (Nobel Prize Psychology or Medicine) (Nobel Prize Economics)

Space

• Inversion Space has built a reusable capsule called Arc that they say will be “the world’s first space-based delivery vehicle,” allowing “on-demand delivery of cargo and effects to anywhere on Earth in under an hour.” Both military and civilian applications are pretty profound, everything from logistical support and resupply, to disaster relief in hard to reach areas. Oh, and it’s all autonomous too. (Inversion)

• Using AI to analyze 20 years of orbital images, scientists tracked 1,039 Martian dust devils and discovered winds reaching up to 158 km/h (98 miles/h)—far exceeding previous estimates from rover measurements and climate models. The tornado like whirlwinds, most common during spring and summer between 11 AM and 2 PM local time, lift massive amounts of dust that shapes Mars’s weather and climate. Better understanding these powerful winds will help engineers design safer landers and rovers—and improve predictions for future Mars missions threatened by dust coating solar panels or buffeting equipment during descent. (Phys.org)

• Impulse Space plans to deliver 6 tons of cargo to the Moon annually (across 2 missions per year) starting in 2028 using their Helios kick stage paired with a new in-house lunar lander—each combo hauling 3 tons per mission without needing in space refueling. Impulse is betting their vertically-integrated rapid development approach can help kickstart Moon based infrastructure and resource utilization. (Impulse Space)

• Scientists finally explained why Jupiter and Saturn have eastward equatorial jet streams while Uranus and Neptune flow westward—all four result from fast rotating convection cells that can settle into either direction, “establishing a direct link between jet direction and atmospheric depth.” The unified model shows a bifurcation where very similar conditions produce different stable states, solving a decades old why similar factors (little sunlight, internal heating, fast rotation) result in jet streams moving in different directions. Understanding this mechanism could help decode atmospheric dynamics on gas giants elsewhere in the universe too. (Phys.org)

• Webb detected seven carbon-rich molecules—including benzene and acetylene—in a potential moon-forming disk around exoplanet CT Cha b, 625 light-years away, providing the first direct chemical measurements of an environment where moons are born. The discovery lets astronomers witness moon formation in real time and compare it to how Jupiter’s Galilean moons likely condensed from a similar disk billions of years ago—with nine more accessible systems queued for Webb’s survey. (NASA Web Telescope)

• James Webb has detected intense auroras on SIMP-0136, a rogue planet about 20 light-years away with no parent star. The auroras are so powerful they heat the upper atmosphere, creating a thermal inversion where temperatures rise with altitude—opposite to what we see on Earth. The planet is also cloaked in silicate clouds, more like floating sand than water vapor. These findings show that even starless worlds can have complex weather systems, challenging our models of planetary atmospheres. They also open the door to studying the unique environments of rogue planets in unprecedented detail. (Universe Today)

• Reanalyzed data from NASA’s 1970s Pioneer Venus mission indicates that Venus’ cloud aerosols are approximately 60% water—predominantly in the form of hydrated compounds like ferric and magnesium sulfates. This contrasts with the previously held opinion that the clouds were mainly sulfuric acid. The discovery was made by examining mass spectrometry data from the Pioneer Venus Large Probe, which had been archived on microfilm. Researchers found that as the probe descended, aerosol particles clogged the instruments, leading to temporary drops in CO₂ readings. By analyzing the temperatures at which these aerosols decomposed, they identified significant releases of water and sulfur compounds. This finding suggests that Venus’ clouds contain more water than previously thought, which could have implications for the planet’s potential to support life. (Universe Today)

• NASA has just announced its 2025 astronaut candidates. For the first time ever, one of them has already been to space (we’ll soon see a lot more of this). Anna Menon is a former SpaceX engineer who flew aboard the Polaris Dawn mission last year, and if all goes well she’ll have the opportunity to walk on the Moon sometime within the next decade. (NASA) (Toby Li)

Engineering, Energy, & Physics

• We’ve known it for a while, but now there’s finally proof. Researchers at UT Austin have mathematically proven that quantum computers can “unconditionally outperform classical computers.” Using a task where a quantum system (“Alice”) sent encoded states to another (“Bob”), the team showed that a classical computer would need 62 bits of memory to match the quantum device’s performance—while the quantum processor used just 12 qubits. This marks the first definitive proof of “quantum information supremacy,” demonstrating that real quantum hardware can exploit its vast exponential memory to solve problems forever out of reach for classical machines. (Phys.org)

• Anduril Industries recently unveiled EagleEye, “the family of warfighter augments that place mission command & AI directly into the operator’s helmat.” It brings command-and-control, live sensor feeds, and autonomous decision aids into a single AR (augmented reality) helmet so soldiers can see threats, sensor tracks, and task nearby drones in real time. It’s the same AI networking Anduril uses for vehicles, giving soldiers immediate situational awareness and direct control—equipping soldiers “with the ability to plan, fight, and survive while connected to every asset in the battlespace.” (video) Take a look at this video, it really is science fiction type technology. (Anduril Industries) (@anduriltech)

• In a nutshell, Heisenberg’s Uncertainty Principle says that you can’t know precisely both the position and momentum of a particle. As you get a better measure of one thing, the uncertainty level on the other rises. But for any natural law, there usually seems to be a way around it if we’re clever and persistent enough—two things that humans are very good at. Physicists in Britain and Australia have managed to sidestep the restriction and measure both position and momentum simultaneously. They did it by shifting “the unavoidable quantum uncertainty to places we don’t care about…so the fine details we do care about can be measured more precisely.” (Phys.org)

• Speaking of breaking (really just bending to the point where you’re playing jump rope with them) the rules of physics to get stuff done, scientists have just figured out how to get around the Carnot principle, which determines “the maximum efficiency of heat engines.” It turns out that the Carnot principle doesn’t really apply in the quantum realm where “particles become correlated, interacting in ways that defy classical physics.” I’m not going to pretend I understand this, but the results indicate that nanometer scale machines “functioning at the atomic scale are capable of converting not only heat but also correlations into usable work. What’s more, these systems can generate more output, allowing the efficiency of a quantum engine to exceed the conventional Carnot limit…Tiny motors, no larger than a single atom, could become a reality in the future” In other words, we can power our nanobots—or nanites for all us Stargate fans. (Interesting Engineering)

  Here’s a good little explainer of quantum physics for anyone interested.

• Scientists have developed an ultrathin LED that emits light almost indistinguishable from natural sunlight. The flexible film, just nanometers thick, uses quantum dots to produce a full, balanced color spectrum while cutting harsh blue light. It’s efficient, easy on the eyes, and can be tuned for different environments. Researchers say the technology could transform indoor lighting, next generation computer or television displays, and even horticultural systems needing realistic sunlight. Not to mention potential uses on the Moon, Mars, and beyond. (Science Daily)

• Why can’t we build things like this anymore? Not only is it absolutely spectacular, but it cuts what was formerly a two hour drive down to a couple minutes. There’s also a restaurant at the top, and did I mention it can create an artificial waterfall? And, the cherry on top of this cake is they built it in less than four years. (Collin Rugg)

• Berkeley Lab has developed a compact laser–plasma accelerator that can generate muon beams on demand, using a device just 30 cm long. The system accelerates electrons with intense laser pulses and collides them with lead targets to produce muons (most of which are flying off in the same direction), enabling non-destructive imaging through dense materials, including rock hundreds of meters thick. Delivering over 40 times more muons than natural sources like cosmic rays, the accelerator could revolutionize muon tomography for scanning volcanoes, ancient structures like the pyramids, and nuclear materials. I wonder if there’s an application here for muon catalysed fusion? (Phys.org)

Medicine & Biotech

• In another case of the truth being stranger than (science) fiction, a team has built a swallowable bioprinter—a pill sized device that can print living tissue directly inside the body. Once swallowed, magnets guide it to a damaged site, like a stomach ulcer, where a near infrared laser triggers it to release a bio-ink that forms a living patch over the wound. In animal tests it repaired internal injuries without surgery, marking a major step toward fully internal regenerative medicine. Future versions could print therapeutic cells or growth factors on demand, letting patients one day “swallow a cure” instead of going under the knife. (Interesting Engineering)

• A gel made from deoxyribose—the sugar molecule in DNA—triggered 80-90% hair regrowth in balding mice, matching the effectiveness of minoxidil (a leading hair loss reversal medication) without hormonal side effects. Researchers stumbled upon the discovery while studying wound healing, noticing treated mice regrew fur faster than controls by boosting blood vessel formation around hair follicles. The naturally occurring sugar could offer a simple alternative without the side effects of current treatments, and potentially restore hair growth after chemotherapy. Human trials are still needed, but this looks promising. (Science Alert)

• Researchers at UC Irvine have discovered that supplementing a specific polyunsaturated fatty acid—not the usual DHA—can reverse age-related vision loss in mice by restoring vital very-long chain lipids in the retina, countering the decline driven by the ELOVL2 gene. This simple intervention boosted visual performance and reversed cellular aging markers without needing to tweak the gene itself. Genetic variants in ELOVL2 were also linked to faster progression of age-related macular degeneration (AMD), paving the way for early risk screening. Longer term, this could lead to easy dietary therapies to prevent blindness in the elderly, and even extend to rejuvenating the immune system. (Science Daily)

• A rheumatoid arthritis drug dramatically improved heart attack recovery in mice by blocking T-cell activation and reducing post heart attack inflammation—preserving heart function. Around 25% of heart attack survivors die within three years due to inflammation-induced heart failure, but this anti-inflammatory approach could change that by targeting the root problem (overactive T-cells after a heart attack), rather than just treating symptoms. (Medical Xpress)

• A new gene therapy called AMT-130 has slowed Huntington’s disease progression by 75 percent in clinical trials—the first treatment to ever accomplish that for the fatal neurodegenerative disease. The therapy uses custom DNA delivered via ‘minor’ brain surgery to instruct cells to produce RNA that destroys the mutant huntingtin protein causing the disease. One patient in the trial, who was medically retired due to the disease, has now returned to work—and researchers believe this approach could eventually work for other neurodegenerative conditions like Parkinson’s and motor neuron disease. (IFL Science)

• University of Bath researchers engineered a tiny peptide that locks proteins in their healthy shape, preventing the toxic clumping that drives Parkinson’s disease—demonstrated in worm models. The treatment “checks several important boxes: it’s durable, and it can survive inside cells without causing any toxic side effects.” The same approach could tackle Alzheimer’s and Lewy body dementia, though delivering peptides to human brains remains a challenge to solve before clinical use. (Science Alert)

• Two new compounds, K102 and K110, could become the first treatments to actually repair the nerve damage caused by multiple sclerosis—not just slow it down. After screening over 60 candidate molecules, researchers identified these drugs that both rebuild the protective myelin sheath around nerves and balance immune responses in human cells and mouse models. Licensed by Cadenza Bio and backed by over a decade of collaboration, K102 is now advancing toward clinical trials—and the approach could eventually work for other neurological conditions like spinal cord injury and traumatic brain injury. (Science Daily)

• MIT researchers have discovered that the amino acid cysteine—found in meat, dairy, legumes, and nuts—can dramatically regenerate intestinal tissue by activating CD8 T (immune system) cells to release IL-22, a molecule that stimulates stem cell growth. The finding could offer a simple dietary approach to help cancer patients heal from radiation or chemotherapy damage, and researchers are now investigating whether cysteine might trigger regeneration in other tissues like hair follicles. Unlike cysteine produced naturally in the liver, dietary cysteine concentrates directly in the gut where it’s most needed—turning an ordinary nutrient into a targeted healing tool. (SciTechDaily)

• Taurine—a common amino acid found in energy drinks and supplements—can actually reverse alcohol induced brain damage by stimulating neurogenesis in the hippocampus, according to new research. Scientists gave taurine to rats with ethanol damaged brains and found it increased cell proliferation by 145.8% and cell survival by 54%, reducing inflammation and restoring cognitive function through protection against oxidative stress. Intriguingly, taurine only triggered neurogenesis in damaged tissue, not healthy brains—suggesting it acts like a targeted repair mechanism that activates only when needed, potentially offering a simple therapeutic approach for reversing neurological damage from alcohol abuse. (MDPI via X)

• Japanese researchers have engineered a “supercharged” vitamin K that’s three times more effective than natural vitamin K at promoting neuron growth—potentially offering a way to reverse neurodegenerative diseases like Alzheimer’s and Parkinson’s. The enhanced compound, created by linking vitamin K with retinoic acid, activates a brain receptor to stimulate neurogenesis, crosses the blood-brain barrier efficiently, and remains stable in living systems. Unlike current medications that only ease neurodegenerative disease symptoms, this approach could actually replenish lost neurons and restore brain function—turning a vitamin supplement into a regenerative medicine. (Science Daily)

• Heparin, a cheap, widely available drug traditionally used for blood clots, has been shown to cut COVID-19 ventilation rates and death risk in half when delivered directly to the lungs via inhalation. The drug offers a rare triple threat: it’s antiviral, anti-inflammatory, and anticoagulant all at once, making it potentially effective against any respiratory infection including influenza, RSV, and pneumonia regardless of whether they’re caused by viruses or bacteria. Researchers are now developing an optimized inhalation formulation and planning trials across Europe—and because heparin is inexpensive, it could provide accessible pandemic defense for low-income countries when the next respiratory threat emerges. (Science Daily)

• Researchers have cracked CRISPR’s password system—”CRISPR needs a password, a specific DNA signal called a PAM (protospacer adjacent motif), located next to the target gene, before it can make a cut”—using a new method that maps these codes directly in human cells by analyzing over 10,000 sequences in every single cell studied, instead of just 1-3 sequences, which had been done before. While traditional methods could typically identify 4- to 6-letter passwords, the new approach can decode at least 10 letters easily, dramatically expanding where CRISPR tools can target the genome. This breakthrough accelerates development of safer, more precise gene therapies by revealing exactly which DNA sequences each CRISPR variant can access—essentially creating a comprehensive map of the genome’s editing possibilities. (Phys.org)

• Scientists at the University of East Anglia have developed a 96% accurate blood test for Chronic Fatigue Syndrome (ME/CFS) by analyzing how DNA folds in 3D—revealing a unique pattern that appears consistently in patients but not in healthy people. Using EpiSwitch technology that reads “epigenetic” markers rather than fixed genetic code, the test achieved high sensitivity and specificity, finally offering objective diagnosis for a condition affecting over 400,000 people in the UK (and many more elsewhere) who are often told their illness is “all in their head.” The breakthrough could also pave the way for a similar test to diagnose long COVID—and the immune pathways identified may guide development of targeted treatments for both conditions. (Medical.net)

• Speaking of long COVID, Japanese researchers have finally identified the biological cause of long COVID brain fog using advanced PET imaging—discovering widespread increases in AMPA receptor (AMPAR) density directly correlated with cognitive impairment severity. The imaging technology can distinguish long COVID patients from healthy individuals, providing the first objective biomarker for a condition affecting over 80% of long COVID sufferers. The discovery not only validates brain fog as a measurable biological condition rather than something “in people’s heads,” but also reveals that drugs suppressing AMPAR activity could potentially treat the cognitive symptoms plaguing hundreds of millions worldwide. (Science Daily)

• Engineered nanoparticles slashed amyloid-beta (the toxic protein clumps in Alzheimer’s) levels by 50-60% within one hour in Alzheimer’s mice by reactivating the blood-brain barrier’s natural waste removal system, rather than directly attacking the proteins. Three shots rejuvenated LRP1 transporters that ferry toxic proteins out of the brain into the bloodstream for disposal, with cognitive improvements lasting six months—treated elderly mice navigated mazes easily and even built nests again. The approach repairs the barrier itself instead of bypassing it, potentially avoiding the brain bleeds and side effects plaguing current anti-amyloid antibody therapies. (Singularity Hub)

• Scientists have discovered dynamic nanotube tunnels connecting the dendrites of neurons in mouse and human brains—creating a “web” similar to how fungal networks connect tree roots underground. These dendritic nanotubes transmit electrical calcium signals, sprout and dissolve within minutes to hours, and appear to shuttle toxic proteins from damaged neurons to immune cells while receiving healthy mitochondria in exchange. The discovery adds a fourth communication method beyond synapses, gap junctions, and vesicle transport—and could explain how the brain coordinates repairs in conditions like Parkinson’s and Alzheimer’s, where these rescue tunnels have been observed forming between struggling neurons and protective microglia. (Singularity Hub)

• Engineers at UMass Amherst have built artificial neurons using protein nanowires from electricity producing bacteria that operate at just 0.1 volts—matching biological neurons and using 100 times less power than previous designs. The breakthrough enables direct communication with living tissue without amplification, potentially eliminating the clunky, power hungry interfaces in wearable sensors that currently need to boost signals before processing them. Beyond bio-inspired computing that rivals the brain’s remarkable efficiency (20 watts versus a megawatt for ChatGPT to write a story), the same bacterial nanowires have already been used to create sweat powered biofilms, disease sniffing electronic noses, and devices that harvest electricity from thin air. (Science Daily)

• Chinese scientists have managed to reverse aging in monkeys across 54% of tissue types using a stem cell therapy. In moneys, this reversal was roughly 3-5 years of biological age, which could translate to 9-15 years in humans. Some caveats: this has not been replicated yet, the scientists used a new epigenetic methodology to measure aging reversal instead of more standard ones, and a method to generate stem cells from each individuals own cells would need to be incorporated into the treatment. All that said, this is pretty exciting. (Bryan Johnson)

Agriculture

• Michigan State researchers discovered that a metabolic compound normally involved in creating plant “sunscreen” can reprogram a protein that senses UV light. By modifying how plants sense light and produce protective compounds, crops could be engineered to grow more efficiently in low light or harsh environments—potentially turning light perception itself into a tunable dial for improving agricultural resilience. (SciTechDaily)

• Scientists have finally solved why photosynthesis only uses one of two symmetrical pathways in Photosystem II (“This complex captures sunlight and splits water molecules, releasing oxygen and sending electrons onward to other molecules in the chain of energy transfer”). There are two almost identical branches through which electrons can move in Photosystem II, but these new findings show that the second (D2) branch has twice the energy barrier, making electron flow energetically impossible. The researchers found the D2 pathway has 100x higher resistance than D1, explaining decades of puzzling experimental results. The discovery could enable artificial photosynthesis systems by revealing how to rewire electron flow—potentially allowing us to engineer more efficient solar-to-fuel conversion by swapping specific pigments to overcome nature’s energy blocks. (Science Daily)

Weird & Wonderful

• Mindblowing is a good way to describe this. Someone in Minecraft “built a 5 million parameter language model,” then they “trained it, [and] equipped it with basic conversational ability.” Creating AI within a simulated world. (@tokenbender)

• No driver, no ticket? In San Bruno, California, police pulled over a Waymo self-driving car for making an illegal U-turn—only to find there was no human driver, leaving them unable to issue a citation. The department joked that their ticket books don’t have a category for “robot,” highlighting a legal gray area as autonomous vehicles become more common. The episode underscores how quickly traffic laws will need to adapt to the realities of autonomous driving. (Rocky Mountain Outlook)

• Paleontologists in Argentina have unveiled a new apex predator: Joaquinraptor casali, a megaraptor measuring approximately 23 feet (7 meters) in length. Distinguished by its formidable forelimbs and “claws like hedge trimmers,” it would have been a formattable predator. Remarkably, the fossil was discovered with the leg bone of an ancient crocodile relative still clutched in its jaws. (Live Science)

• Elon Musk’s xAI is working on a new project called MACROHARD, the goal of which is to “create a company that can do anything short of manufacturing physical objects directly, but will be able to do so indirectly, much like Apple has other companies manufacture their phones.” It’s a play off of Microsoft, and is likely inciting chuckles from junior high school students everywhere. (Elon Musk)

Photos & Videos

Scientists have used data from the Gaia space telescope to create “the most accurate three-dimensional map of star-formation regions in our Milky Way galaxy.” It’s pretty spectacular. (ESA)

Have a look at the beautiful microscopic world captured in Nikon’s Small World 2025 Photomicrography Competition. (Nikon)

NASA’s JWST took a close look at the Sagittarius B2 molecular cloud, which is actually the “most massive, and active star-forming region in our galaxy,” despite being only a few hundred light years away from it’s central black hole. (NASA Webb Telescope)

Recommendations & Reviews

Today I’d like to recommend an excellent essay by on Substack. It’s called Pollution’s Peak, and in a nutshell makes the (correct) argument that while economic development initially has a negative environmental impact, it turns positive at higher levels. “Once again, growth is not the disease; it is the cure.” (Risk & Progress)

That’s all for this edition. Techno-Optimist will be back in your inboxes on November 22^nd. Don’t forget that next Saturday will be the first test editions of AI Updates #1, with Nuclear Energy Updates #1, and Robotics Updates #1 coming on the following Saturdays.

Welcome to the thirty-first edition of Techno-Optimist, your go-to destination for all the latest updates and commentary on space, science, technology, and a whole lot more. To all the new readers since last time, welcome aboard!

A quick bit of housekeeping. For the longer form items at the beginning, I’m going to start just embedding links instead of having them cluster at the end. It’s a bit tidier, and it’ll be easier to find more info on any specific thing that catches your interest. I’ll carry on using the same format for the shorter items.

Alright, we’ve got a lot to cover, so let’s dive in.

Alright, let’s dive right in.

We pray for one last landing
on the globe that gave us birth
To rest our eyes on fleecy skies
and the cool green hills of Earth.

–Robert A Heinlein

Potential biosignature on Mars. NASA’s Mars Perseverance rover has provided what could be a potential biosignature from our rusty neighbor. The original press release was in July 2024, showing spots on a rock that scientists now think are likely the result of microbial activity. Since the initial discovery, researchers have been trying to come up with other non-biological explanations for the chemical signatures observed—that’s how science works—and so far, nothing explains it so well as life. There are in fact ways to produce these minerals without biology, but they involve things like “sustained high temperatures, acidic conditions, and binding by organic compounds,” with “the rocks at Bright Angel not showing evidence that they experienced high temperatures or acidic conditions, and it is unknown whether the organic compounds present would’ve been capable of catalyzing the reaction at low temperatures.”

The best way to answer the question definitively is to send a geologist to Mars with a thin section kit. You’d have an answer within a day or two. Personally, I think it’s a virtual certainty that life previously existed on the Martian surface, and that it’s still extant underground—akin to the deep biosphere here on Earth. It won’t be “alien” life however. Though the microbial species will probably be different, Martian life will be Earth life, due to the billions of years of planetary “cross talk” in the form of meteorites blasted off the surfaces of one planet and landing on the surface of the other. If (when) we do confirm life existed on Mars, it will show that if life exists on one planet in a system it will probably spread throughout that system, or at least to neighboring planets if the conditions are right. Exciting times, we need humans on Mars ASAP to answer these questions.

Atomic energy updates. Starting with fusion: (photo) Some huge news for Commonwealth Fusion Systems. They just signed a power purchase agreement with Italian energy (mostly oil) company Eni to buy over $1 billion dollars worth of fusion energy from Commonwealth’s ARC plant in Virginia—which just recently received site approval. This won’t be happening tomorrow, but the company expects (and I agree) that they’ll have fusion produced electrons on the grid by the early 2030s. This will all be funded by their recent Series B2, which raised $863 million.

Of importance for the broader industry, the U.S. Congress has introduced the Fusion Advanced Manufacturing Parity Act, which will “extend the 45X Manufacturing Production Tax Credit to fusion. This bill helps scale fusion manufacturing here in the U.S. by providing a 25% credit for fusion components, securing critical minerals, and building supply chains.” This will help commercial fusion production to become a reality, and will help prevent things from being stuck merely at the demonstration or pilot plant level.

The U.S. Department of Energy has also announced $134 million for fusion R&D. $128 million will go to the new FIRE (Fusion Innovation Research Engine) program, while the rest will be earmarked for the existing INFUSE (Innovation Network for Fusion Energy) awards.

Now on to fission: The Tennessee Valley Authority has just signed a deal with Texas based ENTRA1 Energy for 6 GW of small modular reactors, which will consist of multiple smaller factory produced modules.

Across the pond, U.S. company X-energy now has a deal to deploy the U.K.’s “first advanced modular reactors and pursue 6 Gigawatts of new nuclear capacity in the U.K.” This is part of a bigger nuclear power agreement between the States and the U.K., which will hopefully include multiple billions worth of private investments.

Another U.S. company, Global Laser Enrichment (GLE) has completed a large scale demonstration of laser based uranium, producing hundreds of kilograms of low enriched uranium and now plans to “significantly progress the scaling and manufacturing of its full-scale plant systems and equipment.” Their method aims to offer a faster, more flexible, and lower energy alternative to centrifuges for both LEU and HALEU nuclear fuels. If the process scales, laser enrichment could lower costs and shorten supply chains for next generation reactors.

Possibly the coolest announcement here, startup Deep Fission Nuclear has raised $30 million to build a mile deep microreactor by 2026 as part of the U.S. Department of Energy’s Reactor Pilot Program. Instead of massive surface plants, the company will lower their modular reactor into a borehole filled with water, where surrounding rock naturally pressurizes the system and eliminates the need for costly containment structures. The underground design reduces above ground construction costs by up to 80%, slashes land footprint area, and boosts safety and security by burying the reactor itself.

AI updates. Google’s Gemini has upped its game with new photo capabilities, including letting users “upload multiple images to Gemini, then telling it how to combine them with a single prompt.” (photo) For instance, taking ingredients for a pizza, and saying to combine them into a meal. Looks pretty good to me! Here’s some suggestions from Google AI to get your creativity flowing.

Gemini 2.5 has also managed to achieve “gold-medal level performance at the ICPC 2025—one of the world’s most prestigious programming contests.” Nor were they the only winner, with OpenAI managing “a perfect score of 12/12 during the 2025 ICPC World Finals…where top university teams from around the world solve complex algorithmic problems. This would have placed it first among all human participants.”

Moving on to AI infrastructure investments: the numbers are an insane preview of just how seriously the big players are taking this. What’s a few hundred billion if you’re one of the winners that gets to shape the future of human civilization? In a single event with President Trump, Meta said that they intend to spend “At least $600B through 2028,” with Apple following that up with another “$600B investment in the United States.” Elsewhere, Oracle and OpenAI signed a $300 billion contract for computing power to run future versions of ChatGPT. Microsoft will be putting another $4 billion into an AI datacenter in Wisconsin—this is in addition to an previously announced $3.3 billion datacenter there which will be up and running early next year. Just for reference, in the last year Microsoft added “over 2 gigawatts of new capacity—roughly the output of 2 nuclear power plants.” Circling back to OpenAI, they also just signed an agreement with NVIDIA to “deploy 10 gigawatts worth [of] AI chips to power the next generation of ChatGPT.”

Here’s a neat idea: AI made full length feature films.

OpenAI wants to prove that generative artificial intelligence can make movies faster and cheaper than Hollywood does today. The team is attempting to make the movie in about nine months instead of the three years it would typically take, said James Richardson, co-founder of London-based Vertigo Films. “Critterz” has a budget of less than $30 million, far less than what animated films typically cost. The production team plans to cast human actors for character voices and hire artists to draw sketches that are fed into OpenAI’s tools, including GPT-5 and image-generating models.

If this works, I’ve got a few good ideas, starting with Arthur C. Clarke’s Rendezvous With Rama.

On the efficiency side of things, Chinese researchers have developed a new brain inspired AI called SpikingBrain-1.0, that “mimics brain neurons for highly efficient training with minimal data.” Exactly how the human brain works. Apparently it can train on just 2% of the data the traditional models require. Over in Florida, engineers there have developed a light powered AI chip they say is 100x more efficient, “dramatically reducing energy use and increasing speed.” But to anyone thinking that maybe those 100s of billions of dollar could be wasted if more efficient AI is developed, never fear, Jevons Paradox virtually guarantees that we’ll just build even more of it.

I’ve sometimes joked (half seriously) that we should replace all our politicians with AI. It seems like Albania has actually done it, appointing the world’s first AI government minister, who will handle public procurement. It looks like something of a stunt, as I’m sure there are humans behind the scenes pulling the strings. But soon things like this won’t be stunts, and AI really will start making the big decisions in governments. IMO, replacing human politicians with AI will be a very positive step. I think.

How close are we getting to AGI? According to Elon Musk, Grok 5 (we’re currently on Grok 4) might have a chance of achieving it. Grok 5 starts training in a few weeks by the way. Even if it doesn’t qualify as AGI, I bet it will start to have AGI-like features. Grok 5, or maybe the model or two after it, will become increasingly difficult to distinguish from full artificial general intelligence. What comes after that for all of us is harder to predict. I hope it’s friendly.

Lastly, here’s a practical assist for those of you who use Chrome. Google has recently incorporated AI into their browser, this is a good “how to” on getting the most out of those new features.

Robots. Figure just had a huge week, making some big announcements. Firstly, they raised an over $1B Series C. Then they announced an agreement with a real estate company called Brookfield to collect real world data to help train their robots. I assume that the individual humans involved also need to sign off on this.

They also unveiled a video of their F.02 robot loading a dishwater autonomously. Not quite Jetsons level yet, but it’s getting closer.

Chinese company DEEPRobotics showed their new mass production facility where they’ll make quadruped robots. Impressive.

Perhaps even more impressive though is a new system from Google DeepMind called RoboBallet, which “can choreograph a team of robot arms with precision, working together without collisions.” This looks like something straight out of a sci-fi, and I’m pretty excited to see what it’ll be capable of building in a couple years.

Last but not least, a Chinese humanoid robot shows some pretty crazy capabilities in terms of getting back up once it’s been knocked down. It’s a little unsure of where its human opponent is, but I’m sure that’ll just be temporary.

Space

• The James Webb Space Telescope has captured its most detailed look yet at TRAPPIST-1e, an Earth sized planet orbiting in the habitable zone of its star. New observations suggest is has no thick hydrogen-helium “primary” atmosphere. However, Webb detected subtle atmospheric signals that could point to a heavier, secondary atmosphere of the type that Earth has. But I should be clear that what the study says is that there are two options: a secondary atmosphere could exist, or the planet could just be bare rock. Further observations are needed to determine which is actually the case. Given what we know about red dwarf stars generally, and this one in particular—it’s very active—I suspect bare rock is more likely. (NASA Webb Telescope)

• Penn State astronomers say the mysterious “red dots” first spotted by JWST may not be ancient galaxies after all, but a new class of object: “black hole stars.” These exotic bodies could be giant spheres of gas powered not by fusion, but by supermassive black holes at their cores “that rapidly pull in matter, converting it into energy and giving off light.” One extreme case, nicknamed The Cliff, revealed spectral signatures consistent with a black hole wrapped in a hydrogen cocoon rather than a cluster of stars. If confirmed, these objects may represent the missing link in how supermassive black holes formed so early in the universe. Reality may be weirder and more awesome than we thought possible. (Penn State)

• Ten years after first detecting gravitational waves, LIGO (the Laser Interferometer Gravitational-wave Observatory) has become a prolific black hole hunter, now spotting a merger about every three days. In its clearest signal yet—GW250114, detected earlier this year—scientists verified Stephen Hawking’s black hole area theorem with 99.999% confidence, showing that,

  “The total surface areas of black holes cannot decrease. When black holes merge, their masses combine, increasing the surface area. But they also lose energy in the form of gravitational waves. Additionally, the merger can cause the combined black hole to increase its spin, which leads to it having a smaller area. The black hole area theorem states that despite these competing factors, the total surface area must grow in size…In essence, the LIGO detection allowed the team to “hear” two black holes growing as they merged into one, verifying Hawking’s theorem.” (LIGO)

• New research suggests that the violent final stages of planetary formation may have been key to making Earth habitable. Scientists from the Southwest Research Institute and Yale found that late stage collisions—accounting for the last 1% of Earth’s growth—delivered water, altered atmospheres, and shaped tectonics across the inner planets. For Earth, these impacts likely set the stage for plate tectonics and a stable atmosphere, while Venus’ more energetic collisions may explain its present volcanic surface. The findings also suggest that a planet’s impact history could be just as important as its size and orbital position when assessing habitability. (Science Daily)

• Astronomers have discovered a new trans-Neptunian object, 2017 OF201, at the edge of the solar system—potentially large enough to qualify as a ‘dwarf’ planet like Pluto (BTW, Pluto is a planet). With an orbit taking 25,000 years to complete, stretching from 44.5 AU at its closest to 1,600 AU at its farthest, the 700 km (435 mile) wide object challenges assumptions that space beyond the Kuiper Belt is empty. (Institute For Advanced Study)

• Blue Origin’s Blue Alchemist project is moving towards “permanent and sustainable Lunar infrastructure,” completing its Critical Design Review. The system uses molten regolith electrolysis to turn Lunar soil into oxygen, metals, and silicon and other metals, and construction materials—”all critical elements and systems needed for near-term, low-cost, and sustainable bases, growing settlements, and eventual cities.” By producing all this directly on the Lunar surface, Blue Alchemist could cut mission costs by up to 60% and enable permanent human settlement. The technology will next undergo autonomous testing in a simulated lunar environment in 2026. (Blue Origin)

Energy, Engineering, & Physics

• A US–China team has built a one step process that converts mixed plastic waste—including notoriously hard to recycle PVC (polyvinyl chloride)—into the main components of gasoline, plus hydrochloric acid, with efficiencies up to 99% at room temperature. If it scales, the method could turn plastic waste into a feedstock for fuel while recovering industrially valuable HCl, potentially transforming one of recycling’s biggest headaches into a practical resource. The irony here is that if this works commercially just as there’s pressure for vehicles to go electric, gasoline might suddenly become a lot cheaper, driving up demand for ICE vehicles. (Interesting Engineering)

• MIT physicists have outlined a way to build the world’s first “neutrino laser”—by cooling a cloud of radioactive rubidium-83 atoms into a Bose-Einstein condensate, where they will rapidly decay in minutes instead of the usual 82 days for rubidium-83. The decay would be mostly in sync, releasing a burst of neutrinos like photons in a conventional laser. If it works, this technology could potentially enable communication systems that beam straight through Earth. (SciTechDaily)

• A recent MIT geophysics experiment used “lab-quakes” (miniature, controlled seismic events taking place inside the lab) to map out exactly where quake energy goes. The result is surprising: about 80% of the energy ends up as heat right around the fault, only 10% causes the shaking we feel, and less than 1% goes into actually breaking the rock (fracturing & grinding). These quakes got so hot, they briefly melted rock and generated molten glass. “In some instances, we saw that, close to the fault, the sample went from room temperature to 1,200 degrees Celsius in a matter of microseconds, and then immediately cooled down once the motion stopped,” said the study’s author. (Interesting Engineering)

• Meta isn’t giving up on smart glasses, with CEO Mark Zuckerberg demo’ing their new “Meta Ray-Ban Display” at the company’s annual developer conference. They’re actually ready to go now too, starting to ship September 30^th. At $799 though, I do wonder how popular they’ll be, given the lackluster performance of previous smart glasses. That said, they do look pretty slick, and incorporate a wrist band to allow for hand controlled actions, avoiding the need to actually touch the glasses to use them. (TechCrunch)

• Flying car company Jetson Aero just completed their first delivery. The buyer was Palmer Luckey of Anduril Industries, who proceeded to buzz around on a low altitude practice flight. Honestly, it looks like a lot of fun, I’d like to try something like it one day. (JetsonAero)

Medicine & Biotech

• Australian researchers have engineered nanoparticles that both detect and shrink arterial plaque, cutting unstable deposits by up to 52% in mouse models. Coated with a peptide that locates inflammatory immune cells inside artery walls, the particles act as a combination therapy and diagnostic tool, “lighting up plaques for imaging while also helping to shrink them.” Because they naturally fluoresce and can be tracked with PET scans, doctors could one day use a single platform to spot dangerous plaques early, treat them, and monitor results in real time. (Future Timeline)

• The Center for Pediatric CRISPR Cures has just launched in San Francisco. Founded by pediatrician and wife of Mark Zuckerberg Priscilla Chan, and CRISPR pioneer Jennifer Doudna, it will create personalized gene edits for children with rare diseases. The model builds on the case of Baby KJ—successfully cured of a fatal metabolic disorder with a personalized CRISPR therapy—showing such treatments are possible. By standardizing and scaling this process, the center aims to cut costs, speed development, and open the door to custom cures for thousands of debilitating and often fatal genetic conditions that today remain untreatable. There are many good uses for CRISPR technology, but I think this might just be the best yet.* (Nature) (Not Boring)

• The Arc Institute has again pushed the frontiers of biotechnology forward, creating the first AI-generated synthetic bacteriophages (viruses that prey exclusively on bacteria)—using their Evo genome language model to design and build functional bacteriophage genomes from scratch. They produced 16 novel phages, some with hundreds of unique mutations and even protein combinations never seen in nature. Several designs overcame bacterial resistance that defeated natural phages, showing AI can generate biological diversity far beyond what is present in nature. The breakthrough suggests phage therapies could be custom built to outpace antibiotic resistance. (Arc Institute) (Arc Institute 2)

• Scientists at Leipzig University have identified a little studied receptor called GPR133 as a key regulator of bone strength. By activating it with a newly created compound called AP503, they boosted bone density in mice and even reversed osteoporosis like conditions. The receptor works by stimulating bone forming cells and suppressing bone resorbing ones, effectively mimicking the body’s natural response to mechanical stress. Because AP503 also strengthens muscle, the discovery hints at a broad anti aging therapy that could protect both bone and muscle health—pointing to a potential one time or long lasting treatment for osteoporosis and age related frailty. This sounds like it could become a vital part of anti aging cocktails that I hope will start rolling out early in the next decade. (Science daily)

• After centuries (millennia?) of snake oil salesmen, there’s now an anti-aging skin cream that actually works. “A 28-day trial showed that pterostilbene cream improved wrinkles, skin firmness, collagen, and pore size far better than a control emulsion. The results highlight pterostilbene as a promising natural ingredient for next-generation anti-aging skincare.” My understanding is that these skin care products with pterostilbene already exist, so ladies (and any maybe some gentlemen), have at it. (Science Daily)

• By supercharging the brain’s mitochondria—the tiny power plants inside cells—dementia like memory loss has been reversed in mice. Researchers were able to flip mitochondrial activity on and off with a drug, restoring memory and motor function in animals engineered with neurodegenerative symptoms. “The researchers also want to take a closer look at exactly how supercharging mitochondria might work over longer periods of time, which will be crucial in figuring out whether or not this is a genuine option for treating diseases such as Alzheimer’s.” (Science Alert – dementia reversal)

• A new 256-element ultrasound helmet that can non invasively target deep brain tissues with a precision up to 1,000 times sharper than conventional methods has been built over in Britain. In early tests with volunteers, the device stimulated the lateral geniculate nucleus—a tiny part of the thalamus—and fMRI scans confirmed real time changes in the visual cortex lasting up to 40 minutes. Unlike existing approaches, it requires no surgery or skull screws, yet allows reversible, repeatable modulation of brain activity. The breakthrough could reshape treatment of disorders like Parkinson’s, depression, and tremors while opening up new doors for basic neuroscience. (Science Alert)

• Researchers at Stanford have pinpointed the reticular thalamic nucleus—a sensory “gatekeeper” between the thalamus and cortex—as a driver of autism like behaviors in mice. In animals engineered to model autism, this brain region was hyperactive, leading to seizures, sensory hypersensitivity, repetitive behaviors, and reduced social interaction. Suppressing the overactivity with either seizure drugs or neuromodulation (could use that ultrasound helmet!) reversed symptoms. Because the same drugs are under study for epilepsy, the findings highlight overlapping circuits between autism and epilepsy and suggest a new therapeutic target for both. (Science Daily)

• A study in mice suggests that beta-glucan, a fiber found in oats and barley, may promote weight loss in a way similar to Ozempic. Animals fed beta-glucan on a high fat diet lost body weight and fat without eating less, showing increased energy expenditure and higher levels of butyrate—a gut metabolite that stimulates GLP-1, the same hormone mimicked by semaglutide drugs. Other fibers tested, like cellulose or resistant starch, had no such effect. While more research is needed in humans to determine effectiveness, it’s perfectly safe so I expect a lot of people will just give it a go themselves. (Science Alert)

• Harvard researchers have built an AI model called PDGrapher that pinpoints the genes and pathways most likely to reverse diseased cells back to healthy ones. Unlike traditional drug discovery, which tests single targets in isolation, PDGrapher maps the complex interplay between multiple genes and predicts which drug combinations can restore normal function. In tests across 19 datasets and 11 cancer types, it not only rediscovered known targets but flagged new ones, with accuracy up to 35% higher and speed up to 25× faster than other models. By focusing on cellular drivers of disease, the open source tool could accelerate precision therapies for cancer, neurodegeneration, and other complex conditions. (Dr Singularity) (News-Medical)

• Scientists at the University of Zurich have shown that neural stem cell transplants can reverse stroke damage in mice—regenerating neurons, repairing blood vessels, calming inflammation, and even restoring motor function. The transplanted cells (‘derived from induced pluripotent stem cells, which in turn can be manufactured from normal human somatic cells’) not only survived but integrated with existing brain circuits, communicating with host neurons. Remarkably, waiting a week after stroke before treatment worked best, offering a realistic clinical window. The researchers hope to start human trials soon. Given the high prevenance of stokes, that can’t come soon enough. (Science Daily)

• There’s been a surprisingly simple breakthrough in the fight to treat fatty liver disease—a problem for up to a third of the world’s adult population. The discovery was made in South Korea, where scientists identified microRNA-93 as a genetic driver of fatty liver disease—and found that vitamin B3 (niacin) can suppress it. Doing so restored healthy lipid metabolism, cut fat buildup, and improved liver function. (UNIST News Center)

• Neuralink is expanding further outside the States, with its first two Canadian patients receiving their implants earlier this month. Everyone currently in the Neuralink program is a paraplegic, with the implant allowing them to control computers and other electronics with their minds—given them increased freedom and a type of mobility. Next year, the company will aim to “restore (limited) sight to the completely blind,” according to Elon Musk. Keep in mind that this is as “bad” as this technology will ever be. It will just keep getting better from here. A decade from now, where might we be? (Neuralink) (Elon Musk)

Agriculture

• Scientists at Cold Spring Harbor Laboratory have mapped the genetic “master switches” that drive plant stem cell growth, using single cell RNA sequencing to chart regulators in maize and Arabidopsis (thale cress, a common model organism used in labs). [Side bar: This process separates each cell, “convert its RNA into DNA, and label it with a tag that identifies which cell it came from.” This allows scientists to determine which genes are expressed over multiple cells] They uncovered hundreds of rare genes tied to shoot development, some linked directly to crop size and yield. By creating an atlas of stem cell activity, the team hopes breeders can more easily engineer resilient, high yield plants for food, feed, and biofuels—laying a foundation for fine genetic editing control over plant growth. (Science Daily)

• On Nishinoshima, a volcanic island south of Tokyo that’s repeatedly wiped clean by eruptions, researchers sequenced the DNA of an extinct purslane (a common, edible succulent plant that grows as a common weed worldwide) population and traced it t nearby island of Chichijima. The plants showed strong founder’s effects and genetic drift—quirks caused by only a few seeds making landfall after volcanic eruptions had previously scoured the island clean. The work offers a rare look at how ecosystems reassemble from scratch, revealing the genetic bottlenecks and chance events that shape life’s earliest footholds in a new environment. (Science Daily)

Weird & Wonderful

• Scientists have pinpointed two genetic changes that helped transform wild horses into cooperative companions about 4,500–5,000 years ago. By analyzing DNA from thousands of ancient horse remains, mutations linked to reduced aggression and improved trainability were found. These shifts likely explain why horses became suitable for riding, farming, and warfare. The study highlights how small but crucial changes in behavior related genes reshaped human history, turning horses into vital allies that carried civilizations forward. An interesting tangent, I assume zebra’s do not have these, given their famous resistance to domestication. Perhaps they could be engineered in? Someone should really compile the genetic basis of domestication in all the species that we have managed to domesticate. Then it might be as easy as a little genetic editing, and bam, a newly domesticated species in a single generation. (Interesting Engineering)

• This year’s Ig Nobel Prize winners have been announced! FYI, the Ig Nobel Prizes are a parody of the Nobel Prizes, but the science is actually real. The goal is to “honor achievements that first make people laugh, and then make them think.” For biology, cows were painted with zebra stripes to see if they were bit less by flies. The answer? Yes, it worked beautifully, with the cows being bothered far less. In chemistry, the experimenters tested “whether eating Teflon…is a good way to increase food volume and hence satiety without increasing calorie content.” Going to be honest, not sure I’d try that one myself. On a much tastier note, the physics prize was awarded to a team that discovered a foolproof way to make the sauce for a pasta dish called ‘Pasta alla cacio e pepe.’ In aviation, bats were given alcohol to see if it impaired their ability to fly and echolocate. Unsurprisingly, it did. Go check out the rest, they’re all pretty funny, and a little thought provoking too. (Ars Technica)

• A new study of wild octopuses captured nearly 7,000 arm movements across multiple habitats, revealing a clear division of labor: front arms probe and explore, while back arms drive locomotion. Each arm can bend, twist, shorten, or elongate—with the animals often combining multiple actions simultaneously—showcasing their extraordinary motor control. Different regions of the arm even specialize in distinct movements, with bends concentrated near the tips and elongations closer to the body. These insights into octopus behaviour and physiology not only highlight their intelligence but could also inspire advances in robotics and neuroscience. (Science Daily)

Photos & Videos

This picture is “the most detailed view of the human brain to date.” Only about 3mm (1.22 in) wide, it took 1.4 petabytes of electron microscopy data to produce. (Science via Nicholas Fabiano)

From a recent Falcon 9 launch, doesn’t this deployment of Starlink satellites look straight out of a science fiction movie? (SpaceX)

Speaking of sci-fi images, check out this one from the most recent Starship launch. Incredible that we get to see things like this. (Elon Musk, SpaceX)

A Falcon 9 rocket launch transiting the face of the Sun. One of Andrew McCarthy’s best yet I think. (Andrew McCarthy)

The Event Horizon Telescope released a series of photographs showing the changes in the supermassive black hole M87* over the course of 4 years. (Event Horizon Telescope)

Recommendations & Reviews

This time, I’m going to recommend a mini documentary marking the 10^th anniversary of the first detection of gravitational waves. On September 14^th, 2015 a signal from merging black holes 1.3 billion light years away reached the LIGO (Laser Interferometer Gravitational-Wave Observatory) detectors, swaying them by less than the diameter of a proton. It was enough though, and from that moment we’ve begun to hear the gravitational “whispers in the cosmos that had gone unheard until that moment.” (LIGO)

That’s it for today, hope you all enjoyed it. Techno-Optimist will be back in your inboxes on October 25^th. Thank you all for reading — and until next time, keep your eyes on the horizon.

-Owen

Subscribe now