💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation are being harnessed to nurture and improve urban life. If you’re new here, try this 90 sec. explainer video.

⚡️ More on Milan + “Community Energy Futures”

For one week every spring, the city of Milan explodes into a spectacle of design and technology. Starting as a massive furniture fair, Salone del Mobile, this annual event has spilled out into galleries, pop-ups, installations, and public spaces all over the city. Fuorisalone, “outside the fair,” has become a global event drawing nearly 500,000 visitors annually into Milan’s design-chic urban kaleidoscope. This is Milan Design Week.

This year, students from the Urban Technology program at the University of Michigan were there to participate. Third-year students Jack Bernard, Pranav Boopalam, and Elijah Stowell presented their project “Community Energy Futures” at UNFOLD, a student design exhibition hosted at BASE Milano. They also delivered a presentation of this work to designers, professors, and students from around the world during the day-long UNFOLD conference on April 23, 2026.

Their project was accepted through a competitive review, organized by the renowned design school Domus Academy. UNFOLD brings together students, faculty, and designers to share the stage through the exhibition, conference, and faculty symposium. UNFOLD is now in its third year.

The theme for this year’s event was “Engage: Friction,” encouraging designers and students to move beyond buzzwords like “seamless” and “solution” to see the generative potential of friction. Twenty leading design schools from around the world were selected to participate.

The conference was hosted at Domus Academy’s campus in the Navigli neighborhood of Milan. The exhibition was presented at BASE Milano in the Zona Tortona Design District. BASE Milano is a multi-function venue that transitions from an exhibition hall by day to an intense EDM-saturated social scene by night: food vendors, rooftop spaces, pop-up bars, DJs, and a LOT of bass.

Bernard, Boopalam, and Stowell shared their project, which uses Interaction Design and UX/UI methods to present multiple, sometimes competing, scenarios that renewable energy microgrid systems enable. Inspired and directly informed by the city of Ann Arbor’s pursuit of a Sustainable Energy Utility (SEU), the project focuses on unanswered questions on the “social” side of this proposed socio-technical system. Ann Arbor’s SEU envisions a community-funded, opt-in system of networked 100% renewable energy resources, such as solar, wind, and battery storage systems. Sounds great! But, how might this system present new social frictions as neighbors begin to negotiate energy-related decisions together? Can these systems, including their interfaces, be designed to support more neighborly and less competitive models of collaboration? These are the questions the project “Community Energy Futures” engages.

Elijah said it was “such an honor and joy to be able to participate in the Unfold conference and to represent UT at a global scale. Seriously, this trip was potent with lessons and memories I will carry with me for the rest of my life.” Jack found inspiration in “seeing the large variety of global design perspectives, and gaining better visibility into how other parts of the world approach design challenges, and what factors shape their outputs. This international context helped me develop a stronger individual positionality as a designer, student, and global citizen.”

The project was accepted based on the screen-based UX/UI design, but exhibiting the work required a physical experience. With support from Taubman College’s SEED Fund, the initial project team expanded to include two M.Arch students to help conceptualize and build a physical model to present the scenarios depicted by the interface. Shaguun Patel and Naman Desai were amazing collaborators, joining what just may be the first joint M. Arch-meets-Urban Technology project to date.

Jack said that collaborating with Patel and Desai on the physical model “has become a highlight of my time at Taubman so far; what an awesome opportunity to learn, build, and create with two amazingly innovative and ambitious people.” He added, “It’s been awesome to be a part of this convergence of skills, ideas, and resources from across the whole college, bridging gaps across programs and majors to come together and produce a stunning final product.”

Elijah agreed, saying the collaboration “really challenged and bettered both sides’ communication skills to be able to find a common language and to trust the other’s decision-making in their expertise.”

The Urban Technology students were among only a small number of undergraduate projects represented. Most were produced by students in graduate programs. Given the multi-disciplinary nature of the Urban Technology degree, these students may not even consider themselves “designers.” Regardless, I’ve assured them that this achievement gives them major design “cred.”

For Jack, his “personal highlight of traveling to Milan Design Week this year was getting to meet and interact with other students from different institutions all over the world. From talking about their work and studies to exploring the city and dancing together, the experience reminded me that despite artificial borders and geographic differences, we’re all connected as humans and as designers. From my perspective, the US can be very isolating at times given its size and independence, but now more than ever it’s important to embrace unity and show solidarity both domestically and internationally.”

Of course, the city of Milan was also a major attraction in this whole experience, full of its own wonders and surprises. “I would say that the biggest surprise was the realization that Milan must be taken and appreciated on Milan’s time,” said Elijah. “By that I mean bulldozing through all of its highlights like a true American in a day will leave you with nothing but sore feet. Take Milan slowly, appreciate and match its relaxed pace to reveal its beauty.”

And if you’re bulldozing, you get thirsty. “I was very surprised at the lack of free/public water in the city,” Jack remarked. “I felt like I brought my reusable water bottle for nothing, as the only way to get water while out and about (or at a restaurant or at the hostel) was to buy a plastic bottle of water for a few euros. I guess most people in Italy just hydrate off of Aperol Spritz...”

On a personal note: This project began in an Urban Technology studio I teach called UT330: Interaction Design and Urban Experiences. Seeing work from this class presented so brilliantly by the students on such a global stage was truly a delight. I joined Taubman College of Architecture and Urban Planning in 2023, and this was the very first design studio I developed and taught for the Urban Technology program. Over the years, I have often remarked that it usually takes at least three tries to get a class right: the learning objectives and projects are dialed in, the projects are doing what they’re supposed to do, and the class yields the results you’re looking for. You can imagine my immense pride that the third edition of this class yielded a project that ended up at Milan Design Week!

Something struck me as I was engrossed in this bacchanalia of design, technology, and spectacle. At some point, I thought of the world’s fairs of the 20th century. World’s fairs saw nation-states put their best minds to work crafting dramatic experiences that showcased competing visions for new ways of living and working, often also inspired by new technologies. In the 21^st century, design weeks, among other events, have taken on some of that role. Rather than nation states, we see designers, architects, technologists, artists, and provocateurs—many of whom are backed by global corporations—providing spectacular and often wildly divergent views of how life today and in the future might differ in so many grand and banal ways. As one presenter at UNFOLD remarked: All societies need their spectacles!

Jack summarized the impact of his experience: “Participating in the 2026 UNFOLD Conference at Domus Academy in Milan was such an exciting and unique opportunity. The experience helped to solidify so many concepts and ideas brought forth in our studio courses back home, and see firsthand how design can play a role in shaping the world around us.”

These weeks: The building is quiet. Students have disappeared for summer adventures, but staff keep moving. UT++ workshop in Detroit. Exit surveys and budget planning. Let the summer projects commence!

Recently I’ve been thinking again about the space between technology and urbanism, and again causing some unintended feather rankling. As I work through this, I’ve returned to the basic attitudes, assumptions, and practices that give shape to the gap between an urbanist’s conceptualization of the world and a technologist’s. Both of these perspectives live inside me, so on a regular basis I am navigating my own thoughts about what’s right as an educator, a practitioner, and a city-dweller. It’s not straightforward.

On occasion of this being the 300th issue of our newsletter, and last issue we will send out before going on summer break, I’m revisiting the table and exploring the implications of each row in more detail.

BTW - if you like this newsletter please do us a favor and share it with someone who would also enjoy it. We’re just shy of 1500 subscribers!

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation are being harnessed to nurture and improve urban life. If you’re new here, try this 90 sec. explainer video.

⚔️ Unpacking Technology vs. Urbanism

Here’s the table mentioned above (from 2021), after which I will address each line item (from 2026):

Theory of Change: Revolution vs. Evolution

This is at the top of the list because it’s the mother of all differences between tech and urbanist perspectives. Clayton Christensen’s concept of disruptive innovation, widely internalized by Silicon Valley, valorizes small firms that topple incumbents. Think Netflix unseating Blockbuster, Japanese automakers challenging the Detroit Three, or Uber (and ride hail as a category) taking ground from taxi franchises.

In each case, permission was not granted and the upstart entered without playing by the same rules as the incumbents. Ultimately when a disruptive contender wins, it’s because the market chose them. Consumers could pay for a taxi ride one day, then pay for an Uber ride the next day, note for themselves the pros/cons of each, and make a decision about which they prefer. In the urban (physical) realm this kind of shopping is not as easy to come by.

Big changes to the city are often cast in literal concrete, costing millions or billions, and become commitments with timelines counted by decades not minutes. Decisions about the urban realm are slow, hard to undo, costly, and almost always mutually exclusive to other possibility. Your favorite park is either your favorite park or it’s a parking lot, or housing, or something else. It usually can’t be all of those at the same time. The choice architecture for a resident evaluating possibilities has different consequences when those choices are urban.

Compared to the taxi example where you can shift back and forth between taxi and ride hail with relative ease, you cannot say yes to the park→parking on Monday and then get your park back on Tuesday. For this reason, incremental changes are rationally safer bets in urban development versus disruptive change. But disruptive changes in cities do happen and can be awesome. Pedestrianizing Broadway in NYC, building Central Park, and removing the Embarcadero Highway in SF, are just a few highlights worth naming.

Financial Risk Tolerance: High vs. Low

The structure of the professions that deliver the built environment, and the work that they do at nearly all levels, is about minimizing risk. No one wants a big urban development to fail, nor to end up being wildly more expensive than planned. All sorts of mechanisms from fixed price contracts to as-of-right zoning are tools to minimize the inherent risks of urban development. Risk tolerance is low because the basic proposition of making something from scratch is so incredibly expensive in the first place. As a result of this low risk approach, returns are also relatively modest. A good financial outcome for a real estate project may be a 2x return on a 5-7 year timeline.

Building technology is also pretty expensive, but the potential investment returns when you get it right are in a different ballpark: 10-100x. Venture capital fuels bets with the promise that many will fail but enough will have a big exit that the portfolio pencils out favorably. Whereas the urban development world is about minimizing risk at every turn, technology is more aligned around finding the right risks to take and leaning into them. Both are rational choices in the context of the capital allocation logics in their domains.

Growth Model: Exponential vs. Linear

This refers to the dominant business models in both camps. Technology businesses are product-based. The proposition is to build a product once and sell it to as many customers as possible. Indeed, the growth model is tied to the financial risk tolerance line above. The reason venture capitalists bother with tech at all is because of the potential to build once and sell many, many, many times. On the other hand, built environment work is almost entirely consulting based, which generally scales linearly with the headcount of the firm. To serve more customers, you need more consultants. This is about the businesses, primarily, but also applies to the outputs of those businesses. It’s hard to speed up urban development and productivity across the built environment has been flat for decades; meanwhile productivity in tech has been growing and is now (maybe?) growing even more thanks to AI.

Risk Mitigation: Prototyping vs. Planning

How do different domains handle executional mistakes? The difference comes to life if you compare the process of launching a digital product versus a physical place.

Let’s say you’re Instagram and you are working on a new feature. You can build it internally and fully test that feature with a closed audience. Then you can launch it for users and conduct A/B testing to see how people react to multiple versions of this new feature, before finally committing to the version that is launched to all users. A loop of build-test-refine lessens the need to make guesses about what will work best, and it enables quicker recovery from misses (which come for even the best designers and engineers).

In the built environment, prototyping is almost never full scale. Models can help the architect test daylighting on a proposed building. Simulations can help the urban designer understand potential impacts to street traffic. Ultimately the way we prototype the built environment is by building it at full scale and living with it. 99.9999992% of everything in every city on earth is a prototype in that it has never been built before just like that, in exactly that location. Learning begins immediately, but the cost of making changes is extremely high and slow compared to the speed of iteration in the digital world.

So, in lieu of the ability to build and test, urbanists plan carefully. This involves circulating planning documents (written or in drawing form) to professional stakeholders for their review and input, and later a similar process with the public to collect input and refine the plan. The process is a “waterfall” where one stakeholder makes the plan good to the best of their ability before passing it on to the next party to do the same, and the handoffs between parties stack up as “interface costs” that make this a slower and more brittle process than the agile approach that’s now common in tech.

Scale: Quantitative vs. Qualitative

In one way, digital and physical react to scale similarly, and that is when it comes to the volume of users. In digital spaces as in physical ones, more users often results in more opportunity (via network effects), but also requires more negotiation of conflict. In this instance, quantity drives qualitative differences.

In a more simplistic sense, the domains of urbanism and technology generally talk about scale from different points of view, however. The tech community talks about scale as growth. Scaling up is about quantitative increases of headcount, compute, bandwidth, and other resources. From a business perspective, scale is something to be achieved (because it implies winning), and from a technical perspective scale is something to be managed (because it’s more difficult).

For urbanists, scale is more likely to be invoked as a question of appropriateness, and usually refers to the relative size of things. One might ask, is the scale of this fountain appropriate to the plaza that it sits within? Does the new extension have an appropriate scale next to the original structure? Is the bar outlandishly sized for this restaurant?

Legal Perspective: Ask for Forgiveness vs. Ask for Permission

The built environment is highly regulated through zoning ordinances, building codes, deed covenants, homeowner agreements, and other legal mechanisms. There are formal means to submit plans and ask for clarifications about what’s allowed, or to request permission to do something outside of the norms. Cities often entertain this, for instance, in the form of developments being allowed more square footage if they provide generous public spaces back to the city. There have not been equivalent review procedures for tech products. Uber did not ask for permission to operate a taxi service because it was designed to not be a taxi service, per the letter of the law. Existing in the legal gray area was a way for it to exist, and eventually to grow such that it was harder to argue with. This is not merely an attitudinal factor, but relies as well on the capital needed to weather legal battles that may arise from the strategy of asking for forgiveness.

Humans: Users vs. Publics

Successful tech products are organized around a persona or customer type. Identifying a segment that you can speak to with your product (the target demographic) is how you win as a business because you are able to focus on their unique needs and how to market to them as best as possible. This ability to focus is a gift, especially to new businesses because it means you don’t need to be all things to all people (though some businesses attempt to become that later in their life—ahem Amazon).

In the urban realm it’s far harder and less common to segment users. For something like housing it’s possible (e.g. “dual income couple with a growing family and are drawn to modern farmhouse aesthetic”), but for urban design, urban planning, and landscape architecture there’s a presumption of serving the public as whole or—at the very least—a variety of user types at the same time, as in a campus situation.

Design Perspective: Human Centered vs. Systems Centered

Tech has adopted a human-centered approach to understanding what users want, designing for their needs/desires, and optimizing to the hilt. When such techniques are applied to making it easier to pay taxes or get a passport, that’s a good thing. When they’re applied to something like JUUL vapes, it’s not so good! In either case, user research goes deep into the personas of your users, including their behaviors, constraints, and motivations.

The design process for architects and other urbanists is less likely to focus on users with this level of specificity, and it’s more common to think about broader categories such as the public, students, and staff.

Of all the line items being considered here, this is one where there’s lowest hanging opportunity for bridging the gap. For reasons I don’t fully understand, human centered design methods never really landed in architecture and planning schools, but the methodologies of user research are pretty readily adaptable.

Commons: For Utilization vs. For Protection

One aspect of this is that distinction between digital commons and physical ones. Digital commons are non-rivalrous: an infinity of people can use the same open source software at the same time without creating any downsides. Streets are rivalrous: if your sidewalk is filled with hotdog vendors, then there’s less room for grannies on benches (or trees, or school kids walking home, or what have you).

Another aspect is about the stance toward physical commons. An urbanist perspective on the design of a street is usually about weighing different types of users and trying to find a fair balance. A tech perspective is more likely to see space for everyone as an opportunity to use it in ways that are beneficial to their business. The most hilarious (and kinda sad) example I think of here is Zume, who operated a pizza-shop-in-a-truck model that would pull up to a parking spot on the street outside your house, cook the pizza, and then give it to you. Technically this is not too dissimilar from a food delivery car parking temporarily while dashing food to someone’s door, but this is an example where the quantitative difference in the number of total minutes spent idling feels qualitatively different and, from my humble POV, more exploitative of the right of way.

What does it mean?

It’s not just two domains of knowledge that are being brought together here, but disparate value systems as well. For our students in urban technology, the space between these two columns of caricature is their territory of operation. What makes me most optimistic about this is the fact that none of the rows on this table are static. For instance, when I started my career in tech, we were still using waterfall development to build digital products. Now agile and lean methodologies have radically changed the way tech development works.

Whether it’s building enough housing to make homes actually affordable, dealing with climate recovery and migration, or finding ways to combat the isolation that comes as a result of sprawl, there are a lot of reasons that the work of city-making feels extra urgent today. And yet, it also feels stuck—and wickedly difficult to unstick.

Tech has been wildly productive in my adult lifetime, so maybe we can borrow those practices to unstick city-making? The possibility of urban development practices learning from tech practices is enticing, but urbanism isn’t going to readily “be more like tech” for all of the reasons listed above. For our undergraduates in urban technology, we ask them to look beyond the unresolved conflicts described by this table and to imagine how both perspectives could meld in new and interesting ways. That’s the privilege of a transdisciplinary degree program. Because of this, the educator side of me is excited!

But as a professor of practice, the professional half of me finds the moment to be awesome in the older sense of the word, mixing wonder and dread in equal measure. There are vast amounts of work to remake the built environment to produce better health, sustainability, and civic outcomes, but it feels like traditional ways of working just get harder and harder.

My sense is that the experimentation needed to undo this stuckness will come from a focus on how the work of the built environment is done, not just what gets built. If cities need radically better outputs, they're going to need quite different inputs. That's why I returned to this table: it's a map of where experimentation is needed in how places and spaces are imagined, made, and maintained.

These weeks: Commencement this weekend! Our second graduating cohort. 🏃

For Urban Technology, the importance of service design is that it provides a clear and coherent set of tools to own the user experience across digital and physical interactions. Having an electric vehicle is one of my favorite examples in this realm. The satisfaction of an EV owner is not determined by the car itself, the phone app used to manage the car, or the network of private and public charging options in your city. It’s the result of all of those and how well they work together.

This is why service design is a required part of the UT curriculum, recently taught by Alex Johnston and Ron Bronson, and also why we’re excited to welcome our new colleague Sheng-Hung Lee who brings additional service design chops to the team. Interview below.

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation are being harnessed to nurture and improve urban life. If you’re new here, try this 90 sec. explainer video.

👴 Interview with Sheng-Hung Lee

Sheng-Hung Lee comes to Taubman having recently completed his PhD in service and system innovation at MIT. Upon arrival, he established the d-mix lab to build out a research program on Design for Longevity, which is focused on making services, systems, and cities age well alongside the people who use them. In addition to the academic credentials, he brings consulting experience from design firms IDEO and Continuum. Sheng Hung’s spirit is as generous as his emailed are filled with well-curated emoji.

Bryan: Where did you first get comfortable working in a multidisciplinary setting?

Sheng-Hung: When I worked in industry. I never worked with the same people on different projects more than once. Every project was new… a kind of recurring startup situation as the projects were so different: How do you design a baby diaper? Redesign an organizational culture? Plan a curriculum design? Or create a future kitchen hood? It was pretty broad and very different.

I have to say I didn’t feel comfortable and confident. It felt more like I was in a safe environment where I could make mistakes. That really gave me a chance to think out loud. It’s still uncomfortable because of all the uncertainty that comes with doing new things, but I could feel secure that nobody would blame me. I could make mistakes, and the more mistakes, the better.

The first class I taught here at Taubman was UT 210 Listening: Design Ethnography Methods, and a student asked me: what does urban technology mean? I said that we’re on the same page exploring this together. I do feel like it’s really aligned with the current way we think about design. It’s not just discipline-driven challenges, it’s multidisciplinary and interdisciplinary. These are complex, systematic challenges, not single issues, and they can not be divided by traditional design discipline.

For example, we are doing a project with General Motors (GM) in this course. The expectation from a company like GM is that they’re trying to design a new EV car or something, but the true question is whether they need to focus on cars or more services. As city designers, we’re not just talking about the individual level — we also think about broader aspects like how the EV charging station can be rethought in relation to social infrastructure (communities). This topic is pretty new for students, for GM, and for myself so we’re exploring it together. I appreciate urban technology as a safe and creative environment to try something novel, something different, and bold

Bryan: Your PhD is in service design, which is still pretty rare in the US. Where did you first encounter service design?

Sheng-Hung: I was trained as an industrial designer and electrical engineer, so everything was super tangible: specs, processes, standards. The first time I became aware of services was when I started my master’s and PhD journey. I started to think that intangible things are really important in our daily lives. To continue with the GM example from before, how do you use an EV in different places? How can you recharge it? Do we have apps? Do we need to redesign apps’ interfaces for various personas? In service design these are called “touchpoints,” and they are where a service is anchored. The ultimate goal is to create better user experiences [that link all these touchpoints]. It sounds very subtle, but it’s that subtleness that accumulates into the overall user experience.

In Europe, service design is well-established and a relatively mature “service industry” with a degree in higher education. They offer master’s and PhD programs, and specific service design journals. In the US or in Asia, I would say it’s still pretty new and people easily get misled: oh, you have sticky notes and Sharpies, so you’re a service designer! But I think these are just tools and a very surface-level understanding. You still need to know how to make things, to know the fundamental design process, and not just talk or think about it, because just talking or producing a service blueprint or drawing a service ecosystem diagram simply changes nothing. You still need the ability to execute ideas and proposals.

Bryan: You implied that service design gets more important when digital and physical meet. Why?

Sheng-Hung: Nowadays, when we say ‘design,’ people think, “let me create an app.” In reality, this risks creating extra apps that don’t really solve the core problem or see the whole picture. As designers, we don’t just create stuff; we also curate experiences. We have to curate all these experiences together. Service design is a great medium to connect the dots. For example, when students want to design the hospital discharge experience, they think about apps, but the truth is our first service interaction in a hospital is with a nurse, with a caregiver in a physical environment.

Bryan: Let’s talk about Longevity, which is another thing you’re passionate about. What do we need to know?

Sheng-Hung: When I say longevity, it’s definitely a new term. I can’t even accurately translate it into Mandarin! When we say “aging,” people think about negatives like mobility issues, fragility, and sickness. When we say “longevity,” it opens up possibilities to thinking about a multi-generational culture, policies, multicultural workforce, and society. It also includes populations across various ages: middle-aged, teenagers, children, and even babies. Longevity isn’t just about creating sustainable products or services. It’s also about designing meaningful, personal, and respectful products, service, and experiences. Longevity also cares about meaning, purpose, and motivation. Because if people are living longer, we also need to talk about the quality of life. Why do you want to live longer?

Bryan: You are everywhere and always have iconic eye glasses so I have to ask, where do you get your glasses?

Sheng-Hung: Every time I travel, I want to visit a local eyeglass shop. The big neon yellowish eyeglasses I wore yesterday are actually from Belgium. There’s a brand called Theo. The color is incredible. I love it!

Bryan: What is your favorite city and why?

Sheng-Hung: I don’t think I have a favorite city. But I would say: my favorite city is the city I want to co-build, co-create, and co-develop with city planners, service designers, system architects, and local communities based on their needs and their vision for the future. At least for me, I’m looking for a place where I can find my community and a place with culture. That’s really, really important to me.

These weeks: Our final UT Studio Talk. Admitted student campus days. A new UT Peer Mentor Program. Spring is here, but so is final project stress and midwest thunderstorms. Commencement on the horizon.

Robert Moses used speed as an advantage. When Jane Jacobs and colleagues successfully defended Washington Square Park from being bisected with a road Moses had proposed, this was the first crack in Moses’s armor. Soon the average velocity of urban works in NYC started to slow way down as his bullish tactics lost their might. Participatory planning rose around that time in the late 1950s-1960s, bringing small-d democracy to urban planning decisions. The churn of those processes creates buy-in at the expense of speed. Good participation takes time, after all.

Or does it?

This week’s issue features an interview with Ariel Noyman, MIT-trained urban scientist who is exploring urban simulation and modeling as decision-making aids. Fundamentally, his work is asking how evidence-based approaches and real-time technologies can empower people to make decisions faster and from a more informed starting point. It’s about “shrinking the gap between prediction and product when we design cities,” as Ariel explains.

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation are being harnessed to nurture and improve urban life. If you’re new here, try this 90 sec. explainer video.

🕵 Ariel Noyman on why Democracy Needs Data-Driven Processes in the Built Environment

As an urban scientist, Ariel Noyman understands urban environments as nouns and verbs—both the qualities of a place that has been made, and the processes from which those places emerge. Planning, design, decision making, and debate. This interest in urban places, fused with technical know how, results in a wide body of work in modeling, simulation, and evidence-based decision making. I met Ariel at Cornell Tech’s Urban Tech Summit last year when he presented Travel Agent, one of many interesting projects at CityScope, and we recently connected by video between Brooklyn and Ann Arbor to discuss his work.

Bryan Boyer: What are you working on?

Ariel Noyman: I’m involved in a slew of work at CityScope under the guise of democratizing data-driven design processes. Democratization of planning processes has existed since the 1950-60s in things like public engagement, participation, civic groups, etc. Data-driven design has also existed for about 6 or 7 decades.

Where we found the white space is really in the convergence of these two. How do you take the complexity of algorithms, models, data, and bring it to the hands of decision makers, planners, and the general public?

The four themes of CityScope are: Insight, which is understanding and visualizing the current state of the city. Transformation, which is about interactive tools for scenarios and design planning. Next is Prediction, which looks at the impact of changes to the built environment, such as moving a building, creating a road, or building a plaza through mostly the subtle aspects. What is the experience that these changes create? Finally, Consensus, which is about how you bring all of that back to the hands of decision makers and planners.

Bryan: You used the word “tool” in the context of democratic decision-making. Are tools what’s needed or is it about process?

Ariel: We are realistic in the way we work. We have academic freedom, but we work with a lot of cities and governments around the world. When you meet the current process, you understand that you cannot just replace it with whatever imaginative, ideal process you came up with from Cambridge. So, in many of the cases where we used our tools, it was to take an existing, complicated, mundane, slow bureaucratic process and try to put some WD-40 in different places around that process so it could be more streamlined.

The ideal consequence would be that the process itself is changing and that you rethink the way we make decisions in our cities. Instead of contracting McKinsey or somebody to do a 5-year research project, you might build a tool that, in real time, shows everything that is currently happening in the city in multiple layers in a tangible platform, and then you can bring all the decision makers around that platform and have a conversation.

We validated in research that these are much faster and much more compelling ways to make decisions. It doesn’t replace the whole process, but it takes one chunk of it and replaces it with something more accessible.

Bryan: One of your projects is Travel Agent, which is a novel use of agent-based modeling in an urban planning context. What do we need to know?

Ariel: Traditionally, whenever you build a complicated building—an airport, a hospital, a public plaza—many design agencies require you to run a simulation: agent-based modeling, where you run computational agents 10, or 100, or 1,000 times to see how they would react to the changes you propose for that site. But we never were able to use these tools to really understand what people feel as they move through the building. That’s because these agents were pre-computed, pre-composed. You give it a set of rules: go right, or left, or forward. If you meet X, you do Y. They’re not people. These are computational entities that run in an orchestrated world, in a deterministic way.

Agents in this more traditional style have the origin and destination for their trip and it is very orchestrated. It’s useful, but such an approach doesn’t tell you much about what people experience. And we know that what’s really important is how we eventually experience a building or the place we move through, and that experience is different depending on your demographic. If you are a 16-year-old girl walking at night in a public plaza, your perception of that space is not the same as [a grown man’s]. If you’re using a cane and your vision is impaired, you have a different experience.

In 2022–2023, with the emergence of GPTs and LLMs, we started seeing experiments around the idea of humanizing computational agents. People got excited about agents that can suddenly become more humane—more sophisticated. This is where the Travel Agent idea fell into place. Someone was creating an LLM-based, text-based city and behavior, and I said, What if we add the actual city to it? What if we add the physical, three-dimensional environment to those agents, and then let them roam through an actual city? And even better, let them roam through a designed city that doesn’t yet exist, that we want to measure and evaluate?

We started by letting the agents run through Google Street View with very open-ended tasks… not an origin-destination but a goal such as, “look for a shaded place to have lunch.” Would they go astray? Would they find a path? And more importantly, can they explain their decision-making process at each step? If you walked 50 feet forward, why? If you turned, why? If you got lost, why?

Bryan: What are some examples of unexpected behaviors that you found useful or surprising?

Ariel: We tested whether the agent could handle multiple sequential tasks, and one of those tasks was: “You got to the train, but unfortunately the train is not working today. Find another means of transportation. While you’re waiting, look for a coffee shop.” We hadn’t set up anything that resembles a coffee shop in the virtual test environment and we wanted to see what they do as they look around. How do they search? The most fascinating thing was that the agent started recognizing the form of the city and reported things like, “This looks like a plaza. I see a tall building that looks like an office. I see the train station. Probably there’s a coffee shop down there.”

Bryan: How do you see the different disciplines of the built environment evolving as computational approaches become the norm, instead of something that happens in an ivory tower lab?

Ariel: We’ve seen a lot of effort over the past decade to use technology to improve the early stages of design. All the big companies, including tech companies not associated with the built environment, were investing in early-stage technologies. And I’m frustrated by that. We teach our students that this is an important step, and indeed 90% of the decisions are made in the first few months of a project, but that’s just the first few months. You then have, in many places, 6 to 8 years until you see the building built. What about that portion of time that is completely unhandled right now?

I’ve been looking at environmental review reports here in New York City. You print one page after the other of the shadow analysis of a building, maybe 450 pages total, and you submit that as an environmental review. The gap between this and what we do with our technologies in the very early stage is tremendous.

My real interest is not so much how we break silos in the early design stages, it’s really how we break these silos of the decision making—of the design process—as it moves forward. There is so much opportunity there, especially now that we’re facing a huge housing and development crisis, and everybody talks about how the process itself doesn’t get them to the goal line.

Bryan: Ten years from now, what do you think urban planners are going to not be doing? What will be made obsolete through technological change?

Ariel: Frankly, I don’t know if we’re going to make anything in the process itself obsolete. Some tools will become better, some techniques faster. But the main thing we can bring to the table is shrinking this whole process into a much shorter and a much more agile decision-making process. That, to me, would be our greatest contribution to planning as it exists today.

In an ideal world we can diminish the 14,000 pages of zoning in New York City, but in reality, that’s not going anywhere. The question is: can we create systems, tools, platforms, and eventually processes that take that same thing and make it legible? Then you don’t need to spend 6 to 8 months just to understand the code, two to three years to make environmental review, three to four years to get a permit, then build for another year in the winter. Let’s shrink this whole thing.

It’s not just about saving money for the developer or the design team. It’s about the fact that we make predictions based on what we know today. If you make a prediction today and your building is there only 6-8 years from now, your prediction is invalid. You predicted a future that doesn’t exist anymore! The elevator, the car, autonomous vehicles: these technologies come into life faster than buildings. In that sense, speed is not just a cost-saving parameter but it is really about shrinking the gap between prediction and product when we design cities.

Bryan: What is your favorite city, and why?

Ariel: Tel Aviv. I’m a bit biased, and I have this sine wave of affection and disdain about it. But the more I work in and around cities, the more I understand that my history as an architect is less important: it’s not about beauty, it’s about function and vibrancy. And there are very few places as vibrant as Tel Aviv. It’s crumbling, it’s messy, it’s smelly—all the things you can imagine of a very highly active city—but eventually it works. It works perfectly 24-7. And I think that’s something we need to aspire to as designers.

These weeks: In the beginning the calendar was empty, nothing but possibility. Next it was a breeze block sort of configuration, structured but with regular gaps. Today it’s a solid wall. Bricks of time stacked on top of each other and waterteabathroombreakstretchstretch as the mortar in between. We’re doing a lot, but there’s lots to do. Every time it seems like UT++ has reached peak effort, a new peak emerges. Also: Town hall with students, coffee hour hosted by the student org, prepping for campus visit days. All smiles.

We still have not formalized anything such as an oath, but if I had to start it right now I would begin with the idea that we must conceptualize of urban technology as always having a third user: there’s the operator of the system (whatever that may be), the direct user, and the public at large. In Wonyoung’s work this would be the landlord, the rental tenant, and society broadly: how does a landlord protect their property and the stability of their business, how does the tenant secure access to housing, and how does the public at large feel confident that there’s a fair way for everyone to find housing?

If we take that framing, then any transaction or decision that’s enabled by a piece of tech cannot be deemed “good” until the outcomes are beneficial for all three parties. If city-building may need to speed up, as I’ve written about previously, the making of algorithms also needs to slow down and make space for better deliberation. Finding the balance is about understanding the limits of technology like we explored with Xiaofan some weeks ago, and Wonyoung’s work lands this question for us in the domain of housing. More below!

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation are being harnessed to nurture and improve urban life. If you’re new here, try this 90 sec. explainer video.

🧑‍🔧 Wonyoung’s Focus on Housing Technology

Wonyoung So and I taught together last semester so I got to see the urban technology program vicariously through his calm and curious eyes. One day he was lecturing on putting values like sustainability and justice into action—what do you concretely do if those are things you care about?—and the next he was helping students hack an API, refine a UX mockup, or clarify an argument. Fun stuff. He recently launched the Reparative Technology Lab to work on inequities embedded in urban technologies, so now we can add determined and ambitious to the list of adjectives that describe Wonyoung.

Bryan: What is reparative technology?

Wonyoung: I like to decompose it, because I want to be really practical about this work. To create reparative technologies, one approach is exposing the mechanisms of technology that create or contribute to the inequalities that we have.

A second approach is to change the outcomes of a given algorithm. The problem is actually the question of what a fair algorithm or algorithmic fairness is. I realized it’s almost impossible to create fair algorithms without addressing the underlying relationship of historical inequalities. The starting point is that we know there’s an inequality that technology actually contributes to. Then we can shift our orientation from creating fair algorithms to creating reparative algorithms, or affirmative algorithms, as some people say. In one of my papers, I identified that treating people explicitly differently is almost the only way to fix the underlying inequality—not just “de-biasing” the algorithms and technologies but by changing outcomes. An example would be giving more down payment assistance to people who are historically discriminated against.

Bryan: Is it fair to say that you can’t just eliminate bias, you have to actively create reparative value?

Wonyoung: I tend to highlight this intentionality a lot in my work. In discourse around fairness, one of the underlying assumptions is that technology is neutral, and by trying to be more neutral and balanced, we can create fair technologies. In my work, I try to counter that by highlighting the intention at the front, and being more proactive.

Bryan: What’s an example of reparative technology?

Wonyoung: One good example is that the state of Washington recently created the Washington Covenant Homeownership Program. They did a study to figure out the historical discrimination—in this case, via restrictive covenants—and then used this argument to legislate a statewide reparative lending program that distributes down payment assistance based on that research.

Bryan: Does reparative technology require a new attitude toward risks?

Wonyoung: I wrote about this in my work on risk-based pricing. One of the fundamental irritations for me, in terms of risk, is that by doing risk assessment work, the consequence is that people who should be most protected end up paying more.

The mortgage lending context is a prime example. Because of the racial wealth gap, people who end up having less down payment due to historical injustices end up paying more in mortgage interest because of this risk assessment. It comes from the different value proposition of those who have resources and opportunities and want to protect their interests. From the lender’s perspective, this assessment makes financial sense. But on the other side, people who are actually assessed end up paying double, which doesn’t make sense for them. They’ve been discriminated against by the system, and then to engage with these contemporary technologies or algorithms, they end up paying more for the risk they have been left with from the past. Technology works beautifully, skeptically speaking, at detaching all of the context from a decision (such as historical inequities) and applying all of the risk onto the individual.

Bryan: What is your theory of change for urban planning? How does it create change?

Wonyoung: Planning is fundamentally thinking about institutional change, coming from institutions and public support. That involves a lot of democracy. When people talk about democracy, their definitions are really different, and my theory of change might differ from others. I almost think democracy is a synonym for equity.

When you think about democracy in planning, especially around taxation, people who have more physical resources such as owning homes, usually have more of a say, according to Fischel. That dynamic is happening at the local and municipal level, and democracy is actually shaped by those who have more resources and opportunities. So my theory of change is that, in order to have institutional-level changes, we ironically need to think about the people who are most marginalized, in order to have a democracy that makes change possible.

It translates into the technology context as well. When you think about democratic AI or democratic technology, what we need to focus on is those who are marginalized by those technologies.

Bryan: That makes me think of the accessibility advocacy from decades ago — if we make the city work for people who are in a wheelchair, it’s going to work great for everybody else too.

Wonyoung: Exactly.

Bryan: What is your favorite city and why?

Wonyoung: I really miss Seoul, which has a really reliable transit ecosystem. There are a lot of bad things about Seoul, but I really like the good things: the river, the mix of nature, the two big mountains in the city.

These weeks: It didn’t fit above but this is a cool project on the cartographers of North Korea. UT++ reached a new milestone. Faculty committees, studio discussions, and rapidly planning some cool stuff for this summer. More soon, as we shift posture from giraffe to anteater.

Having grown up in California and spent lots of time in SF as a dotcom youth, I took the occasion of the trip to write some reflections about the city by way of orientation. An abridged form of those notes below.

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation are being harnessed to nurture and improve urban life. If you’re new here, try this 90 sec. explainer video.

🌁 Dear Students, Welcome to San Francisco

“A desk is a dangerous place from which to view the world.”
—John le Carre

The reason we travel should be clear to each of you, having begun your UT education on a bus. If you study cities, you have no choice but to get out and walk across them; learn their patois; smell them, even. Even if the urban condition is not your central subject of exploration, there’s no substitute for meeting the world where it is. If you must view the world from a desk, make sure it’s someone else’s desk. And that’s exactly what we will do this week, as we visit our slate of thoughtful organizations across the city.

This is not a set of meetings. Do not treat these three days as a checklist. Every minute of every day you are immersed in a context of new shapes and habits. Find them. Observe them with all of your senses. Wonder about them. In the space between our meetings, I invite you to find a reason to love San Francisco—an exercise that’s worth repeating any time you visit a place beyond home.

Calamity, whether seismic or economic, has always been part of San Francisco’s story. A graph of the vitality of this city is as spiky as its own topography. The first big shock came in 1906 when the city suffered a massive earthquake that then let loose days of fires. Fifty-six percent of the city was rendered homeless and 28,000 buildings lost. Just prior to the quake, SF was an ascendant capital in the fledgling country of America, and life was good. Here’s how Market Street looked in that precipice moment. As the camera moves toward the Ferry Building, look at how vibrant the street life is! Just 10 years later, the harsh memory of the quake and fire had been dissolved to such an extent that American author Inez Haynes Irwin, writing in her book Californiacs (published in 1916), offered a facetious warning for the stranger arriving to San Francisco:

“If you ever start for California with the intention of seeing anything of the state, do that before you enter San Francisco. If you must land in San Francisco first, jump into a taxi, pull down the curtain, drive through the city, breaking every speed law, to Third and Townsend, sit in the station until a train—some train, any train—pulls out, and go with it. If in crossing Market street you raise that curtain as much as an inch, believe me, stranger, it’s all off; you’re lost. You’ll never leave San Francisco.”

This gives you a sense of what the place felt like in the decades after the Gold Rush, which is when the urban form of SF exploded into a proper city. At that time, this was a gateway to the Pacific and the nations beyond, as well as the westward terminus of Manifest Destiny. Then, as now, it was hard to ignore San Francisco.

Though most people now arrive by car or plane, for a long time the front door to SF was via the the Ferry Building. It will be part of our story too. One hundred years ago the Ferry Building was how many people arrived to the city. Today the Ferry Building is an active transportation terminal but is more well known as a culinary destination, which is the identity it has claimed since the removal of the Embarcadero Highway that used to so rudely isolate SF from the Bay. Here’s how it felt in 1924, pulled from a travel guide of that vintage:

Stand at the Ferry Building, looking up Market street, and imagine the beginning of the city that spreads before you. First of all you must realize that this point of observation would, in those days, have been offshore, on the shallow water of Yerba Buena Cove. To the right is the scarp of Telegraph Hill, from which ships coming through the Golden Gate were sighted, and to the left is the lesser Rincon Hill, which is being cut away to provide a light manufacturing district. These marked the headlands of the cove, and the waterfront curved inland as far as what is now the site of the Donahue monument to mechanics at Market and Battery streets.

Seeking survivals of the past, you must realize that San Francisco is one of the most modern of the comparatively old American cities. Most of the area that saw its beginning and early history has been wiped clean by fire. The San Francisco of today may be said to date from its rebuilding following 1906, since which time something like a half billion dollars’ worth of new construction has been done. Yet something of early San Francisco remains, either beyond the reach of the devastation of eighteen years ago or in miraculous islands of safety in that sea of fire.

In our own era, Rebecca Solnit is perhaps the sharpest observer of life in contemporary San Francisco. In Wanderlust she writes about coming back to SF with fresh eyes after living in another state for some time:

San Francisco has long been called the most European of American cities, a comment more often made than explained. What I think its speakers mean is that San Francisco, in its scale and its street life, keeps alive the idea of a city as a place of unmediated encounters, while most American cities are becoming more and more like enlarged suburbs, scrupulously controlled and segregated, designed for the non-interactions of motorists shuttling between private places rather than the interactions of pedestrians in public ones. San Francisco has water on three sides and a ridge on the fourth to keep it from sprawling, and several neighborhoods of lively streets. Truly urban density, beautiful buildings, views of the bay and the ocean from the crests of its hills, cafés and bars everywhere, suggest different priorities for space and time than in most American cities, as does the (gentrification threatened) tradition of artists, poets, and social and political radicals making lives about other things than getting and spending.

Once the Embarcadero freeway came down (the result of the last large earthquake, all the way back in 1989), the city could breathe a little differently. Gay culture and counterculture in general still provided the city’s reputation, but just 45 minutes south in the area around Stanford University a new wave of culture was emerging: the personal computing era, followed shortly by the internet era. The combined impact of these movements are what would ultimately define modern San Francisco, and take it from being a regional financial capital as it was in the 1980s, to the global gravity well that it is today.

By the time the world wide web went mainstream in the mid-1990s, there was a bewitching energy combining electronics/computer tinkering; mixed with “information wants to be free“ hackers; mixed with an open source culture powered, in part, by the prevalence of university professors and students in those early days of the internet because those campuses are where it was accessible first. Two movies from this period describe the time between the PCs being normalized and the full-on internet era: Hackers shows the east coast scene with a hard edge and great soundtrack (fun fact: it was also the inspiration for the UT studio furniture). Sneakers tells the story from the vantage of San Francisco and the Bay Area writ large with a more humble (and funnier) telling of hacker culture. In that moment, computing was obviously consequential but had yet to rewire the economy and reconfigure San Francisco’s geography. Then things went from nerd hobby to next economic reality real fast. Here’s a timeline:

• 1991: First website published by Tim Berners Lee (in Switzerland)

• 1994: Time Magazine cover story describes “The Strange New World of the Internet.” Very few people have heard of it at this time.

• 1995: Netscape, the company that made the first commercial web browser, IPOs and begins the frenzy of the “dot com era.” If you want to see a visual metaphor for what this period of time felt like, watch this completely unrelated commercial from Sony (watch it even if you don’t care about any of this; it’s beautiful and filmed a few blocks from our hotel).

• 2000: Pets.com, a website to buy pet supplies shipped to your door, IPOs and then collapses all within the year. This is the death knell of the dot com era. The bubble bursts after five short years.

Wired Magazine was the banner of San Francisco culture in this era. Kevin Kelly, a co-founder, wrote about the period outlined above in a manifesto of sorts called We Are The Web that is worth a read. Spoiler alert: he ends it with the conclusion that even if the web boomed and then busted, it was not going anywhere. Was he ever right!

“In retrospect, the Netscape IPO was a puny rocket to herald such a moment. The product and the company quickly withered into irrelevance, and the excessive exuberance of its IPO was downright tame compared with the dotcoms that followed. First moments are often like that. After the hysteria has died down, after the millions of dollars have been gained and lost, after the strands of mind, once achingly isolated, have started to come together – the only thing we can say is: Our Machine is born. It’s on.”

From roughly 1995-2001 the South of Market (SOMA) neighborhood of SF was colonized by what is now called “tech.” (See here for images of what it looked like before that). Through the dotcom era, the cultural ideals of openness and futurism often clashed with big money. It was a cliche—and a truth—that tech founders fought with “the suits” who inevitably had MBAs and wore a strict uniform of khaki pants and blue button down shirts. It all felt like there was impossible amounts of money to be had, but “big money” was then in the millions, not today’s billions, and money was freedom to work on hard problems and build cool stuff, not an invitation to rule the universe.

One of the best ways that I can describe how tech was different at that point in time is to tell you about ‘jwz,’ Jamie Zawinski, and the DNA Lounge. jwz was an early Netscape employee who decided to spend his windfall by buying a nightclub so he could hear more goth music. At this point in time, people in the tech community used their resources to do things they found interesting. Today it’s more common that people use their resources to do things they think the economy will find interesting.

There was a generally egalitarian mood in the tech community then. Anyone could participate because no one was an expert: there were no phds in the web in those years, no degrees of any kind related to the internet, really. That pervasive feeling of possibility is what convinced me to drop out of college and spend two years working in startups (Deepleap and then KnowNow, both RIP). All of the opportunities I had then, and a good degree of them since, came from the networks that I was part of. That sounds like I was shaking hands and shuffling resumes, but what it actually meant was finding something I was crazy excited about, going to and throwing parties with other people who were equally excited, and obsessing over what we found interesting together. If you’ve ever heard me say in the context of UT, “optimize for what you want to do”, it’s because that period of time launched so many people I know in directions that were meaningful and in some cases immensely lucrative. Just for the record: you don’t need to drop out of college to do any of this.

A couple decades later and tech’s role in society is clearly different. Egalitarian tendencies of the tech community are still present, but they’ve been dwarfed by the weight of real economic interests and a libertarian impulse that we will return to in a moment. Our timeline continues:

• 2004: Gmail launches, and with it the “Web 2.0” era. This is what we called social media when it was still friendly.

• 2012: Facebook IPOs. In retrospect this is the beginning of the end for the web as a place of optimism.

• 2018: Cambridge Analytica’s abuse of social media for the purposes of political manipulation is made public. The innocence of the internet is finally obliterated.

By 2018, Meta/Facebook was the 5th most valuable company in the S&P 500 and 50% of the top 10 companies were tech. The economy had become digital and companies leading this charge had been rewarded with immense financial resources. People had to stop calling Google and its peers startups: big tech joined big oil and big tobacco as morally compromised endeavors. The one-season TV show Devs makes a compelling vibe out of this combined sense of unbridled optimism fused to unsettling levels of power.

It is in this era that the geographic center of gravity of Silicon Valley began shifting from SOMA in SF back to the so-called peninsula in and around Palo Alto. Google, Facebook, Apple would eventually spend in the neighborhood of $5-10B on their corporate campuses in Mountain View, Menlo Park, and Cupertino. Such growing footprints one hour south of SF eventually spawned the “Google Bus” phenomena of private shuttles taking people from their homes in the city to their offices in the burbs. At first these buses used public bus stops (and public roads, of course) but paid no fee, which led to substantial protests. Solnit wrote about the whole affair alluding back to the era of the 49ers:

“I think of it as frontierism, with all the frontier’s attitude and operational style, where people without a lot of attachments come and do things without a lot of concern for their impact, where money moves around pretty casually, and people are ground underfoot equally casually.”

For a cinematic take on the economic reality described here, watch The Last Black Man in San Francisco. While you listen to its haunting soundtrack, look up the composer and see who he is related to. Tech is now the economy and the economy is now tech.

The Californian Ideology, written by scholars Richard Barbrook and Andy Cameron described the intellectual underpinnings of this era. The ideology holds that technology is an inherently liberating force, that the market will distribute its rewards justly, and that legacy institutions—government, community, place—are friction to be optimized away. In this worldview, the Google Bus isn’t a provocation; it’s an efficient solution to a transportation problem. Plush corporate campuses aren’t an affront to a region suffering a housing crisis; they’re monuments to innovation. The ideology converts winners into “visionaries” and losers into casualties of progress. Heady if you’re on the right side of the equation, and potentially delusional if you’re not.

The history of the Bay Area shows us that technology, when abstracted out of the places and situations in which it is created, creates a fever dream. The Dotcom era crashed, in part, because investors forgot about the literal physics of shipping goods like pet food. It took another 15 years for that infrastructural problem to be solved. The era of social media is faltering now because its protagonists forgot that the human need for community cannot be satisfied by comments alone. It may take another 15 years to repair the holes that this has torn in our democracy, but the work of repair is more ideological than technological.

While you are here, remember that in addition to big tech, meaningful social projects including gay rights, disability rights, environmentalism, slow food, free speech, and anti-war movements all were born or amplified under the Monterey Cypress and Eucalyptus of San Francisco. In this city of nearly constant fog, bouts of luminous clarity are not to be missed. That’s as true for the generational events listed above as it is for the daily moments where sun pierces through all manner of meteorological conditions to light up the city, if even for a beat.

These weeks: Leadership trek, guest lecture by Alastair Parvin (notes forthcoming), continued work on UT++, signing admissions letters because that wave is coming to shore soon. Fog in SF, fog in Detroit.

When Jack Bernard, junior in UT, came up to me after a recent lecture by Mariama N'Diaye, Director of Atlanta’s Innovation Team in City Hall, and asked if he could write the newsletter, I was glad to invite him as a guest author. Continue below for Jack’s reflection on our visit from Mariama and her unique take on civic innovation.

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation are being harnessed to nurture and improve urban life. If you’re new here, try this 90 sec. explainer video.

🪴 Growing an Ethic of Care in Local Government

Mariama N’Diaye is the director of the Innovation Team at the City of Atlanta, and she stopped by to share with us what she’s learned from institutionalizing collaborative and care-based design in governance and what “civic innovation” really means to her. N’Diaye’s background combines a rare set of perspectives, having studied urban planning and business at MIT and then doing a thesis on civic design. All of this combines into a passion for deeply rethinking how government could work to serve its people. Stay with me to learn how the City of Atlanta is working to deeply consider outside perspectives and embracing civic design as a mindset to create a more ethical, responsible and competent approach to governance.

Even before her guest lecture started, one of N’Diaye’s guiding principles became apparent just from observing how she introduced herself, interacted with others, and met everyone with a smile and a handshake. Putting “relationships over requests” is not just something she preaches, but what she embodies and practices, even here on campus. Any business student could tell you that your network is your net-worth, but in a civic design context, relationships are the raw material of co-creation, whether that is of products, policies, or outcomes in the public realm. This is not just to build buy-in or to check a community engagement box, but to create space for people to feel seen and heard throughout the design process. In this way, integrating care into civic work is not just the prioritization of a moral value, but can be the foundation for innovation in a more just and pluralistic society.

Here at the University of Michigan we have an entire research lab dedicated to understanding barriers to delivering effective public services, still searching for many answers in 2026. Despite this topic taking up some academics’ entire dissertations, improving the way the government works also seems to be a topic that everyone—from the clerk at the grocery store to the billionaire CEO of a failing social media platform—has an opinion on. Each person seems convinced that their idea of a seemingly simple solution A) has not already been thought of or tried, and B) is comprehensive enough to solve these complex problems entirely. “If they just did X, then Y would never be an issue!” Are these ‘armchair governance’ kinds of comments pure criticism? N’Diaye helps us reframe the question and instead asks what value can be taken when interpreting criticism as a way of showing up, and speaking up as an act of care. Seen from this perspective, frustrated citizens represent an untapped source of energy that can contribute to building and shaping the work of government—if that government is ready to listen, engage, and co-create.

As a prominent scholar of political theory and care-based ethics, Professor Joan C. Tronto presents five “ethical qualities of care” to outline what it takes to understand and empathize with those you are working with. Responsibility, responsiveness, attentiveness, competence, communication and plurality are all values N’Diaye puts forth in her work, and uses in a civic context to understand how to best meet the needs of the community in the outcomes her team creates. A care-based approach is critical to civic work due to government’s scale. The term “innovation” is a piece of coded language often associated with picking up, turning and shaking out the current situation to envision new ways of doing things and disrupting old norms. But even if a system is broken, inefficient or in need of change, people take comfort in what they know and often attach value to the status quo. N’Diaye seeks to understand such community values before pushing for change. The result she seeks is one where outcomes are in-touch with the needs of a neighborhood or community of people.

This doesn’t mean N’Diaye’s team works at a snail’s pace, though. In an age of rampant techno-solutionism (increasingly in government contexts), prioritizing listening and understanding over immediate disruption leads to far better outcomes for all parties involved. To create a space for marginalized voices to be heard and understood is to create a place where solutions emerge that uplift and empower, not drive deeper an existing divide or disconnect. A more meaningful solution that aligns with community values might take longer to produce, but will require less remediation in the long-term by virtue of early investment in trust-building.

Leading with care means understanding where people are at, and building tools and solutions that meet them there to move the needle forward, not jump past them entirely. For example, a flashy new website or digital product will do nothing to help solve an issue if the vast majority of the user base finds it unhelpful or offensive. Instead of disruption, N’Diaye sees civic innovation as “moving forward together,” and understanding: “Who has tried to work on this? What is the history? Who can we travel with to try to fix this meaningful issue?”

While good user experience of public-facing tools and services is important to consider in civic contexts, N’Diaye also flipped the perspective and discussed how technological tools can be used to improve internal operations, help civil servants better understand their problem space, and build solutions faster. Instead of treating technological development as a silver bullet to fix all issues facing the public sector, she framed technology as a tool to help with the real problem solving—performing engagement, defining issues, and building capacity.

How can we deploy a stronger communications strategy that leverages digital resources to bridge information gaps within the community? How can we develop a new data standard to unify strategy and operations among different city departments? How can we use digital tools to better understand stakeholders, their needs and conditions? How might we use technology to back up claims, hold people accountable, and push development forward at an accelerated pace? These are some of the questions N’Diaye and her team seek to answer. None can be answered sitting alone at a computer.

It’s easy to hire a full stack developer to vibe-code a new intake form that uses a trending typeface and pastel colors, but N’Diaye’s definition of civic technology is not interested in band-aid solutions to individual problems. Design-based civic innovation requires knowledge and understanding, or “Context over Disruption” as N’Diaye put it in her presentation. To gain this critical context, teams must first build the infrastructure and tools needed to understand their people and the problems facing them. This provides a framework to shift from a reactionary and top-down model of public services, to a proactive and uplifting approach to civic tech.

If there was one main idea I walked away with from N’Diaye’s lecture, it was that “innovation” is a complicated, multifaceted, and sensitive topic that requires a lot more work than just the code behind an app. There are obviously lots of problems in government, just like there are lots of problems in any organization, but the solution-making process has much more intimate and delicate stakes in the public sector. Innovation here is not just for innovation’s sake… to understand “government efficiency” we must first understand the role of government, and its goals in serving the general populace.

The premise of “co-design” is innovation with, not just for the people to be served by any solutions created. Building relationships and integrating community into design processes is the future of civic innovation as N’Diaye described it. Government doesn’t get to choose which people to serve and which to ignore as outside the “target demographic.” All people use services in unique ways with individual needs. The biggest idea of Mariama’s talk was her suggestion that the most important capability of the public sector is to bring individuals with professional and lived experiences together to understand common needs and create reliable solutions.

These weeks: Career fair. Leadership trek nuts and bolts. More friendly faces in the studio for talks. And a random 60° day in the mix.

When I saw a proof of concept website float across my feed recently that was using video footage from a bike ride to conduct a similar assessment of building conditions—this time in Ireland, not Detroit—I was excited by how much things have changed in a decade. Cheaper hardware makes it possible to give lots of people video recording devices and GPSs. Cheaper compute makes it trivial to process the hundreds of frames that even a short bike ride can produce. LLMs enable a form of qualitative analysis with scale and speed. Add all of this up and it prepares the pre-existing means of mobility in cites that includes bikes, cars, and buses to become potential platforms for ambient sensing. The Spatial Dynamics Lab at University College Dublin is doing exactly that. This week I interview Brian Rogers, Research Scientist at UCD, about his work on making bikes into the ultimate urban sensor.

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation are being harnessed to nurture and improve urban life. If you’re new here, try this 90 sec. explainer video.

Interview with Brian Rogers on Urban Sensing

On a zoom call representing 100% of all popular spellings of the name Brian/Bryan, I spent a rainy morning talking about cycling, urban sensing, and tech stacks with Brian Rogers. While he’s now at the UCD Spatial Dynamics Lab as a Research Scientist, Brian’s career includes a background in digital media and advertising, so he brings a unique set of tools to the urban tech effort—namely, an affinity for telling accessible stories via compelling websites.

BRYAN BOYER: How do you make data with a bike?

BRIAN ROGERS: A theme that's emerging in my work is the deployment of the bicycle as a scientific instrument and asking the question: what can we equip a bike with in order for it to produce usable and insightful data from cities? The door into that work for me was through cycling safety. I became interested in the question of how to make cycling safer for folks in the urban context because that is the largest barrier to people taking up cycling as their preferred mode of transport. An example is the Collision Tracker project completed with support from local newspaper, the Dublin Inquirer, which is an effort to uncover the true picture of cycling safety through crowd-sourcing reports. The more devices I might attach to the bike in order to help with the cycling safety question—namely the GoPro camera to collect imagery or other sensors—the more applications started to emerge. Someone from the architecture side of the lab here asked if we could use that technology to identify building conditions instead of hazards for the cyclist, and that turned into a project called Colourways with Tiago Tamagusko, Lasith Kottawa Hewamanage, Titouan Desnoyer, and Sonia Soubam.

BOYER: When you initially thought you were going to look at cycling safety, were you using computer vision to look for potholes or something like that?

ROGERS: Yeah, so that project used emerging computer vision techniques, namely video LLMs, as a tool that could potentially help us identify or rate the safety risk for cyclists in a given urban environment. So, what we were looking for is a set of parameters that include surface quality, the level of separation between you and other road users, and the type of vehicles and users that you’re sharing the road with. We built upon CycleRAP from the International Road Assessment Program who have a system to codify safety for cyclists based on observable characteristics of the environment.

BOYER: You’re using the LLM to evaluate against their rubric?

ROGERS: The LLM takes an image as an input and as we are set out we ask the LLM to see if it can identify and rate the presence of certain hazards. What we noticed is that these systems are sort of opaque, aren’t they? They’re hard to debug and tweak. As developers of the system, if we have a checklist of things that we’re looking for in the scene, but we also want the LLM to output an explanation of what it saw, and why it outputted the score that it did. This allows us to get some sense of why it chose the score that it did. From that analysis we might have 20 data points each. Ranging from very safe to very unsafe, and then we’ll run another layer of analysis to come to a general safety. Of course, we’re not just running this on one image but on an array of images that are geolocated across a street, so the last thing we’ll do is layer up those images and output a [set of scores every few feet for an entire street.]

BOYER: Your data is dependent on when the video was recorded, right? Do you think about combining nighttime and daytime, or clear weather and rainy days, to get a variety of conditions?

ROGERS: Is averaging, as you suggest there, actually useful? To me, as a cyclist, or to someone else [like] a friend or family member that I might recommend a route to, is the average safety of that route an effective way to recommend whether they take it or not? Or should you always be taking into account when they’re going, time of day, time of the year, whether or not there’s something going on in the city at that time. I think it may be an attractive idea that you can average safety, but there might be pitfalls to it. So what it means for our project is that we’re more likely to do a morning commuter study on this street with this tool, for example. And that will tell us, for this group of people at this time, they’ll experience this level of risk or safety.

BOYER: Do you weight safety factors differently for different personas or use cases?

ROGERS: Yes. Interesting that you mentioned weighting. One of the emerging ideas in cycling safety and road safety at the moment is that of perceived safety rather than objective safety. You and I might have different risk tolerances as we ride a bike through the city, and we’ll certainly have different risk tolerances to someone who’s only begun to cycle. Or someone who’s maybe a little bit more elderly, and whose balance isn’t 100% anymore. So it’s becoming more of a focus of our project to try to make an output that’s malleable or applicable to specific personas.

If you’re in the active travel safety team in your municipality, and you are interested in reducing the level of risk experienced by cyclists, you can make a huge mistake by baselining to the average cyclist, right? Especially if your goal is to increase participation in active travel [because] active travel sees higher uptake when you cross a certain threshold of safety for everyone. So when the risk-averse person is comfortable on your network, then you’re going to see a reduction in private car ownership, and you’re going to see CO2 savings at a much higher rate than if you made it, for example, acceptable for a guy who’s used to mountain biking.

BOYER: What’s something you put on the bike that didn’t work, in terms of data collection?

ROGERS: The GoPros themselves were not a home run! You do face issues there, especially when you want to do a citizen science project and equip members of the public with cameras so that they can collect data on your behalf. You can run into trouble with GPS signal lock for a GoPro where maybe someone will begin to record their commute, but for the first number of minutes you can actually get a very low-resolution GPS signal, and you’ll find that when you look at the data that the [location is not accurate until the signal] finally locks in. Then that piece of footage is unusable. People will surprise you by the way that they use these things. They’ll turn them on too early, and then you’ve got footage of the inside of someone’s garage, or they’ll turn them on too late. It’s just never a ‘clean room’ when you step into the citizen science side of things. You need to really take account of the full system, including how users charge the video recorder, how they upload the footage, and things like that.

BOYER: Colourways then led to using ground level imagery for building vacancy assessment as part of The Building Stories project at University College Dublin. What’s that about?

ROGERS: Building Stories is a project at the Spatial Dynamics Lab here in UCD that is trying to solve one of the largest civic challenges in Ireland, which is to make a universal, up-to-date vacancy and dereliction registry of buildings. This is particularly of interest at the moment in Ireland because we have a housing crisis. Particularly in urban areas, folks are under enormous rent pressure. Alongside that problem, we have visible dereliction and vacancy in our cities. I’m sure maybe Detroit has a similar phenomenon in places. The approach for collating this enormous database is to use existing public data sets like utilities such as gas usage or water usage that could help paint the picture of whether or not a building is in use. There’s other ideas in there too, like whether or not a building is receiving mail. As you can imagine, the team ran into a lot of questions around collaboration with public and private entities. We have data sensitivity issues and GDPR considerations, so there was a desire to find other data sources that could help inform the synthetic view of the data.

Where I came in was around this question of building condition and whether or not we could assess the condition of an individual building and attach it to the row of the database. With our work on computer vision and video LLMs, we started to ask what parameters would indicate to us that a building is vacant? In our cycling project we looked for potholes and missing bike lanes, and what would be a visual giveaway that a building has been neglected? You’ve got window and door condition, facade condition, is the brickwork maintained? Do you have peeling plaster? Is the paint in good condition? Roof condition is very important, but a challenging one for us, given our ground level view of the video recording. Then you have another category called ‘signs of life,’ such as whether there is a Christmas wreath on the door for example, or another timely signal of habitation. We put together a dozen initial characteristics and we experimented with different video LLMs to get a sense of how well they would be able to recognize each characteristic. Then we have this output of a parameterized score sheet, which we can translate pretty easily into an overall assessment of vacancy confidence.

[To test this approach] I went to a small town close to the border of Northern Ireland called Clones. I took out my bike from the back of the car and rode around the town. It didn’t take too long, and we collected footage from each of the main arterial streets of the town and then ran an analysis on that data. That output is what you see in the proof of concept site.

BOYER: Was the municipality of Clones involved in this or was it a guerilla action?

ROGERS: Before we set up the indicators, we spoke directly to the municipality and to the team who carry out vacancy surveys. They were able to help us with some of the indicators and contextualize what they consider vacancy. Briefly, they showed us that what constitutes vacancy in a certain town might not hit the bar in the next town. If you have a well-funded main street and one door starts to peel, it might get flagged very quickly by a local authority. Whereas in [some] other neighborhoods or towns that wouldn’t be a big deal. So how does your system account for that? It’s a matter of perceived and objective outputs.

BOYER: Hardest question last: what’s your favorite city and why?

ROGERS: Dublin's a place I moved from in 2013 to move to New York for nine years. When I moved back I got to see it in sharp contrast to that city. When you grow up here, Dublin is the New York of Ireland, right? It's the biggest, loudest place you can go. But coming back from the U.S, you're reminded of many of the characteristics of a town that Dublin still has: its walkability and size are, for me, sort of in a sweet spot of the city-town spectrum. Second place goes to Paris. As an example of decarbonization on the mobility front in the past few years, you can't look past Paris. It is such a cool and interesting place to go with so much going on beyond the mobility changes. In my work, some cities become sort of case studies for their bike infrastructure, don't they? But maybe if the bike lanes are the most interesting thing in the city, that’s not a great sign! That’s not a problem that Paris will ever have.

These weeks: UT++ proposal drafting, tee crossing, and eye dotting. Guest speakers in the studio. Snow on the ground. Thanks to Alex Alsup for the Motor City Mapping history lesson in the intro!

The new website can act as a forcing function to check in with what an organization is really trying to do and what people need it to be in the world. I’ve seen the pressure of having to put an organization’s online presence into a new navigation menu cause yelling matches in a conference room, all stemming from different mental models of what the org does and how it should relate to the outside world. In that example, the website is little more than a mirror reflecting the org back to itself: is the organization ready to see itself as it truly is today, and as the outside world sees it?

Something similar is happening right now at a larger scale as AI and automation in general invite (or force!) the professions of the built environment to rethink their role in society and responsibilities to the public. The question is not just how planners do their work, but what work they should be doing.

Should planners spend the efficiency of AI to answer 20th century needs a little faster, or can this moment of disruptive change be used to automate things we now understand to be relatively simple and free up capacity to focus more deeply on 21st century urban concerns, such as affordability and climate? If that is possible, what are the new ways in which society should expect to convene, discuss, debate, and build consensus on urban issues?

These questions are the kind of awesome discussion you have when you ask Pamela Robinson from Toronto Metropolitan University, hey, what’s going on with planning these days? Through our discussion I came to see an emerging vision of what we might call ‘supportive disruption.’ In contrast to unbridled tech for tech’s sake, supportive disruption is how we might describe the kind of change that breaks with the past by unapologetically embracing present technology in order to center the pressing needs of the future.

Full interview below!

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation are being harnessed to nurture and improve urban life. If you’re new here, try this 90 sec. explainer video.

Interview with Pamela Robinson

Pamela Robinson teaches at Toronto Metropolitan University; writes about urban planning and open data for academic audiences and the general public; sits on the board of the Metcalf Foundation; and does research at the nexus of technology and planning. She was generous enough to spend a morning chatting with us on Zoom between the Motor City and the Big Smoke.

BRYAN BOYER: In 2026, what do planners do?

PAMELA ROBINSON: I think a planner’s principal job is to work with communities to guide and shape the change of those communities over the short term, medium term, and long term. We’re bringing expertise from a bunch of different backgrounds: technology, policy, design, and things like transportation, housing, parks and open spaces, social well-being and social infrastructure. Our job is to put that knowledge and expertise into service to community wants, needs, and ambitions in a way that—we hope—builds on people’s potential to thrive and be happy in the places where they live. It’s a big mission statement, eh? But that’s the goal.

BRYAN: What does technology mean for planners in 2026?

PAMELA: It’s both friend and foe. Practicing planners have lots of questions about it. When I’m wearing my planning educator hat, it’s the thing that students want to use at school but are the most worried about for future work. Planners are turning to large language models for tasks a junior planner might normally do.

I’ve had lots of senior people say, “I just want to train AI to write all of my planning justification reports,” and my answer is, well, you learned how to write those because somebody mentored you. You can’t just skip over the junior planning job and expect to have seasoned planners who can actually work with those justification reports [in the future]. You need to invest the time in the next generation of planners. If we don’t, we’re going to have a pipeline problem.

In terms of what we are seeing in the profession, the place where I’m seeing the most uptake is the use of AI-informed tools to track development applications, whether it’s to help residents pre-check their applications or as a transparency tool. In some places, municipalities are experimenting with using the tools to sort which kinds of permit applications need a site visit versus which ones are okay to approve. The City of Edmonton has some interesting work happening there. It seems to be the most common use case.

BRYAN: Isn’t AI in that context mainly just bridging the gap between unstructured data (PDFs, forms) and structured data, rather than doing anything unique to the built environment and permitting?

PAMELA: AI is doing something we should have done on our own sooner. It is allowing people to start asking, “What really needs attention?” or “What’s probably okay [to approve] right now?” The arrival of technology ignited that conversation, but as planners it is our job to continuously look for better ways to do our work.

One existential worry I have about that is we’re asking the wrong fundamental question. Given that the planet is on fire, that we have significant human rights challenges in cities, and that young people feel quite despondent about where they’re going to live because housing is so unaffordable, I think we need to be asking ourselves, is the acuity of the planning process the wrong acuity? The process we use was created at a much simpler time.

We’ve continued to modify and iterate on a process for shaping land that came from a much different time. I would argue maybe we should be using our attention and our effort on much more significant things. In a city like Toronto, for example, we have this thing called the Committee of Adjustment, which sounds so Kafka-esque, but it’s where you go if you want to modify the zoning bylaws because you want your porch bigger or your shed to move. I feel that in a city that’s so vibrant and dynamic, but also has some significant challenges, I don’t think we should be so anally retentive about how big someone’s shed is relative to the other things that are happening.

I think some kind of urban preciousness is at odds with the issues that we really need to tackle. The profession needs to confront those big issues around where we are going to spend our time and energy and what really matters. The arrival of AI to sort and dispatch the easy things is a bit of a crutch to keep us from actually doing some harder work.

BRYAN: Tell us more about the “harder work…”

PAMELA: What’s the cost of investing public time and money on small, planning issues? Some of those small things might matter, but sometimes they don’t. If we were going to start now and build the planning process [from scratch], I’m not sure we would take the resources we have to focus on minor variances to the zoning bylaw for homeowners. In a city like Toronto, where half the city rents, those are significant municipal investments in half the population when there are other issues that the other half of the population is worried about like access to stable, affordable rental housing.

It may not be a direct transfer, but if we took that capacity and redeployed it to figuring out the best way to get the big things done, could we make more progress? If we continue to use the same old processes, and those things take time and capacity and money, maybe we should be asking ourselves, “Is this the right use of limited time, capacity, and money? How might we work differently?”

BRYAN: How would you use AI, open data, or other technology in a new approach to planning?

PAMELA: It needs to start with the politics of sharing land. We have not paid attention to producing housing that is accessible to people for way too long, and now we have a crisis. We need to figure out how to get more units that are affordable and stable into people’s hands. We need to get more people better housed, faster. Full stop.

We have this interesting situation in Canada right now where some of our existing rental housing stock is starting to age out. There’s a need for significant reinvestment: building owners are selling, large financialized landlords are buying it all up, and so we’re losing housing stock. Preservation of existing housing stock is a top priority. Part of the challenge is political will, part of it is finding the money, and part of it is creating incentives to reinvest in that housing stock.

What’s the role that technology and open data plays there? I’m not entirely sure.

There are people who see tremendous potential in gentle density increases through things like accessory suites, laneway houses, and adding on through modularity. I’m interested to see how this scales. What happens with labor there? I recently learned about an NYC skyscraper that’s being built with modular units being brought on site that sticks with me. I was sitting with some engineers, and I was asking them, “Do you think this off-site building will extend the life of construction workers? And does it make it easier for construction workers to live in more affordable places because their trade expertise can be off-site? What are the greenhouse gas emissions of bringing all this big stuff in versus pouring concrete on site?” There’s a role that good analytics can play in answering these trade-off questions. But I don’t, yet, see a ready technology solution that scales. We need to set the agenda politically, and then explore how the technology and data tools can support those ambitions, rather than leading with a technology-first solution.

BRYAN: This is exactly how we think about it in Urban Technology at U-M, but/and I’m hungry for a concrete example of how you see tech changing planning?

PAMELA: I had this conversation with a national organization recently that’s sitting on a trove of Main Street data across cities in Canada. They have this amazing tool that’s rich with data about physical space, business makeup, and how people spend money in neighborhoods. We got talking about whether or not we could make a Main Street digital twin, and if there was a way to use the twin technology to connect up all business improvement areas across the country.

We have all these individual organizations focusing on local neighborhoods, but is there a way to find opportunities in the aggregate data, or a way to find opportunities to connect municipalities that are facing the same challenges? For instance, this could mean using a twin to identify 17 places that don’t talk to each other, but are all encountering the same challenge. Maybe this tool can help cities see new, connected opportunities that aren’t easily seen because the ways in which decisions are made is highly fragmented. I see real opportunity there. If these small organizations could see the bigger, shared picture, they might be able to find some opportunities to work together at scale.

BRYAN: You could also potentially use data from “look-alike” settings that share certain similarities of physical, economic, or cultural factors to make better predictive decisions—is that right?

PAMELA: My prediction is that in 2026, we’re going to start to see the entire community involved in city building (planners, elected officials, community leaders, ordinary citizens) recognize that the conversations about our future need to start much sooner, and they need to be ongoing. We can’t work to our best collective civic potential if we’re only reacting to development applications, right? This idea of who we are going to be together must be an ongoing conversation. It can’t just be episodic and in service and reaction to things that are changing over time.

The other piece we’re going to struggle with moving forward is that we are no longer considering the government to be an authoritative place to get information that’s truthful and accurate. People in the city building professions are going to have to deal with all kinds of mis- and disinformation that we haven’t encountered before. The rise of AstroTurf campaigns as part of anti or pro development projects is a good example of that. How are we going to know what’s real resonant perspective versus generated? But also, how are they going to know what’s real from us [as professionals]? Where does the public go for trusted sources of information?

Planning has always existed with the assumption that there’s an obvious place to go for trusted information, and that assumption is no longer going to hold true. Residents can make their own plans quite easily now, relative to how it used to be, and so this notion of working together becomes even more important. Planning professionals are going to have to get better at working in real time.

BRYAN: What is your favorite city and why?

PAMELA: This is weirdly a really hard question! Right now I’m in a long relationship with Toronto - I’ve lived here for more of my life than anywhere else. Part of my affection is for the city I hope we can be, not for the city we have now. We have a lot of work to do! One of the gifts of teaching urban planning in Toronto is that our students come from across our city and into our region. So we all learn about the nooks and crannies together. If readers come to visit I highly encourage you to come visit our new, incredible waterfront park. Biidaasige Park is ecologically rich and an innovative climate adaptation intervention with the renaturalization of the Don River. So many North American cities have engineered rivers—come see what happens when we welcome nature back.

Postcard from UT Studio

We made a poster for all of the events between urban planning and urban technology. These are in-person, so I’m sorry to tease you but I wanted to share the poster here because 1) I’m proud (and also a little embarrassed it took us this long) to have finally gotten to a place where we have a full slate of enlightened guests and 2) the poster is beautiful and designed by pinni, a studio in Detroit.

These weeks: “pro forma” said to the tune of Riiiiicola, because that makes the spreadsheets go down easier. Studio talks. UT++ presentations and proposal drafting. Surprise snow!

TOC below!

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation are being harnessed to nurture and improve urban life. If you’re new here, try this 90 sec. explainer video.

🔖 Table of Contents

Over the summer, Joshua, a senior in UT who also currently runs the student-led Instagram account, and Charlie did the yeoman’s work of tagging all of our newsletter issues over the years. Thanks to that effort, we can now present some groupings of posts that offer a good cross section of how we think about urban technology.

#AECO (13 issues) - The acronym stands for Architecture-Engineering-Construction-Operations and captures the full lifespan of the built environment. While this term is more used in a professional or economic context, we’ve starting using it here to refer to the work of making and managing the built world. These posts are generally about how AECO is getting all mixed up thanks to ongoing digital transformation in the sector.

#AI (9 issues) - Hard to avoid writing about this topic at this time. Trying to stay sane and hype free.

#Cities Intensive (17 issues) - A core tenet of our curriculum is that you must learn about cities by visiting them, so we created the cities intensive and spread opportunities across the curriculum to get on a bus and go see the world. These posts are some of my personal favorites.

#Civic Technology (5 issues) - There is almost nothing in urban technology that is solely a private sector or solely a public sector concern. Whether projects are executed in public private partnership or there’s a regulatory function overseeing what’s occurring in urban space, a civic ethos is critical to the humanistic way that we approach UT.

#Community Engagement (4 issues) - If there’s one area I am optimistic about the opportunity for technology to enhance, it would be community engagement. As we grow our research efforts, I am hoping to see experimental approaches to community engagement that uses tech to tell stories in richer ways, synthesize and sort feedback, and generally enable deeper listening.

#Data Governance (6 issues) - Still bracingly hard to get right after all these years of tech growth in general, civic technology, and the rise of data science.

#Design (9 issues) - … is the way that we teach students to invent possibilities—technical, cultural, political, social, and economic—that did not exist before. We have a unique and still emerging approach to teaching design so lots of ongoing thoughts here.

#Faculty Interviews (20 issues) - This is a great place to start if you want to understand the people and passions behind our urban technology degree program.

#Infrastructure (12 issues) - Connected infrastructure means digital infrastructure; which means a whole new crop of interfaces to infrastructure; which means a whole new set of decisions and behaviors are being afforded. UX is getting a whole lot more important.

#Industry Interviews (19 issues) - So it’s not entirely industry—we also interview academics and even the Lieutenant Governor—but having these conversations is one of the ways we stay current.

#On the Road (17 issues) - Generally these are stories reported from trips to places like Copenhagen, Oslo, Al Ula, and New York.

#Program Updates (21 issues) - Meta issues that describe the building of our humble little degree program. As Joshua poetically described it while reading through all of these, “it was like watching the weeds come out of a garden” from beginning to current day. That’s progress!

#Prototyping (7 issues) - We prototype to know how products/services/systems behave in the world, and how people behave around those things behaving.

#Sustainability (7 issues) - “What’s the thing you want to be working on for 50 years?” is a question I sometimes ask students. Sustainability is a good answer as far as I’m concerned. See also: Climate.

#UT Curriculum (34 issues) - These are often describing why or how we teach certain ideas and perspectives. This is inside baseball for sure, but maybe you like baseball?

#What is Urban Technology? (13 issues) - The very first issue started with this question and we’ve kept asking it in earnest, every year at least once. One of the things I’m hoping to do this semester 🤞 is to turn the question into a set of lectures with a definitive answer. A student said to me (paraphrasing), “shouldn’t you have an answer to this question?” And we do. Many answers!

Further tags: Architecture, Augmented Reality, Autonomous Vehicles, Data Visualization, Design Fiction, Economic Development, Feedback Loops, Foresight, Interaction Design, Land Use, Mobility and Transportation, Platform Urbanism, Smart Cities.

🌍 Favorite Cities

Regular readers of the newsletter will know that we try to ask everyone we interview about their favorite city and why. Here’s the current leaderboard based on 45 answers and counting:

These weeks: Building is filling back up. Finalizing our winter lecture series. Planning for a trip to CA in a few months. Working on UT++. Cold out, but not as cold as it could be!

On Mobility: Provocations from the Physical, Sensory, and Digital Realms, was edited by Allen Sayegh and Isa He of the Harvard GSD Real Lab. In the words of the official description, it explores the future of mobility in a collection of essays by multidisciplinary experts from academia and industry. Moving beyond reactionary solutions, this book envisions a proactive and integrated approach that harmonizes technological advancements with human needs. Accompanied by evocative illustrations that both challenge and enlighten, the essays cover a wide range of subjects, from urban planning and architecture to software design. This book reimagines the concept of mobility, aiming to craft environments that support and enhance our future mobility experience.

Contributors are: Sebastian Birolini, Bryan Boyer, Steven Johnson (text and amazing, often hilarious illustrations), Paolo Malighetti, Jean Christophe Naour, Dimitris Panikolaou, Andrea Sigrnori, Kyuman Song, Carole Turley Voulgaris, and Brad Weed.

To entice you to check out the book when it’s available, I’m including the full text of my essay below. It describes how AV mobility will change the sounds, sights, smells, tastes, and physical feel of cities. Yum.

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation are being harnessed to nurture and improve urban life. If you’re new here, try this 90 sec. explainer video.

🚕 The Five Sense of Autonomous Vehicles, and essay from the book On Mobility: Provocations from the Physical, Sensory, and Digital Realms

Using the five sense as a framework to think about the implications of autonomous vehicle technology, is something I was anxious to do all the way back in 2016 while working on autonomous vehicle scenarios for Bloomberg Philanthropies. At that time, Anthony Townsend, Varun Adibhatla, Melissa De La Cruz, and I would trade AV concept videos on a daily basis, generally dumbfounded by how they seemed to naively (optimistically?) ignore basic aspects of human experience. I’m thankful for Allen and Isa’s invitation to present at the Mobility Workshop Symposium put on by GSD REAL Lab and University of Bergamo in 2022 and then contribute to this book in 2024. Below is my original text, slightly tweaked for this format.

The 2010s were a boom time for AV concept videos, with so many flooding the internet that it could be hard to differentiate between demonstrations of functional technology, like Google’s koala-faced “gumdrop,” and quasi-functional “wizard of oz” style videos showing physical mockups of vehicles bodies with LEDs and projectors simulating levels of intelligence not yet engineered. Still more videos eschewed physical representations altogether, favoring renderings and illustrations trapped in the uncanny valley like Ford’s City of Tomorrow or Volkswagen’s SEDRIC (SE-lf DRI-ving C-ar).

Predictably, these videos focus on the benefits of the technology but it’s hard to discern from them what one can expect from the lived experiences of an AV city. The videos tout autonomy given to elderly, blind, and others left behind by car-dominated landscapes; tease possibilities for new forms of commerce; and often show cities buzzing with humans set loose by abundant and (somehow) free access to mobility.

What is less clear from the emphatically acted and digitally botoxed faces of rendered humans in these corporate films is how autonomous mobility will change the visceral experiences of moving through urban space. The answer includes more than vehicle experience design. Public policy and economics will be co-designers of the next generation of mobility experiences. What’s allowed, encouraged, discouraged, and disallowed in or around these vehicles will quite possible include dramatic changes to the way mobility is experienced in cities.

Stemming from Toronto’s well-regarded scenarios for AV ridership in private and shared configurations (Driving Changes, 2015), many experts have adopted the idea that if AV mobility can be harnessed as a force for good in cities, shared AVs providing “on demand” rides are the best scenario for widespread benefit to be realized. From the policy perspective, shared rides are defensible and desirable, but the experiential implications for individuals are less not as well explored:

What will replace the smell of gasoline on roadways or that potent mix of musty brake pads and damp that permeates subways round the world? How will the soundscape of AVs evolve? Will the advertisements inside roving AV media boxes obfuscate the need for glass windows entirely?

A review of AV concept videos from the last decade provides an opportunity to explore these questions while contemplating the five senses of the autonomous vehicle: how does AV mobility change the sounds, sights, smells, tastes, and physical feel of the city?

Sound

“Buckle up, David,” the Waymo One implored professor David King and me after getting into the vehicle on a chilly 2022 evening in Phoenix. The experience of each make and model AV is different from others, just like Google Calendar is different from Outlook. Even if they have roughly the same functionality, it takes some time to orient yourself to that specific software. The paradigms, patterns, and patois of each AV species are unique.

The sound of the AV is that of robots trying to deconfuse their human passengers so that they may comply with the requirements of whatever behavioral protocol the legal team has determined prudent. Once that duty is complete, “buckle up” will give way to “buy now.”

Cheery AV concept videos like Ford’s City of Tomorrow (2016) show urban environments flush with mobility options, implying access and thus affordability. Renderings, unlike products, don’t need to have a business model fully baked. It’s hard to imagine that the final list of ingredients that make AV mobility affordable will be free from personalized advertisements. After all, if AVs are roving rich media boxes, then why wouldn’t their environments mimic what’s already happening on the media surface in your pocket today? Sci-fi films like Minority Report and The Fifth Element provide Hollywood glimpses of what it might be like to inhabit a relentless audioscape of commercial jingles and enticements, but Keiichi Matsuda’s HYPER REALITY is still the most vivid of the genre. The visuals are overwhelming but close your eyes for a second and imagine that audio as the soundtrack to your commute.

Sight

Experimentation in AV concept videos and physical prototypes show a fixation with vehicle-to-pedestrian interactions as an important new field within Human-Machine Interaction. Whether implied like the “face” of Google’s gumdrop or extremely anthropomorphic and literal such as the over-sized eyes on Jaguar’s Aurrigo autonomous pod system, exterior eyes give iconography to autonomy. While putting a face on the algorithm might help, that alone does not communicate clearly, particularly in the urban melting pot where interpretations of expressions vary by culture.

Taking this a step further are the likes of Mercedes Benz exploring how external light displays could be used to notify pedestrians of what the algorithm is thinking or, as in the use of blue lights next to the standard red tail lights on their production cars today, merely that an algorithm is thinking.

Drive.ai is an example of making vehicles one notch more verbally communicative and adding dynamic text to the exterior of the vehicle. As with artificial eyes, the efficacy of these approaches in cities where language, literacy, and eyesight are far from homogenous will be the subject of much study by human-machine interaction experts in coming years.

Some experts are pursing an approach where the negotiation between vehicles and pedestrians is not entrusted to the vehicle at all. London-based Starling Technologies, working with design firm Umbrellium, proposed an interactive pedestrian crossing in 2017 that uses LED lights embedded in the road surface to dynamically show pedestrian crossings, safety buffers, and other safety features. While this concept was developed to improve safety around piloted vehicles, it is a potent demonstration of the potential for programmable infrastructure to create rich communications in cities. If concepts like this mature, the sights of the AV city may be more about existing horizontal and vertical surfaces that that of new gizmos trundling down the street.

Taste

Freed from having to tend the wheel of the vehicle, the idle hands of AV riders will almost certainly engage in the same activity that has been common on trains, airplanes, and boats for ages: snacking. The taste of an AV is very likely to be tuna sandwiches, shrimp chips, mocha frappuccinos, and whatever else passengers ensconced in their own private bubbles might enjoy.

Design concepts like the one released by vaunted design firm IDEO in 2014 show a vehicle for four passengers who have the ability to share a cabin while being ensconced in their own bubbles thanks to swiveling chairs—a ‘halo’ above their seats implying hyper directional private audio, and unspecified video devices that let passengers communicate or escape. Kind of.

It’s when private bubbles of taste enjoyment create intersecting Venn diagrams of scent that things will get interesting and, quite plausibly, (emotionally) spicy. What happens when one passenger’s shrimp chips conflict with their neighbor’s breakfast banana?

Smell

The aroma of AV mobility is likely to be jumbled in general, not limited to the mixing of potentially incompatible food odors. AV concept videos emphasize the newfound autonomy that their products will bring to groups like the elderly, suggesting that ‘2-nonenal,’ the organic compound thought to be the cause of that familiar “old person smell,” will be mixed up with teenage perspiration pouring out of bodies exploring the freedom to navigate cities on their own. This on top of the baseline melange of personal hygiene products that today’s public transit riders know all too well. All of that, now compressed into a space smaller than your average bus or train car

Humans will not be the only passengers according to Toyota’s E-Palette concept in the Cities of the Future video from 2019. Leaning into the flexibility implied by the name “palette,” Toyota’s shuttle pods could ferry passengers by day and carry logistics shipments by night. Early morning riders, just after the switch from cargo to commuters, might find themselves enjoying the smell of a cabin scented by loads of mandarins freshly delivered to a grocery, or the gasoline-like whiff of printers’ ink after a load of print on demand brochures. Mmmmm.

The video tells us that in the AV city human and non-human are packet switched across the landscape, meaning that the smell of such a place may involve a tinge of self-pity and sadness as human primacy in cities is reduced in the name of logistical efficiency.

Touch

Outside the vehicle, the programmability of AVs paired with the ubiquity of smartphones frees AV mobility networks from the mandate to have fixed points of embarking and alighting. Rather than meeting at a bus stop, your AV ride might summon you to a specific corner that has been calculated to be the ideal meeting point for you, your fellow passengers, and the vehicle. This programmability is enticing for the convenience it provides to riders, but somewhat complicated for those same individuals while they’re waiting for a ride. At a pop-up AV shuttle stop, where will you touch down—where will you sit?

The ‘touch’ of the urban AV may be one of strained calves and sore feet as the lightness of physical infrastructure in a programmable city translates to new demands on the body.

Inside the vehicle, however, the promise of “new forms of local communities” awaits, in the words of Muji’s GACHA concept video proposed for Finland. Soft rounded benches aim to engender conversation and connection will also allow for manspreading or other unwanted contact between bodies in the same shared space. New etiquette of sharing physical space within close quarters will emerge. Hopefully.

Is it polite to offer a hand to someone stepping into the AV shuttle who’s carrying a load of gear from their softball game, or is it creepy? If you do offer friendly help to someone in need, will you then regret being locked into small talk due to the awkward intimacy of a vehicle that holds at most ten people? Not in Finland, where small talk is notoriously scarce!

Rides Worth Sharing

The questions raised above show just a sampling of the complexity of designing new experiences in an urban setting, where nearly everything is experienced in multiplayer mode. Identifying elegant solutions to the design problems inherent to AV mobility will entail balancing individual interests and needs with those of diverse others, in the form of co-passengers both human and non.

As localities navigate their own path toward an AV future, economic and policy needs will inevitably shape AV product offerings through regulations and taxes. For instance, the US federal government seems unlikely to enforce a requirement for shared mobility in AVs, but some of its cities very well may. Likewise, countries with ambitious climate goals and clearer lines of authority will act more boldly to enforce shared ride modalities on AVs. As these questions are debated and laws are written, the design of AV mobility experiences will also be shaped culturally, and that means locally.

To the extent that shared mobility is important to achieving global carbon reduction goals—and it is—an important consideration for architectural, urban, industrial, interaction, and service designers who take part in this mobility transformation is to recognize that their work is the make or break difference between experiences that are functional, and those that are holistically and sensorially compelling. The role for designers across all disciplines in the mobility transition is to create humane and dignified experiences worth sharing.

These weeks: Everyone’s offline. Gone (ice) fishing! Have a nice new year and we will see you in 2026.

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation are being harnessed to nurture and improve urban life. If you’re new here, try this 90 sec. explainer video.

🎙️ Xiaofan Liang’s work in AI and Urban Planning

When Xiaofan came for her job interview a couple years ago she told me about doing transit planning advocacy work with a neighborhood, using computational methods to analyze businesses, and having studied design thinking as an undergraduate. This rare mix is exactly the kind of mashup that gets us excited here in urban technology. Her research is technological in means, community-empowering in orientation, and cross-disciplinary in execution. We met up on a surprisingly warm autumn morning to discuss her recent work.

BRYAN BOYER: Tell me about your current data center projects.

XIAOFAN LIANG: I have two projects. The first is more qualitative. I call it Data Center Complexity Visualizations. It’s geared toward local planners or local governments who are considering a data center. There are a million different things you have to consider, right? But this tool presents it in more of a cognitive way, in an infographic or interactive media format to help people unpack the complexity of this topic. Kate Crawford’s “Anatomy of AI” was an initial inspiration. We wanted to take information that usually takes 20 minutes to read and help people understand it much faster.

I start with a visual anchor to introduce the topic, similar to: “imagine a site plan of a data center and the different infrastructure components to it. Click on each of the infrastructure components to view the implications associated.” For example, here’s the water line. You click on it, and it tells you the whole narrative of water associated with the data center. I’m working on this project with Matthew Wizinsky and a few students, including some UT students.

BRYAN: It sounds like what you’re making is essentially a decision support tool.

XIAOFAN: Yes, that project is more about education and understanding. There are three things we need to know before we make decisions about things like data centers: stuff about people, stuff about infrastructure, and stuff about governance. That was our starting point. Here’s a sketch of what I described, made by Matthew Wizinsky at Taubman. It’s still a prototype, but it includes the infrastructure elements and the related implications. We’re building this now.

BRYAN: What about your other project?

XIAOFAN: The other project takes more of a computational approach to data centers. Zoning is really a bottleneck for local developers when it comes to determining where a data center can go. It’s also a bottleneck for local governments because they have to figure out the rezoning process. For many communities, they actually address the implications of data centers through zoning. They will revise their zoning ordinance to say things like, “We want a closed loop water system,” for instance. Communities exercise power through the zoning process.

Historically, working with zoning has all been very manual, physically reading documents and finding specific text. Even now with the opportunity to use AI, it’s computationally expensive and not environmentally friendly. So, I want to envision a computerized infrastructure for zoning where people can interact with that as an intermediary rather than the raw zoning text. It’s much more efficient for any kind of inquiry.

We want to use a knowledge graph as a representation for the zoning document because I think it’s the best way to illustrate zoning’s interlocking logic. If you use a traditional natural language processing method, you only look at a specific chunk of text, and you’re going to miss out on the interrelationships to different sections. We recently manually generated a knowledge graph about Ann Arbor’s zoning related to solar energy. The yellow items are the use type, the red ones are zoning districts, the green ones are rules, and the dark green ones are rule conditions. I hope to be able to automatically generate this for most of the zoning document, and then make inferences based on the structure. In this example, you can see all the rules associated with the solar energy system.

BRYAN: When you have that as a graph and you’re testing whether a proposition is allowed, you can procedurally walk through that graph and test every condition?

XIAOFAN: Yeah, you have the ability to structurally compare and contrast different regulations [based on the graphs representations of them]. For local jurisdictions, it can be very beneficial to learn from peers. The graph can be used to view other areas with similar regulations, structures, and logics, and this can inform how they impose regulations in a similar, or not so similar, way.

BRYAN: It sounds like you’re building a machine that converts zoning ordinances into structured data using AI.

XIAOFAN: Yes. The data center and renewable energy are test cases for seeing how this works. Then there’s a question of how well it can generalize to other kinds of inferences people want to ask, whether that be about battery storage or affordable housing, etc.

BRYAN: You teach a few great courses here at Taubman. What is your Urban Networks course about?

XIAOFAN: It’s a conceptualization of cities functioning as a system of networks, which is not a default framework for many people. The default for planners and architects is that cities are place-based systems. A network-based perspective is actually not the most natural form for a lot of analysis. There isn’t really an agency that governs networks, though there is an agency that governs a place or specific infrastructure. So it’s both a conceptualization as well as associated analytics and tools to unpack that conceptualization: a city as a network system.

At the syllabus level, I’m trying to introduce students to a way of seeing the city as a network, with a little bit about social networks, spatial social networks, mobility networks, infrastructure, and more.

BRYAN: What about your course, Urban AI? What is that class about?

XIAOFAN: The motivation for this course came from the sentiment surrounding AI I have observed in our college and in planning students overall. We’re at a stage where many people are stuck on the negative parts, like environmental impacts, and don’t know how to move forward. It can be depressing. People don’t have the knowledge or agency to act towards it, and I want to move past that.

I want to enable planners to play a more active role. In terms of pedagogy, I want people to have a survey of a topic, but go deep enough that they can build something. I want them to learn things by technique so [they are] able to apply these things to a particular context. Last time I taught the course, I used OpenAI’s ChatGPT API because it’s easy to use, and I want to show people how to use Hugging Face models as an alternative practice. We also learn the fundamental knowledge about AI and how to use it in this class.

This class has a pretty radical AI policy, which is that you can use AI in whatever way you want, including if you want to find an answer to an assignment. But you have to write a weekly reflection on how you use the tools and your thoughts on it. The concept is to give students an opportunity to be explorative so they become a better user.

The two main things that I want students to take away from this class are: What are effective and appropriate application of AI in urban contexts? What are AI’s social, environment, and economic implications in cities? These concepts are very important for Urban Technology and Urban Planning students. To understand these two questions, you must also understand human values, ethical concerns, and social risks.

BRYAN: Can you give me an example of a final project?

XIAOFAN: They can all be found online at my website!

BRYAN: What’s your favorite city, and why?

XIAOFAN: The answer has shifted from time to time depending on how old I am. If I were an 18 year old, I would say San Francisco. Everything is very new, you have crazy ideas flying around all the time, new tech terms, and I was drawn to that. Then, I started to dislike San Francisco and be more critical of the tech hype. I started to get more interested in cities with more complex social dynamics or more diverse landscapes. Now, I like cities in Asia because they have vibrant urban dynamics, like Bangkok, Hong Kong, or Guangzhou.

These weeks: New t-shirts arrived. The bustle that comes from exam week being just over the horizon. Final reviews have actually begun. First time having a fit check at a UT final review. Crisp and cold outside.

As one of the original “superdutch” architecture firms that seduced the world in the 1990s, MVRDV works on everything from housing to fancy retail, from market halls with housing tucked above to housing with sports fields stuffed inside. Today they’ve stayed true to their roots as a highly conceptual and experimental design studio, but are also investing in creating their own digital tooling and workflows via the MVRDV Next team. Next helps the firm do things such as facilitate the design and construction of very special buildings, integrate carbon simulations into the earliest stages of design, and explore new possibilities for urban rooftops, by making data accessible and even manipulable as part of the design process.

This week we had the opportunity to interview two members of the team to understand the why, what, and how of MVRDV Next.

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation are being harnessed to nurture and improve urban life. If you’re new here, try this 90 sec. explainer video.

🐖 Interview with Jaka Korla and Leo Stuckardt of MVRDV Next

While touring the MVRDV offices back in February, Leo Stuckardt narrated the evolution of the firm and how they’re now developing their own technology. Jaka Korla had given me the highlights over lunch the day before at a conference at TU Delft, and the office visit provided depth and visual aides. As the three of us talked, there was an electricity in the way that they described their work, and I got the sense that MVRDV Next’s is not the kind of innovation team that gets stuffed into the basement. Instead, the team is truly embraced as part of the MVRDV’s continual effort to redefine itself, and includes important contributions to projects within the broader firm, as well as experiments and installations. Later we met via Zoom to conduct this interview and, due to a technical glitch, we lost the distinction of who spoke each line. For this reason you will see most of their replies below labeled simply as “MVRDV.”

Bryan: Where did MVRDV Next come from?

MVRDV: In the beginning, Next was basically one person, one hour a week while spending the remaining time on architectural projects. This scaled up to eight hours a week, [and then] to a full-time role, and now we are at a point where we have a team of between 8-15 people.

As of now, all of us bring in our own individual fascinations and I think there’s deliberately a lot of space to nurture people’s fascinations. We give them time to develop those fascinations which loosely fall into two fields: climate and computation.

Maybe one tidbit that’s also worth mentioning as part of the origin story is that Winy Maas has the catchphrase of “what’s next?” that he uses often to express a certain forward-looking attitude. MVRDV Next is a department that embodies that mentality both in spirit and in name. So [the team name] is an abbreviation for new experimental technologies that have been around since 8 years ago when we formalized this as a department.

Bryan: What are the backgrounds of your team members?

MVRDV: We have tried to experiment with that. For now, we’re 15 people trained as architects and we also have collaborations, for instance with computer scientists. We feel that the work we do within the practice is still so volatile in a way… like jumping from the really applied design work, which also needs an expertise on floor plans, to coding, that it doesn’t allow us to hire non-architects.

It’s also a question of working culture. The architectural workflow is so chaotic, fuzzy, and nonlinear that it’s—in our experience—easier to teach an architect how to script than to teach someone who has been educated in programming to adapt to the design culture here which to them is pure hell and chaos.

Bryan: Often the technical innovation teams within architecture firms have an origin in providing rendering services to colleagues. Was image-making ever part of MVRDV Next?

MVRDV: It was nice to read that question because it made me remember the moment where, very early on, we realized that it was the one rule that we would make within MVRDV Next: we don’t do renders!

I think I was really quite practical. At the time before [real time] render engines came up, it was just very slow to do a rendering. We were the first team at MVRDV that exclusively worked with screenshots.

Bryan: But your process still produces images, so what is their function if not showing the proposed end result of the firm’s architectural process?

MVRDV: Yeah, maybe to rephrase it: we’re not against rendering, we’re against photorealistic rendering. This is probably one of the biggest questions right now, especially in the context of the whole sustainability discourse: how do we communicate the complexities of material supply chains, of how things are fabricated, of the environmental conditions that they create? How to visualize all of that in a practice that sells [its services] based on visuals is, I think, important.

Within the European context we are experiencing a shift in the narratives that are needed [by people who hire us] away from justifying an end result to narrating a process. There is an overall tendency towards discussing processes, procedures, and so on. There’s still, of course, the question of how honestly [process images created for internal teams of experts] can be used [with clients and stakeholders], and there’s forms of propaganda and convincing that are needed to push, of course. We cannot discuss it one-to-one, but in general there is a shift. Audiences, I would say, are becoming less interested in the kind of glossy rendering end product.

Bryan: Why do you think that is? Why do you think that the operational images have more purchase with decision-makers?

MVRDV: We try to focus on more operational aspects of images. I would think primarily it’s because of the way that teams are set up nowadays in public tenders and also public-private partnerships. Even in strictly commercial projects it is increasingly much more about team composition than about one visionary architect. It’s much more about compromise and collective decision-making and so on. So naturally the question arises of how to design such processes of collective collaboration.

It’s also a question of complexity. Design processes now have so many decisions embedded in them based on different metrics, quantifications, etc. so [it is important to] communicate that clearly and to make clients or stakeholders understand how a conclusion was reached. I think [this is now] almost as valuable as a representational image.

Our focus in MVRDV Next on process and method is so hard that sometimes in the office we’re regarded as the department with the bad graphics. We’re known for, and proud of, our rough in-progress, operational images!

Bryan: Let’s talk about concrete projects you’ve worked on. What is the Valley?

MVRDV: Essentially, it’s a mixed-use development of three high-rise residential towers on a commercial base. The building has this architectural concept of a rock-like formation [like a] valley. That creates an irregular building envelope, so in collaboration with Arup we built a computational form-finding algorithm that would consider a lot of things, such as daylight exposure in the winter, shading in the summer, apartment sizes, modularity of the facade and so on.

[As the building went into construction] we built a custom digital model that represents the entire building facade in standardized tiles, symbolized by letters, like a strand of DNA. We underwent 45 iterations of 40,000 tiles and refined the result for aesthetic, structural, and cost reasons.

What’s interesting is that in a project of this scale you have a complex team. We had around 45 parties between client, us as the architect, cost calculator, structural engineers, facade engineers, contractors and all that. Before we built this algorithm, there was a discussion on the facade and the contractors said there was nobody who could do it as designed. The option was either we find a way to make it, or we go to a regular pattern, which would have been counter to the design.

This was a very good case for us. It’s no longer a question of “how many hours do we save” or “how much cost do we save?” but actually the evidence of value was that this project [could not be realized at all without MVRDV Next].

Bryan: So you can design it, but what does that mean for the person on a scaffold with a hammer? Are they pulling up Revit on a tablet or something to know which panel goes where?

MVRDV: Well, our idea was quite simple. We described [the facade] as a text so basically there’s a 40,000 character string that describes each option of these facades. It looks somewhat cryptic, but you could imagine as a construction worker you can say, “give me an A tile, a B tile, A, A, C, B, A, etc.” It’s intuitive. In the overall project there are only four unique types of tiles.

Bryan: Tell me about Roofscape.

MVRDV: If The Valley is focused on facilitating design and construction, Roofscape is much more about systems, software, and how to deploy those tools to improve urban planning.

Roofscape was an external commission from Rotterdam municipality, and it was commissioned as a research project into new digital methods for rooftop intensification through policy making. Essentially, it was about helping public servants understand how much [architectural program (functionality)] they can place on which rooftops, and what exactly would be the impact for the city.

The project tackled questions like: What data would we need for this analysis, and what data is available? How can we systematically calculate which programs can be placed where on this kind of complex “rulescape” of a city of rooftops with different heights, slopes, sizes, and shapes? We were also asking, how can we simulate future scenarios through some kind of digital tool that tackles all of this complexity and [makes it] intuitive and accessible to different stakeholders? These include urban planners at the municipality and public servants, but also the general public.

Bryan: Just to ask a straw man argument here… why is it better to have a 3D map with all the colors on it instead of a spreadsheet?

MVRDV: With elements like rooftops, the majority of them are owned by private parties and homeowner associations. For them, spreadsheets are not very intuitive. The second reason is that the intention of this tool is not to prescribe the best possible use, but it is about [interactively] showing different scenarios of possible futures. Depending on how you set priorities, or how you would model subsidy schemes or new regulations, different special outcomes will be [visualized].

We’re still a bit on the abstract level, so we’re also looking into ways to add detail. We have a separate publication called the Rooftop Catalog, which shows visual concepts for what the rooftops could become.

Bryan: Are you trying to make software as a product of the firm or is this an elaborate form of internal tool-making?

MVRDV: We’ve presented this particular project in countless conferences, and it’s very engaging. A lot of people are motivated but usually at some point very soon down the line it turns out that it’s hard to find a business case for it. Especially with public authorities. They [might] love it, but they don’t have budgets allocated to it. Many municipalities don’t have a team for rooftops altogether. And if they do, they might not have software development budgeted in, which means they will be interested, but then you have to wait a year at least for the new budget. It’s just, you know… it’s slow.

[Another thing is that] for MVRDV, every architectural project [is a] prototype intended to challenge conventions and propose new approaches. Five minutes from here we have built the Market Hall which has more than 250 exemptions from the Dutch building codes and is a prototype for a different way to build and design housing. The way that we naturally think about software is also as prototypes - and Roofscape is a prototype too. We don’t have the expertise in-house, nor the intent, to build a stable bug-free software with all the tedious debugging and compatibility checks and so on. We would love to see it grow into that, but probably [worked on] by professional developers.

Bryan: Beyond providing a service to your colleagues in MVRDV, what kind of role does Next have within the overall organization?

MVRDV: The key role we play is to test or prototype another form of a practice within a big organization. We have a lot of expertise in design, but if there’s something as big as the climate crisis, it’s very hard to change the way you design [with] 300 people. In a way it’s faster to prototype that within a team of 10 and see what things work, and what things don’t work, and scale them gradually. That way is also very effective to communicate to the outside: “Hey, we are actually dedicated and we do try to change every month with regards to responsible climate aware construction techniques,” for example.

Bryan: Finally, what’s your favorite city and why?

Jaka: I guess it would be an Asian city… something like Bangkok, Shanghai or Tokyo, for their dynamism and chaos and because they feel a bit like an experimental petri dish of future urbanism.

Leo: I really struggle to nail it down to one. I would maybe say Moscow, where I lived for a while during my postgraduate studies at Strelka. The size and the scale of city makes it diverse, rich, and engaging, and the concept of multiple layers of cultures and languages coexisting, which is very visible in Moscow as a kind of border between Europe and Asia. It makes me think of the book, The City & the City. It’s a detective story which is situated in a fictional city that is two cities that are existing in the same place. Different languages, different styles, different cultures, but coexisting in what he calls “cross-hatching,” so I think that’s what I would say about Moscow.

These weeks: Town Hall with students; everyone’s sick, just recovering, or just about to be sick; very pleasant visit to Weiser Hall; successful trip to Cornell Tech Urban Tech Summit with a contingent of 8 of us! 🏃

This summer I had the pleasure of joining Evelyn Lee’s Practice Disrupted podcast to talk about our degree and how it relates to shifts in architecture and the work of the built environment more broadly. I could not be more pleased to have urban technology at Michigan in the same feed as Jesse Devitte, Robert Yuen, and so many other smarties.

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation are being harnessed to nurture and improve urban life. If you’re new here, try this 90 sec. explainer video.

🎙️ Practice Disrupted: Architecture, Technology, and the Future of Cities

Urban technology speaks to multiple different communities of practice that exist out in the world, and our goal is to bring them into dialogue through three commonalities: the blended digital/physical life that we live; the way tech often drives transformative change (for better and worse); and how tech outcomes are influenced by the quality of design—how it’s formulated, not just how it looks.

Some of us are interested in mobility, some in real estate, others in land use and planning issues, etc. I’m into all of it, but have a soft spot for “AEC tech” that includes the software that architects use to organize and facilitate their labor. Reader: there’s lots of room for improvement!

Because of this soft spot of mine, it was awesome to speak with Evelyn who not only has a great podcast, but also is president of the American Institute of Architects for 2025 (!!!), and clearly has 6 more hours in her day than you and me. We talked about the state of digital transformation in the built environment, the UT curriculum, how it relates to architectural education, chain drain, and more.

Click the image below or this link here to hear the interview:

These weeks: Preparing for a trip to Cornell Tech’s Urban Tech Summit with students and faculty, developing another excursion for the spring, going back to square one on project BCWYAF. 🏃

This week we’re going to explore that question by dipping a toe into construction technologies. The topic is on my mind because I am preparing to moderate a panel at the Cornell Tech Urban Tech Summit on November 11-12 (see you there?!), but it’s also a useful lens through which to talk about theories of change in urban technology more broadly.

To do so, we need to imagine an unlikely head to head: Buckminster Fuller vs. Home Depot.

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation are being harnessed to nurture and improve urban life. If you’re new here, try this 90 sec. explainer video.

🦺 Buckminster Fuller VS. Home Depot

For people excited to participate in creating future tech, a key question is how you make it possible for users to adopt your inventions. Does your new tech require collective investment of time, money, and intentionality? Or is it something that individuals can try out by themselves, on their own time, with their own personal budgets? Are you selling to the CEO or the end user?

Buckminster Fuller is one of those thinkers whose quotes are catnip for people that work on the future. He was convinced that the future was going to be so different. Bucky summed up his theory of change thusly: “You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete”.

Perhaps his most famous invention is the geodesic dome, which you’ve definitely seen before, perhaps under the brand name of Disney EPCOT. Other inventions that Fuller promoted were less successful. His experimental deployable housing called the Dymaxion House (1930-1945) is now a nerd fascination with no uptake in broader society. I argue that these failures are because housing is more than merely technical innovation. It is also a social-cultural artifact. Housing is part of your everyday life and it’s a real pain in the butt to move down the street, let alone move into a whole new conceptual model of housing.

Let’s say for the sake of argument that you are 100% bought in on the Dymaxion House, but is your spouse equally excited? Will the local home repair person be able to help you maintain it? Will the mail delivery person know where to put your mail order catalogs? “Making the old model obsolete” sounds cool when your model wins, but damn is it hard to win in this kind of totalizing context! There’s just a ton of people you need to convince to also participate in your new reality.

Bucky’s Dymaxion house is an example of a new platform (specifically, a housing platform) that involves different manufacturing and supply chain, business model, maintenance plan, and maybe even different furniture choices once you move in. Plus it was expensive at $6,500 in 1945, roughly double what a conventional house would cost. Not only do you have to agree with the particular affordances of this unique housing platform, you also have to be willing to drop some serious coin for the privilege of being guinea pig inside Bucky’s new model. No wonder this never caught on; it shifts so much risk onto the customer!

Fast forward to today and consider the Bosch BLAZE Laser Measure that you can buy at Home Depot for $54. Devices like this help a construction worker (or homeowner) take fast and accurate measurements. Today they’re so ubiquitous on the job site that just about every construction crew has one.

This phenomena is relatively new. From personal experience on construction sites I’d say maybe the past 15 years have been the adoption curve where it went from a laser being a special tool that was always put back into its protective pouch to now being something that gets thrown into the toolbox with everything else. They’re cheap and easily accessible, so why handle with care? Indeed, the pulsed semiconductor laser diode that many of these devices use has been getting cheaper and cheaper thanks to Moore’s Law:

Home improvement stores are filled with devices that use batteries, lasers, speakers, and LEDs, all technologies that just a decade or two ago were too expensive to be included in such products at these price points. Since construction sites are dirty and dangerous, when you see tech like lasers and Li-on batteries in this kind of environment you know it has reached the top of the diffusion curve.

What’s neat about lasers on the job site is the theory of change that this reveals. Buy a new screw gun and it works with your existing screws and doesn’t need to interact at all with your other tools. In the aisles of Home Depot you can buy a screw gun from Ryobi one day, return a few months later and add a job site lamp that uses the same battery pack, and so forth because Ryobi is Home Depot tech that has been designed for piecemeal adoption. As you accumulate tools, they all happen to work together through interchangeable batteries and similar design patterns that make them easy to use individually and together. One day you wake up and there’s an entire garage (or truck) full of neon lime green tools and you have bought into a version of Bucky’s “new reality,” but you did so bite by bite, credit card payment by credit card payment.

These days when I go to Home Depot I am eagerly looking for devices with embedded AI, but I’ve yet to find any.

Matt Jones is a designer who has been thinking about machine learning and AI in devices for a long time and wrote a prescient essay titled, Be As Smart As A Puppy. The premise is to “[make] smart things that don’t try to be too smart and fail.” This is exactly the kind of standard that’s required when you’re in the matter battle on a construction site.

Matt advocates for designing to the abilities of technology with clear eyes instead of pretending that tech like AI can do everything. Puppies are not going to solve obscure math problems for you but they do have some agency in the world! Back in 2010 when Matt wrote this essay he and his colleagues were inspired by the bizarre AI-powered toys they were seeing—the likes of which are still commonly available (and just as creepy). AI was kinda clunky then and it’s much less so now.

Home Depot Tech is smart as a puppy. It does one thing and does it reliably. It has bright colors so you can see it on the job site. It’s made to get dusty. It’s cheap enough and good enough. You can buy it in every state. Home Depot doesn’t win because it’s best, it wins because it’s most accessible.

What the Buckminster Fuller examples above show us is that it’s not just the technical capacities that need to be designed for, but also the economic, political, and cultural dimensions that must be considered. Home Depot Tech works because individuals can buy into it one by one until eventually every job site has a laser and—in theory—dimensional accuracy and speed improve across the sector as a whole.

Back to that question at the top: why aren’t houses built by robots? As an experiment I mapped a handful of construction technology (contech) companies to whether they sell something like a “product” or “platform.” Here it is:

Products can be adopted and utilized in existing workflows without needing to convince anyone else to give approval. For instance, Dusty Robotics sells little robots that scoot around a concrete slab and trace out where the walls should go, in lieu of a crew of people “snapping lines.”

Platforms include examples like BotBuilt, which is a offsite construction company that uses robotics to do the work in a factory. In a classic platform kind of way, to gain the benefits of the platform (such as speed and lower cost) you have to give up some freedom. Maybe you can’t use any kind of material but instead pick from their list. Or perhaps you just need to work within the platform’s dimensional constraints (e.g. all ceilings max out at 9 ft).

Change driven by products is incremental and can take a long time to add up to a holistic paradigm shift. When products fail, they do so because of rejection. Change driven by platforms is more transformational but it takes a larger amount of capital (and buy-in) to initiate. When platforms fail it’s because they are unable to get enough capital to build the platform itself or because they cannot attract a critical mass of users to sustain the platform. Products allow you to make a purchase decision at one point in the tech stack. Platforms require that you buy into the whole stack.

So pick your challenge: do you want to be responsive and iterative in your product development or educational and accommodating in your platform proselytizing? See how the map above is bottom heavy? That’s telling.

The built environment sector is full of aspiring Bucky Fullers and not enough people trying to build Home Depot Tech. Platform bets are essentially winner takes all, so they’re very risky. You can invent the coolest construction methodology in the world, but if you cannot get people to buy into your platform then it’s going to have little effect. For that reason, I’m hungry for more product-based innovations—more Home Depot Tech!

When it comes to robotics, the devices are still comparatively expensive and that may be why the dashed blob in the upper right is empty: the robots are too expensive or fragile to sell as products that live on a job site. Cameras, connectivity, and computing that enable the top left of the map are more diffused already. I’m interested to see what happens when construction equipment continues to get as smart as a puppy.

Where is the drill gun that knows exactly how much pressure to apply to the nail, given the material it’s being shot into? Where’s the Friend.AI necklace that gives you framing tips when it sees you building a wall? Or what about a phone app that helps contractors chat with a set of architectural drawings to get the info they need instead of having to jump off a ladder to review printed documents?

None of these ideas are going to change construction overnight, but because you can buy into them one by one, and I can do so at my own pace as well, Home Depot Tech will be making construction a little less dull, dangerous, and dirty while we all wait for one or more of those big Fullerian bets to materialize.

Links: OpenSpace, CivDrone, Dusty Robotics, Propeller, Built Robotics, Clayton Homes, Bensonwood, AUAR, BotBuilt

🖼️ Postcard from Chicago

Photograph by Luther Hoy while on the Cities Intensive field trip to Chicago this week.

These weeks: Autumn is really sticking around. UT++, Cities Intensive. 🏃

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation can be harnessed to nurture and improve urban life. If you’re new here, try this short video of students describing urban technology in their own words or this short explainer video.

👷‍♀️ Eat Your Own Dog Food Before AI Eats Your Lunch

When a talk starts with “I like to call myself a full stack architect,” you know that there’s something interesting to follow because the speaker, in this case Kristen Padavic of Cedar, is able to natively bridge the urbanist/technologist divide. Kristen is an alum of Taubman and she stopped by the UT studio on a recent visit to campus.

Cedar describes themselves as an “architectural design firm built to partner with the modern real estate developer” that has created a platform for evaluating sites and designing housing. In my own words, they combine the feasibility function of the Testfit / Digital Blue Foam / Delve (now part of Google Earth) type products with a set of architectural designs for ‘housing as a product,’ kinda sorta like Juno, and in-house design expertise.

Have you ever tried to pour peanut butter out of a jar? That’s about how fast the real estate-to-architecture-to-construction pipeline often feels. Cedar uses AI/ML to speed up site selection and feasibility analysis in a holistic way that combines zoning, tree coverage, topography, and so forth: the kinds of things you need to know about to really understand exactly what can be built, sure, but more importantly what can be built at reasonable cost. Kristen describes the benefits here as accruing to efficiency and quality both: “80% of what architects do is rote, rule-based crap. If we can automate that, then we get to spend 100% of our time being great designers.” The lesson here is to automate typical situations so you can embrace the edge cases. That lines up with our perspective that the AI era makes judgement even more important. Judgement is the last mile of decision-making.

The result is that they are architects who work at the speed of developers, allowing them to partner with developer clients in ways that previously would have been serviced by a small cast of discrete consultants. Cedar’s internal tooling lets them work so fast they can just go ahead and calculate in advance how many sites there are in Austin where their housing products are viable (18,000) and same for all of Texas (1M). Because they’ve designed the housing products as a suite of different options, they know that on these 1M sites it’s not just theoretically possible to build multifamily housing, but that they could file for a permit to do so tomorrow.

As Kristen explained, this kind of business was unlocked by a confluence of technological and urban trends. The tech side is easy: first ML and then AI have made possible the complex analysis this kind of work demands. The urban side is about zoning changes that open up new places to build housing. This includes changes such as up-zoning, ADU laws, and removing parking minimums. In sum, these changes mean infill development is an area of expanding opportunity in many communities around the country, but/and that infill development is harder to execute. The feasibility is trickier and the design more challenging than doing the equivalent project on virgin land somewhere on the sprawling fringes of town. Cedar’s bet is that the extra complexity that comes with infill development can be managed through well-designed software.

Usually developers live and breathe by the zoning ordinance and know very well how to squeeze the most out of it, but because of the recent wave of zoning changes, Cedar finds that many developers are still catching up. Cedar’s tooling makes it irrelevant: just use the model to test out scenarios on a site. In fact, they can test building scenarios the client wasn’t even asking for, such as single stair buildings (multi-unit apartment buildings with one fire stair instead of the 2+ typically required) which are newly enabled by code changes in some jurisdictions. Because of that newness, single-stair buildings can be perceived as risky, so Cedar’s proactive exploration of analyzing sites for their ability to embrace an important code change in a way of de-risking new building types for developers. This is not a solution to affordable housing in the US, but it’s an ingredient for sure.

Software replacing consulting hours is important because it lets you explore scenarios rapidly. Who else can benefit from scenarios? Just ask my colleague Rob Goodspeed, author of Scenario Planning for Cities and Regions. Cities, of course, and Kristen is already doing so: “We [have worked] with city planning offices in Texas to help them contemplate code changes” to answer questions like “how much more housing would be allowable if we made this zoning change?”

At the moment Cedar does some feasibility analysis by hand using their internal design team, so they’re effectively building their own AI training set on actual projects. This reminds me of working in a Kleiner Perkins incubator space on Sand Hill Road back in the year 2000. It was there that I learned the phrase “eat your own dog food” as we tested the software we were building by using it first among ourselves in-house. Kristen’s work shows that a new aphorism is necessary: it’s time to train your own dog. There’s a race on for who will build the most capable models. As Anthropic, OpenAI, Meta, and Google duke it out for general purpose models trained on the open internet, domains with deep requirements like architecture and planning are not being served very deeply—though it’s worth noting that this could change soon, as Autodesk is working on a foundational model for AEC.

There are not many companies that embrace technological tool building and design work in the way that Cedar does. This blend of eating your own dog food and training your own dog makes so much sense, but is exceedingly hard for a pure consulting firm (which is what most architecture firms are) to find money and time to dedicate to innovation work. “Everything about architecture is about the billable hour,” Padavic explains, “which is why it’s tough to innovate. If it’s not directly relevant to the project, you can’t bill [the client] for that! You can innovate within a project, but not general purpose exploration” and product development.

Ironically, the business model of architecture and planning firms is a barrier to innovating in how architects and planners work. To create radically better ways of working it requires new talent (engineering/dev experience), new business model (software vs. consulting), and a legitimate collaboration of urbanists and technologists, and Cedar is doing exactly that.

Thanks for visiting, Kristen!

These weeks: UT++ drafts. Light work on project BICOE. Got the vestaboard working again! 🏃

These ad hoc notes taped up on buildings and in the public realm are the most prevalent form of post occupancy analysis, which is the term of art for systematically checking to see if the thing we built performed in the ways that we designed it to perform. Reader: very little of the built environment is studied systematically after completion to verify things like actual ease of wayfinding, actual energy performance, actual productivity or serendipity or whatever benefits of the design were promised by a dashing designer in the board room. Architectural design is a profession with a surplus of charisma masking the fact that it exists in a condition of data poverty. The data simply doesn’t exist.

Over the summer we interviewed teams within design firms about how they’re building their own digital tools to take advantage of data they collect or simulations they build, all with the goal of enabling better design decisions. For all the attention paid to AI automating office jobs right now (including design roles), what was most exciting from our research this summer were the examples where a human was in the loop. In fact, that may be the whole point. As firms move to adopt data-infused workflows, this represents a significant culture change whether they know it or not: new data means new tools, means new ways to evaluate work as good or bad, means eventually new kinds of talent are going to thrive in these environments.

Below the jump we are sharing an excerpt from our discussion with Aaron Woolverton at Design Workshop, a landscape architecture firm working in the mountain west and beyond. One thing I found notable from our talk is that he did not refer to monolithic software from the likes of Autodesk, a behemoth in the AECO space. Instead, the future Aaron describes is one of firms building their own workflows, stringing together existing libraries with unique vibe coded modules, merging data sets, and having enough talent in house to make it all bespoke. An ecology of software to design an ecology of spaces, perhaps? Count me in.

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation can be harnessed to nurture and improve urban life. If you’re new here, try this short video of students describing urban technology in their own words.

🌳 Six Questions for Aaron Woolverton

Aaron Woolverton is coordinator for the 16-member DigiLab, which is part of the landscape architecture firm Design Workshop. The DigiLab team explores digital tools and data to enhance design workflows within the Workshop’s projects, spanning from intimate gardens to large-scale master planning efforts. Over the summer we got Aaron on the horn to learn about the digital workflows he’s building.

Bryan: My observation is that a lot of digital work in architecture design firms is script-based and doesn’t have a UI per se, but you’ve spoken about building tools for your colleagues, so the UI becomes critical—is that right?

Aaron: Yeah, that’s totally right, and in a way, that’s kind of what DigiLab is too. We’re trying to create a more standard and universal approach to integrating [tools that use data to accelerate the design work]. Working with data is very quick if [the tools are] set up correctly. [We work on] formulating best practices and [building] a universal workflow. I would credit Chris Landau because he’s the one who built Human UI [which is a tool to create user interfaces that we use in our] workflows. There are maybe like 10 people in the firm who would open a programming environment like Grasshopper and feel comfortable, so being able to build a UI is important.

Bryan: Right, because scripts are still pretty much for power users but UI makes computational power accessible to more users. What kinds of problems are the new tools and interfaces you’re building helping the firm solve?

Aaron: A lot of our work and a lot of what we’re trying to move into has to do with questions like thermal comfort. We’ve been using tools like Ladybug to map and analyze how our designs [would] effect and be effected by the environment [such as wind dynamics]. If you’re sitting down on the park bench and there’s a wind speed greater than maybe 6 meters per second per second, it is going to be considered uncomfortable by most people. So we’d like to be able to create an interface that updates as you’re moving trees around, or moving a building site around, to get a sense of how like wind will move through the site.

Bryan: What is the missing data set that if you would create if you had a magic wand?

Aaron: It would be all about plantings from more of the landscape point of view. Some exist, like the ERA database, and it is actually pretty extensive. It tries to cover the native species throughout the country and stuff like that. But very often, especially with landscape architects, we’re working with adapted plants due to a variety of reasons—it could just be aesthetics, it could be [that] adaptive plants actually can perform better in [a given locale].

Bryan: What are some examples of the missing fields then, if you were to add them to the database you mentioned?

Aaron: It has things like soil properties including ph levels, but I would want more of a mutualism kind of category. What does this plant work well with? Which plants could it provide to or receive nutrients from? Who are its best pollinators? How does it performs as part of an underground network, [but that’s] relatively new science!

Bryan: What kind of expertise do you think DigiLab will need in 3-5 years?

Aaron: With everything we’ve been talking about, UI/UX [is important]. I think people who who can write in python and writing code would be helpful, [but maybe even more so is] literacy and knowing how to get the most out of [AI tools] to facilitate a process. I think a lot of people in my firm think our group is about visualization and graphics, so it I try to like demystify a little bit of that and speak more about data integration [in the design process]. Because of this, people who really understand data [structures] and databases are also who we will be looking for.

Bryan: What’s your favorite city, and why?

Aaron: Philadelphia! 100%! I am a Philly lover. It’s so walkable.

These weeks: UT++ workshop. Cities Intensive field trip to Detroit. More thinking about studio space. Planning some travels for the spring. 🏃

Last time we ruminated on ‘what is urban technology,’ but as much as I like writing on the subject, the truest answer is not mine to give—it’s for the students to bring to life as they pursue internships and careers. Meet a few of them below as they share snippets from summer jobs.

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation can be harnessed to nurture and improve urban life. If you’re new here, try this short video of students describing urban technology in their own words or this short explainer video.

📢 Summer Share Out

We started an annual Summer Share Out event to give students a venue in which to talk about what they did during the break in a supportive and casual environment. The point is to help the whole mix of first-years up through seniors think about internship and career options, build connections between the cohorts, and celebrate our accomplishments as a community of learners. Below are six snapshots from willing souls who attended the event this week.

🪔 Jack, Seattle City Light

As part of the Performance and Policy team, Jack helped Seattle City Light figure out how to create a unified approach to assessing staff performance. He also had the experience of figuring out how to get unusual software (Python and Pandas) installed on a city-owned laptop so he could ply his data science skills in the role.

Q. What is the best thing you learned?

A. Nuances of Electric Transmission & Delivery; Current Problems w/ Bureaucracy.

🚙 Pranav, General Motors

In the Enterprise Innovation and Software Product Strategy teams Pranav worked on prototyping new digital products at GM. Working in Cursor, Figma, and Lovable, his focus was on tech for 5-7 years out.

Q. Which 1 UT course was most relevant?

A. UT 402: Reading & Writing the World taught by Violet Whitney.

🏗️ Katie, RE360

Katie’s work at RE360, a Pittsburg, PA-based real estate development firm, was all over the place in the satisfying kind of way that is unique to small businesses and growing startups: from making maps in Rstudio, to helping improve internal processes by building intricate spreadsheets, to cutting new keys for residents.

Q. What is the best thing you learned?

A. Speak up and ask for what you want, [and you may be surprised when you get it]! I also learned a lot about how the [real estate] industry works.

🏗️ Kenyatta, Toll Brothers

When he entered college Ken was already working in real estate as a broker, so I was not surprised in the least when he told me he was working for one of the largest home builders in the country this summer. Biggest take away? How manual the work of real estate still is, and thus how much opportunity there is for people like urban technology students who are not afraid of automation.

Q. Which software did you use the most?

A. Excel!

🪄 Emma, Cities Reimagined

Emma is the first person to have an internship that’s 100% service design, and with UT-instructor Alex Johnston of Cities Reimagined in Detroit, no less. Emma worked on prototyping new tools for process automation in government and philanthropy, and somehow had enough time to also lead a Jane’s Walk to a very old pear tree.

Q. What is the best thing you learned?

A. Facilitation

👾 Tyler, Cedar

Tyler was in Austin at Cedar, a unique architecture firm that combines architects and software engineers. They create bespoke software that let’s them move more rapidly and with better quality assurance from land planning up to permitting. While prototyping, "I was supposed to work in Figma but I just asked, can I make them?" and did then build a working version in Cursor.

Q. What is the best thing you learned?

A. How to build fast and iterate quickly until desired outcomes are created. Also how to translate between AEC and software people.

We’re proud of you, students! Dear reader, if you’re hungry for me, here are links to similar summer-recap posts from 2023 and 2024.

These weeks: Cities Intensive field trip around Ann Arbor came off without a hitch. More UT++. RECon. Settling into the semester (despite space issues). 🏃

Yesterday we held a welcome lunch for the frosh, who’ve been here just a week. The room was busy with first week of college teenage energy. For those of us working behind the scenes, there was also some “first” energy as we’ve been preparing the past 15 months for the transition to curriculum 2.0 and this week, with these students, we launched it.

Every September for the past couple years we’ve been restating our purpose here in Urban Technology, so let’s keep that tradition running and dive in below. We also share a new video (that you should forward to your top 372 friends please).

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation can be harnessed to nurture and improve urban life. If you’re new here, try this short video of students describing urban technology in their own words or this short explainer video.

🧐 What is Urban Technology?

In 2024 our definition of urban technology centered on the digitalization of the built environment. Back in 2023 the definition was more fixated on the effects of new tech (computation, connectivity, automation) in cities. At the start of this effort in 2020, the definition centered disambiguating urban tech from the broader category of tech. Each time we indulge this exercise it’s a way to provide alternative views into our area of study. Each definition is true.

Our 2025 answer to “what is urban technology?” is somewhat more broad. Working on a new video over the summer, with the pressure on to create a succinct definition, we landed on this: urban technology is a new liberal arts for the urban era. This reflects the way in which our curriculum brings together bits from urban planning, information science, and design.

For our undergraduates, we believe it’s most important for students to build a strong base of knowledge about cities and to fill a toolbox of blended know-how that includes practices from tech and design. Our student Rongyu once described it to me as “full stack,” and that still feels correct. These ideas were on my mind as we developed the video with Lunar North, a local animation studio.

In the smart cities era there was a boosterism that promoted the idea of digital management as the solution to municipal woes. Then and now, that approach felt to many of us on the faculty here as a mechanistic and ultimately fraught conception of the world. Smart cities as peddled by the likes of IBM around 2010 looked like cybernetics 2.0, reviving the dream of a city managed by remote control, perfectly efficient in style of Stafford Beer.

More recently, critiques by the likes of Evgeny Morozov have warned of the danger of solutionism. Just because you introduce a computer doesn’t mean things are automatically better. If anything, that adds a whole new set of risks, such as expansive surveillance as noted by Shoshana Zuboff.

The solutionist and new-luddite framings are both problematic on a burning planet. We try to cultivate a mindset that includes critical thinking and as bias toward action. The intersection of those two is particularly alive in the design process, where the non-linear swing between defining a problem and exploring solutions is useful. When something that “seems like a good idea” is made into a prototype, it’s usually a lot easier to tell if it actually is a good idea: if it will be used, if it will be appreciated, if it introduces unforeseen risks or harms.

Technical systems used as tools—from refrigeration, to tractors, to TCP/IP— are fundamental to supporting a planet with billions of people. That’s where our time should be spent: how can a community know each other better through digital media? How can new interfaces help neighbors be more caring? How can the dull and dirty jobs of sharing physical space be automated to make time for the hard and tender work of being human? That doesn’t mean more tech is always the answer but it’s usually there somewhere. That’s also why the curriculum here has tech as only one of three pillars.

Here we are in 2025: a liberal arts for the urban era. Enjoy the video ad let us know what you think by hitting reply to this email.

🖼️ Postcard from Taubman

These weeks: Trees starting to shiver a little. New faculty making sense of new hallways. Switching up the title on these posts from now on to be more descriptive, but we will keep that weekly count going in the subtitles. UT is actually kinda large now. Welcome, first years! 🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃🏃

💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that data, connectivity, computation, and automation can be harnessed to nurture and improve urban life. If you’re new here, try this short video of current students describing urban technology in their own words or this 90 second explainer video.

👅 The Function of Taste

Emily Kutil and Matthew Wizinsky closed out UT 103 this semester with the Incomplete City exercise, culminating in a wall-sized mural of a city that students imagined over the span of a few days. This is always a high point of the year. It’s boisterous, it’s huge, and it integrates the teaching of UT 103 as well as UT 102 led by Phil D’Anieri. In addition to the faculty who gathered this year to give feedback we also had proud parents in the audience. This was a first for us, and a heartwarming one!

The Incomplete City assignment is about simulating the negotiation and low-grade conflict that makes cities into organic agglomerations that even the most forward thinking planning cannot fully predict. Students begin by making villages for specific user group personas and then over time they combine those villages in successive waves of growth until it’s one big city. This year the final collage featured a poly-centric network of parks and plazas dedicated to the city’s different user groups, such as ‘Wheelertown’ serving wheelchair users and skaters alike.

As has happened in the past, this time there was a healthy dose of easter eggs were embedded in the imagery of wall, as well as some evidence of how strongly suburbia looms in the collective imagination of college students in America. The Detroit region’s fanciest palace of capitalism—Somerset Collection mall—even made an appearance, and it stopped me in my tracks.

One of the recurring themes that we wrestle with in this exercise and across the curriculum is how to help students develop their abilities of discernment. From my own positionality of being a Farm Boy Who Moved to the City as Soon as he Could and Never Looked Back, I will come right out and say that to me suburban shopping malls are among the last of urban gestures I would reach for when thinking about how to locate commerce in a city if I have any choice in the matter. So while I must admit that I had a negative reaction to Somerset gracing the Incomplete City this year, the thing that actually gave me pause was noticing myself being a little grossed out by the presence of a mall in what could have otherwise been a utopia and then using that emotion to split the question of judgement into different facets, like a beam of sunlight through a prism.

These are:

Analytical discernment: because we want urban technology students to create meaningful and successful products, services, businesses, and policies in the world after they graduate, we expect that our students can engage in an analysis of the economic, political, and spatial questions of a thing in a place. We want them to be able to ask: based on this place, utilized by these people, is this thing appropriate? That’s the basic question that we ask again and again in our design studios as a way to evaluate the creative outputs that they come up with. Since the Incomplete City is in their first year, the role of analytical discernment in this exercise is admittedly at the beginner level.

Aesthetic discernment: because we want urban technology students to be inventive and creative, the role of aesthetics for our students is to have an informed sense of what they value, and therefore be able to align their concrete imagination with those values. We want them to be able to say: I think this needs to happen, and it needs to be like this for it to be most effective. The “like this” includes things like tone, character, scale, proportion, legibility, and the ineffable “smell” or “flavor”—those gooey aspects of design decisions that contribute to whether an idea is culturally adept or out of touch.

Intuitive discernment: because we want urban technology students to be biased toward action, the role of intuition is to be able to keep going. To move. If I’m honest with myself, this is not really a separate category of judgement so much as an amalgam of the analytical and the aesthetic, sped up. It’s the gut check of “are we headed in the right direction?,” and because of that it’s absolutely the hardest form of discernment to develop. Intuition is fed by experience.

Across all three facets of discernment described above, the function of taste is to help students separate good enough from truly special and this is important because I want our students to be aspirational, for themselves and for those who experience or become part of the things they create. Our students should make technology that enables the life they want, not the life they think is merely possible. I suppose I want them to be dissatisfied with the status quo. Maybe they’re not mall-haters like me, but there should be something they aspire to that’s indescribable using the references of today. If our students are truly in touch with what they value and find aspirational, then AI cannot tell them what is ‘good,’ and that’s a good thing.

We support the development of our students’ sense of taste by expanding their set of references for urban space and technologies. For instance, when we take them to the arboretum in Ann Arbor or Lafayette Park in Detroit, these are examples of truly special places created in response to unique geospatial, historical, and human conditions. They’re not special in uncomplicated ways by any means, and grappling with compromised greatness is part of the exercise. This teaching is also uniquely important to the life stage of our students as undergraduates. While I expect that every one of them has a favorite park when they enter college, I hope that by the time they leave they have discovered a new favorite, even if it’s something they’ve only read about or heard about from a peer. This is an aspect where AI may have some usefulness, such as by suggesting references for students to learn more about (though the smart students will write prompts to the effect of, “give me a list of inspiring parks and leave your florid commentary out of it TYVM”).

Taste as a way of segmenting ideas and enabling aspiration is tricky because it can easily become a Kantian idea of universal aesthetics without any means of acknowledging difference. On the contrary, the reason I want students to develop a sense of taste is so that they know themselves better. Our goal needs to be cultivating the ability of our students to know what they like well enough to also know when the thing they like is not the right thing for a given place, or community, or moment in time. Call it the Somerset Principle: Many things have a time and a place in which they are appropriate, but no things are appropriate in all times and places.

As I think about the fall semester and my own teaching, my very personal take is that coding and writing, and even analyzing and “reasoning” are cheap in an era of widespread machine intelligence. In response, I’m excited to explore the ways in which we support students’ development of self-awareness and the cultivation of an informed sense of taste. These are the fuel of human decision making. When we dwell on appropriateness, machines may have suggestions but they do not have answers.

These weeks: Have a great summer, readers. See you in September! 🏃