It’s Time To Prototype
Why I’m starting a Lab for Imaginative Prototyping with AI at Harvey Mudd
Thank you for being here. Please consider supporting my work by sharing my writing with a friend or taking out a paid subscription.
Prototyping is second to none as a way of exploring ideas. At root, that’s because a prototype is an embodied question. This definition, which I originally heard from Scott Witthoft, explains the power of prototyping. A prototype is not just about the artifact being created; it’s about the intention with which it’s being created.
It’s not just about the thing, it’s about the question behind it.
As we contemplate the ever-evolving frontier of AI, both in education and beyond, we are in desperate need of more prototypers. We need more people who are building things and asking questions of and with them.
To embody a prototyping mindset is to ping pong back and forth between phases of imagination and critique. We need to start by setting aside the impulse to say “no, but” in order to engage with a generative stance of “yes, and.” But we need more than an open, imaginative stance. After initially suspending our impulse to analyze ideas to death in order to create something new, we need to turn the corner and embrace our critical mindset. A full prototyping process is concerned not just with creating, but also with interrogating the artifacts we create. It embraces the need to assess what these prototypes mean and how they might shape us and the world around us.
Chasing Dynamic Equilibrium
It’s easy to find places where one or the other impulse takes hold. For the imaginative side, there is perhaps no place as famous for this kind of thinking as the “move fast and break things” mentality of Silicon Valley. This mentality, often consonant with the Andreesen-Horowitz mantra of “it’s time to build,” represents a certain distortion of the idea of the prototyping mindset. It represents the “if” and the “can” without the “should,” the “ought,” and the “why.”
But there are plenty of places where the critiquing mind also gets more cycles than it deserves. There are many pockets of the AI conversation these days where even the hint of a suggestion that AI could have a potentially beneficial outcome or application is met with immediate resistance and forceful pushback.
The prototyping mindset offers us a way to find the healthy tension between these two ways of seeing the world. It offers us a way to walk and chew gum at the same time, to be open-minded about the potential for AI to make our lives better while simultaneously considering whether it’s worth the cost. To experiment with generative AI tools in our work while keeping in mind the costs that it incurs—both to ourselves, the people around us, and our environment.
The Lab for Imaginative Prototyping with AI
As I continue to work in this space, I’m excited to find pathways to further embrace a prototyping mindset and invite others into cultivating their prototyping mindsets, too. One of the tangible ways I’m doing this is by creating a studio for Imaginative Prototyping with AI at Harvey Mudd.
The core thesis of the Laboratory for Imaginative Prototyping with AI (IPAI—”eye pi”) is that the best way to explore the promise and peril of AI is to build prototypes with it alongside others who are like-minded in their desire to couple imaginative thinking with thoughtful critique. I recently put out a call for students at the Claremont Colleges to apply to join as fellows next semester. We’ll do four main things together:
Solicit
Build
Critique
Share
1. Solicit
As IPAI fellows, students will apply a prototyping mindset to building with AI. We’ll start by soliciting projects from the community and designing prototypes to address them. As we build, we’ll embrace all the tools of the engineering design process and the broader class of design thinking. We’ll ask questions and for each project, design a one-page brief summarizing the problem, the core question, and the main indications of success.
2. Build
Then, once we have this, we’ll build, leveraging the full set of new AI tools at our disposal. The lab will provide the fellows with access to both the physical resources and AI infrastructure needed to rapidly build. One of the most exciting aspects of our current moment is how fast and easy it is to build. Never has it been faster to go from zero to one and ship a new digital tool or app.
3. Critique
After we have an initial artifact built, we’ll enter the crit phase. One of the formative experiences I had during my PhD was taking a night class in graphic design at Art Center in Pasadena. I felt like a fish out of water. Never before had I been in a classroom where class began by pinning your work to the wall and walking around as a class to offer insights, feedback, and critiques.
That’s the vibe I’ll be shooting for in the critique phase of our prototyping process. We’ll put our AI prototypes on the metaphorical wall and critique them. In addition to offering thoughts from our own personal reflections and vantage points, we’ll consider how the thinkers that I have been inspired by and write about often here—thinkers like Neil Postman, Ursula Franklin, Ivan Illich, Jacques Ellul, Marshall McLuhan, and others—might critique and offer suggestions about the prototypes that we’ve built.
4. Share
After the critique, we’ll share out the results of our discussion. We’ll aim to highlight both the technical details of what we’ve built and our analysis of the societal impact. Our goal will be to write these in a way that will be accessible to the general public rather than aiming for a particular academic community. We’ll experiment with a variety of formats from blogs posts, to YouTube videos, and podcast conversations.
How You Can Get Involved
I would be delighted if you are interested in this idea and want to support it. Here are a few ways you can get involved.
Are you a student? If you’re a 5Cs student and you’re interested in this work, apply to be an IPAI fellow. Applications are posted through the HMC Undergraduate Research Opportunities site and are open until December 1, 2025.
Do you have a project? If you have an idea for a project that you think could benefit from an AI prototype, please submit it or email me directly!
Does your company have a project? In addition to working on these projects through the spring semester, I’m actively recruiting industry partners to sponsor projects both during the spring semester and over the summer months. I’m hoping to line up each of the IPAI Fellows with internship opportunities for the summer, where they will embed in companies in small teams and stay connected with me for mentorship and resources throughout. If you or someone you know is interested in partnering on a project or hiring IPAI Fellows as interns, please reach out!
Do you just want to follow along? If you’re simply interested in following along and keeping up with what’s going on, stay tuned here. I’ll be sharing more about the work in progress over the coming months, and will also be directly distributing updates once we get into the rhythm. If you’re interested in being copied on those email updates, fill out the form and select that option.
Got a thought? Leave a comment below.
Reading Recommendations
Interesting write-up from Anthropic on Claude Skills. They’re basically specialized prompts coupled with scripts and resources. Seems like a natural next step in the journey of finding the tip of the spear where LLMs are maximally useful. I still have a bias toward my mental model of LLMs as glue.
Thumbs up on this tutorial from
on setting up your Claude Code status line. Simple quality of life improvement!This post from
about her journey of exploration of AI is excellent. As always, Elizabeth is asking good questions and writing in a winsome, approachable, and provocative way.If I think LLMs are likely to de-form me rather than form me into the kind of person I want to be, the kind of person the world needs, I should be a running a mile in the other direction. What do I do with the fact that this LLM, working on very little data, in fact helped me live up to my values?
I also enjoyed this post from
on her exploration of “Measured AI” over on her blog. I stumbled on it because she linked back to my e-bike for the mind post, saying, “When you use an e-bike for the mind, you don’t build any pedaling muscles.”Indeed, an e-bike for the mind can be used without building any muscles (i.e., if you only use the throttle). But the nuance and complexity I was trying to draw out in that post is that this is not the only way to use an e-bike. Likewise, it is not the only way to use AI. What makes things interesting is that an e-bike can in fact be used to help you build more pedaling muscles than you could otherwise, for example, by enabling you to bike when you might otherwise need to drive.
The Book Nook
Very close to being finished with this excellent read from Postman. More to come later, but very timely and appropriate for many of our current questions around AI and education.
The Professor Is In
I’m waiting for final confirmation on whether the Lab for Imaginative Prototyping with AI will be funded as part of Mudd’s Innovation Accelerator, but the early indications are positive. Stay tuned for more!
Leisure Line
Hard to beat Porto’s Bakery. Went for a little Saturday morning outing to the West Covina location (which is amazing). Somehow I managed to get a picture of this pumpkin muffin before it quickly disappeared.
Still Life
Snapped a photo of this beautiful rainbow while driving around town this weekend in the rain. If you look closely, you can even see the (very faint!) double rainbow over the top.
Love this project. Can't wait to read about how it goes.
Here is a question: to what extent do you see this work as related to cybernetics of the sort that Dan Davies writes about: the Ashby/Beer tradition coming out of The Ratio Club, a tradition that includes Gregory Bateson and the Norbert Wiener who wrote "A Scientist Rebels." I am thinking about what words to use when we try to imagine alternatives to the optimization mindset that distorts our understanding of what large language models can and should do.
Prototyping –"The simplest way to find out is to make the thing and see," as Ashby put it–is importantly different from "move fast and break things" in order to dominate markets and become trillionaires. Introducing "ought" and "why" into the design process feels crucial, though keeping it front and center as ideas get commercialized is the harder and more important work.
Given how tangled and amorphous, and largely ignored by the academy, the term "cybernetics" has become, it is completely fair to pass on this question. I'm interested in examples of what I think of as cybernetics happening on college campuses, and wondering what you think of the label.