This is the place you build it.
A fellowship in San Francisco, with video cohorts, for people who want to do more with AI than learn a tool: to change how they think about it, and what they do with it. Three paths. $180 each, with financial assistance on the application.
Judgment is built by grappling with real things, questioning them, and talking them through with others. Each path does that from a different angle. Pick one, or all three.
A working session for people running real AI agents, tools, and loops. Members demo their setups live, say what worked and what broke, and we dig into it together.
Advanced is for people who run agents, pipelines, or loops every day; demos go to full depth. Intermediate is for people who want to do much more than simple prompts.
A deep-reading seminar on the texts today's AI arguments came from: Descartes, Turing, Hobbes, Wiener, Schelling. Short readings, sent ahead. Two hours at the table working out what they say.
No lecture and no expert at the head of the table. A moderator keeps the argument honest; everyone has read the same pages. No technical background needed; the readings are the only prerequisite.
Cases from 120 years of technologies being sped up, redirected, or stopped, and why. Work through the record, then argue where AI's path will be decided.
Built for people who bring a field with them, whatever it is. The history is the shared ground; your expertise is the contribution.
Not in San Francisco? The club's Intermediate cohort already meets by video. A Wednesday-evening video cohort will run for one of the two seminars: the seminar and the cases share the slot, and the one with more applicants on September 22 runs. If you applied to the other, you'll be offered the one that runs. The Advanced club's Wednesday lunch cohort opens once six people ask for it. Times are Pacific. Nobody pays until a cohort is confirmed.
Mixed cohorts, on purpose. Past cohorts have included engineers and product people, AI researchers, UI designers, founders, lawyers, recruiters, executives, and students. What they had in common was the reason they came: to expand how they think, with the tools as the means. That is still the only entry requirement.
Everything else about AI, you can rent. The models, the tools, the agents are the same for you as for everyone else, and they get better whether you do or not. Your judgment is the part that's yours.
There is no course for judgment. It gets built with experience and in company, which is why the method matters as much as the material.
“I was already building with AI tools before the Bell AI Fellowship. What it changed was the approach. I came out running on a system I built myself, and I use it every day now.”Michael Albers · former Chief Product Officer, Yahoo!/AOL
Pick the path, or paths, that fit how you learn. One five-minute form covers all three. Every application is read for one thing: whether you want to think differently, and will bring what you know to the table. Titles don't decide it. Places go to the people who will make the strongest cohort, and a strong cohort is a varied one, and a deep one.
A founder and investor who began as a congressional technology-policy analyst, and a product leader from Yahoo, Visa, and Verizon.
Serial entrepreneur and investor. Founded four venture-backed companies; investor or board member for five IPOs and four unicorns. Analyzed technology policy at the congressional Office of Technology Assessment and charted climate-tech trajectories with the Gigaton Throwdown. Co-founded Bell AI to build the kind of learning community he wishes existed when he was starting out.
LinkedIn →
Product builder who has led $1.6B businesses and products reaching 500M+ users at Yahoo, Visa, and Verizon. Engineer turned product leader, GM, and founder. Co-founded Bell AI because good judgment about AI is built the way she has always worked: hands-on, with other people.
LinkedIn →