Pittsburgh, Austin, the Cosmos, and the Local
A report on recent gatherings
Last month I visited two excellent cities– Pittsburgh, PA and Austin, TX – to talk about AI and education with some of the most strategically optimistic people I know. My takeaway is optimism: there is so much exciting work by smart people to retrofit and build anew that the institutions that survive AI disruption will be stronger than many standing today.
I was invited to University of Pittsburgh by Michael Madison, Lisa Parker, and Bridget Keown, conveners of the incredibly forward-thinking Ethics and Economics of Expertise (E3) project, the University in the AI Age and the Institute for Bioethics about AI. I’ve known Mike from his substack, Everything In Between, and I am a fan of his longstanding work on the “knowledge commons,” newly relevant in the AI era. He coordinates PASTA, the Pitt AI Scholar-Teacher Alliance, and directs the Future Law Project at Pitt Law. Mike’s superpower is coalition-assembling, bringing together centers and institutes to think at the micro and macro level about what AI means to mission. His focus is on institutions as sites for producing, storing, and distributing knowledge. He wants to make institutions better.
Pitt is one of the most serious AI adopters in American higher education. Pitt faculty are asking pointed questions about what AI means for higher ed beyond another layer of bureaucracy, which is what I see happening at other universities, like Dartmouth. Yes, Pitt was the first university to secure an institution-wide agreement for Anthropic’s Claude for Education, back in October, but important change will come through faculty innovation. Pitt is clearly a place that wants its faculty to speak up and speak out. It’s incredibly refreshing.
My public presentation, “AI and the Last Mile of Knowledge,” was about the structural weaknesses of higher ed, which got a great writeup (“AI didn’t cause this but AI is revealing it”) by Naomi Weiss, Pitt News reporter. I drew on my thinking about using the Rumsfeld Matrix for higher education, asking how much universities should be focused on “known knowns” versus the quadrants with unknowns in them.
Everyone at Pitt cares about quality teaching. Pitt is known for commitment to excellent mentoring and I saw it everywhere. Faculty tend to remain at Pitt for decades, if not their entire careers. They are committed to undergraduate and graduate education. They are paying attention to whether students are actually learning.
Pitt is likely to see only minor AI disruption compared to other universities because faculty (and department chairs) at Pitt are dedicated to showing what they can do that LLMs can’t do. Everyone was open to the possibility of disruption: short semesters, accelerated curricula, athletes using their new power to demand courses that politicians might want to shut down. The conversation about athlete power to shape their own education was paradigm shifting, for which I thank Bridget Keown and Michael Sawyer particularly. Shout outs to Gayle Rogers, Jeff Aziz, Jane Leibschutz, Arthur Kosowsky, Shelome Gooden, Jessica FitzPatrick, Jonathan Woon, Jen Waldron. Plus an incredible tour of the city by Mike.
This was my first real visit to Pittsburgh and for culture, vibe, and energy, it should be up there with Miami and Austin as new tech hubs. I had a quick lunch with Tyler Cowen, in town for a Bruckner concert. Go to Pittsburgh, people! Go visit the Mayors of Pittsburgh Wikipedia page and read about the city’s leadership, including Mayor David Lawrence, ranked among the ten best mayors in American history. Read about all the bridges. Ask Mike Madison for a tour!
If my trip to Pittsburgh offered the smartest thinking on AI and education from inside a university, my trip to Austin offered some of the smartest thinking on AI and education from outside. I went for two events: a Cosmos Institute gathering and a visit to the GT School, a sister school to Alpha School, in the news for innovative, accelerated learning. The builders at Cosmos and the founders of Alpha, with their eye on new organizations for human flourishing, have the freedom to build from scratch, while universities must retrofit and reform from within.
Brendan McCord, the founder of Cosmos, is a visionary outside the academy seeking to build a new kind of intellectual community. He is a deep thinker and a pragmatist. He graduated from MIT, spent 610 days underwater on a submarine, earned an MBA from Harvard, and has been thinking about AI and how AI can support human flourishing. His concept of philosopher-builder is ideal for this moment. The three central concepts driving Cosmos are truth-seeking, human autonomy, and decentralization. I am 100% on board with the second two. But I prefer “truth modeling.”
In between Cosmos events I gave a talk at GT School, at the invitation of the excellent Pamela Hobart. Pamela Hobart and I have been X friends for years and finally met on Zoom for a conversation last fall that jump started my thinking about acceleration and education. Can you design school that enables acceleration in some subjects but not all, so it is still a “school”? I wrote about it in “the Two-Minute Mile Problem.”
My talk, “Measuring Potential: Dr. Julian Stanley and the Science of Accelerating Gifted Learners,” was about what I learned from my former Johns Hopkins University professor and mentor, who transformed how America identifies and educates gifted students. I received a letter in the mail from Dr. Stanley in 1976, asking about my plans after performing exceptionally well on the math PSAT. This outreach was part of his legacy project: the Study of Mathematically Precocious Youth (SMPY), begun in 1971 with a question: how many gifted children exist in a given population? What are their options? That letter changed the trajectory of my life and I am grateful every day.
Stanley had realized that administering the SAT to twelve- and thirteen-year-olds was a serendipitously efficient screening instrument, capable of sorting millions of children at low cost per capita, to surface talent across the country. Stanley’s research supported the development of a menu of options: acceleration (skipping grades), early entry to college, and popular Center for Talented Youth residential summer programs (CTY). SMPY’s fifty-year follow-up data, the longest-running study of intellectual talent, showed that children who received appropriate challenge reported higher life satisfaction and professional fulfillment decades later. Asked as adults what mattered most about their education, the overwhelming answer was curricula aligned to their readiness, not their age. They wanted to learn at the pace and depth they could handle, among gifted peers.
Pamela and her husband, the excellent Byrne Hobart, gave me a tour of GT School, which is impressive. I would have loved it. The conversation afterward about education with parents and fellow educators, including the excellent Matt Bateman and Matjaž Leonardis, was also high-level and instructive.
The purpose of the Cosmos Institute gathering was to bring together people who care about human flourishing and to orchestrate new conversations (following Chatham House Rules). It was terrific to spend time with philosophers Zena Hitz, founder of the Catherine Project, and Rebecca Lowe. I met the excellent Avantika Mehra and Harry Law, finally, as well as the always impressive Jason Crawford, founder of Roots of Progress. Many good AI language conversations, particularly with Anastasia Uglova, Ivan Vendrov, Peter Danenberg, and Elian McCarron. Too many excellent conversations to list them all. (Is there photo evidence of Rebecca Lowe reading the Declaration of Independence as a break-up letter? In fact there is.)
I met Joe Liemandt, principal of Alpha School, which now operates fifteen locations. I asked Liemandt if he knew of Stanley’s work. Stanley’s approach was of course very different from Alpha School as it is currently headed. Stanley was more interested in life satisfaction than entrepreneurial success. But gifted education is a big tent and can accommodate diverse priorities.
I led a session at Cosmos on “What Kills Courage in Institutions …and How to Build Against It,” for builders building to cultivate independent thinkers. What institutional failure modes should an academy design against? I decided to forego the usual talk about heterodoxy versus political polarization, lack of viewpoint diversity, to ask: what do you know or not know about what is going on above you?
Institutional opacity and thin communication channels are stickier problems than lack of political courage. Most young people aren’t trained to think bureaucratically, particularly in an environment that focuses on “founding” and start-ups as the best and most noble paths for smart young people. So questions of hierarchy, of org charts, memos, reporting layers, chains of command, aren’t a priority. I have been startled throughout my career by middle-aged people having no idea what their boss is worried about, let alone their boss’s boss. These things are alien to far too many intelligent people.
I began with Tae Kim’s excellent 2024 biography of Jenson Huang, The Nvidia Way, which focuses on T5T — “top five things” — a protocol where any employee at any level can email Huang directly with the five most important things they’re working on or concerned about. Huang reads them all and gets access to what’s happening at the edges of the organization without filtering and sanitizing. Tae Kim’s book is brilliant. Yet I have yet to hear about another organization celebrating the adoption of T5T.
Smart young people should understand better how to use their voice in institutions. I noted three other helpful books: Graham Allison’s Essence of Decision, (1971), a study of the Cuban Missile Crisis through three analytical models, about what decisions look like from the outside, depending on where you are in the hierarchy. What information reaches you? What games are people above you playing with each other? Then there’s Ben Horowitz’s The Hard Thing About Hard Things (2014), about running a company when everything is going wrong, making decisions when there is no good option, when you are responsible for the outcome. And of course Albert O Hirschman’s Exit, Voice, and Loyalty(1970), about what to do when the institution you work for declines in quality. You can exit (leave), exercise voice (complain, protest, push for change from within), or remain loyal (stay and accept).
Many people took a photo of this slide and I hope they’ll read some of these books and expand the conversation.
Brendan McCord, Mike, and Pamela all were excellent hosts (thank you!) and the multiple conversations clarified my thinking about how I intend to focus my time. I look around and see how big institutions exert power and gravitational pull even when decaying and declining. Brendan’s questions “what are you building?” and “what are you building for?” seem somewhat less urgent to me than Mike’s “how can we make institutions better?”
At Pitt, Mike Madison has spent years building initiatives that enable the university to examine its own mission in real time, producing, storing, and distributing knowledge in the AI era. It takes labor to encourage an institution to ask the right questions about itself. Mike is building the channels through which information about the institution can reach the people who run it. Keown and Sawyer’s conversation about athlete power was the most paradigm-shifting moment of either trip: the idea that students with real leverage might reshape what a university offers, from the demand side. That is what it looks like when an institution lets its people think out loud about structural change. Mike doesn’t need the books I recommend because he already does the work they describe. Pitt’s leadership has made room for it, which is rare and which is the thing most worth studying.
Courage does not scale. The work of making institutions better is the work of making courage less necessary and I’m going to commit to that. It’s slow work, local and often unglamorous. It happens locally, in coalition meetings and faculty conversations and one invitation at a time. I love it.
You're the educator I wish I'd known.