All podcasts / People I (Mostly) Admire / Summary

172. A New Kind of University

2025-12-06 - source - Read full transcript
Steve Levitt (host)Michael CrowMorgan Levey

Key insights

University prestige built on selectivity is optimizing the wrong variable.
Crow argues that measuring institutional quality by how many applicants are rejected values inputs, not outputs, and that scaling the entire higher-ed system around exclusivity necessarily creates a second tier of 'second-class' institutions serving everyone else, which produces low graduation rates and discouraged faculty system-wide.
access-vs-selectivity
A.S.U. explicitly measures itself by inclusion rather than exclusion, using historical admission standards from already-successful public research universities.
A.S.U.'s charter commits to admitting every student who meets a baseline standard (calibrated against what schools like UC Berkeley or Michigan used when they were already successful), not to minimizing the acceptance rate. About half of incoming freshmen have B averages; the university then adapts teaching and technology to serve that broader, less pre-selected population, rather than screening it out.
access-vs-selectivity
Colonial-college-style institutions (the Ivy League and similar elite schools) are structurally unscalable and cannot alone carry the country's higher-education needs.
Crow traces Ivy League design to the British Oxford/Cambridge honors-college model layered with German-derived engineering and science programs, a structure built for small, elite cohorts. He argues the U.S., approaching 350 million people, needs institutions engineered for scale and speed instead, which the traditional model cannot provide.
higher-ed-institutional-design
A.S.U.'s transformation was a case of disruption from inside an incumbent institution, not from a scrappy outside entrant.
Clayton Christensen reportedly told Crow he had never seen disruptive innovation happen from the core of an existing large enterprise rather than the periphery. Crow attributes this to higher education's unusual structure, where centuries-old institutions from every prior 'wave' still coexist (unlike the private sector, where Schumpeterian creative destruction replaces old players), so building something genuinely new required reshaping an existing university rather than replacing it.
higher-ed-institutional-design
Reorganizing around what faculty actually do, rather than legacy department names, unlocked growth that rankings-chasing could not.
Crow describes eliminating 85 schools/departments and creating 40+ new ones; the geology department example (previously about 20 majors) was dissolved into a new School of Earth and Space Exploration alongside astronomy and astrobiology faculty, after which enrollment in the same underlying field grew to more than 500 majors, because the framing matched how students and faculty actually understood the work.
leadership-and-culture-change
Crow's leadership playbook centers on purpose, faculty empowerment, and constant re-articulation of mission rather than top-down mandates.
He describes four practices: giving faculty a shared sense of institutional purpose (so they aren't just mercenaries chasing the next offer), empowering faculty as designers rather than resource-fighting bureaucrats, direct personal engagement (walking campus daily, talking to whoever approaches him), and continually re-explaining the university's societal role to keep people from sliding into institutional negativity.
leadership-and-culture-change
Story-driven virtual reality dramatically improved outcomes in historically difficult intro science courses by adding emotional engagement that abstract lecture teaching lacks.
A.S.U. partnered with former DreamWorks producer Walter Parkes to build 'Dreamscape Learn,' immersive VR labs (e.g., an orbiting alien zoo for intro biology, a Himalayan chemistry rescue scenario). Crow reports roughly a 40 percent improvement in learning outcomes and an average two-grade-level improvement in biology scores, attributing the gain to storytelling creating an emotional stake that traditional labs never provide.
higher-ed-institutional-design
Scaling engineering and STEM access does not require lowering the bar, it requires building infrastructure to bring underprepared students up to it.
A.S.U. has 33,000 engineering students and now graduates over 7,000 engineers a year, up from about 900 a few years earlier, by admitting anyone who cleared the general admission bar into engineering and then building learning tools to develop math skills rather than pre-screening for a 1500 SAT score, which only about 2 percent of students have.
access-vs-selectivity
AI's value in education depends entirely on whether the student is already engaged; it does nothing to solve the disengagement problem itself and can make it worse.
Levitt pushes Crow on this directly: an engaged learner benefits enormously from AI tools, but an unengaged learner can use the same tools to skate by without learning anything (e.g., having AI write a required essay). Crow agrees this requires redesigning what is asked of students, not just adding AI to old assignment formats, since simple assignments just get outsourced to the model wholesale.
ai-in-education
Crow wants A.S.U. to be a producer of curated, bounded AI tools for students, not merely a consumer of general-purpose AI.
A.S.U. is building fyi.ai with Will.i.am, a small language model conceived as a personalized, agentic tutor with defined boundary conditions, in contrast to open-ended large language models. Crow's framing: knowledge delivered through education has to be curated, contextual, and verifiable, which requires purpose-built tools rather than handing students unconstrained general AI.
ai-in-education
People I (Mostly) Admire is ending after five years and over 170 episodes, with Levitt moving to a regular guest-host role on Freakonomics Radio.
Producer Morgan Levey and Levitt announce the show's close in the final segment. Levitt says he wants to preempt the moment work start feeling like an obligation rather than doing it early, and that he wants a format (Freakonomics Radio's policy-oriented reporting) better suited to driving concrete change than long-form interviews. The final PIMA episode two weeks later will flip the format, with Stephen Dubner interviewing Levitt.
podcast-transition

Books referenced

Companies

Techniques and frameworks

Summary

This episode profiles Michael Crow, president of Arizona State University since 2002, whose tenure has taken A.S.U. from a middling, "party school" reputation to eleven consecutive years as U.S. News's most innovative university in the country. Crow's central argument is that American higher education inherited a fundamentally British, small-elite structure from its colonial colleges, one that scales poorly and forces the entire system into a zero-sum selectivity race. Rather than compete on exclusivity, Crow built what he calls the "New American University" model: an institution designed to be simultaneously research-excellent, radically accessible (admitting by objective standard rather than by scarcity), and directly accountable for measurable community impact. He frames this explicitly against elite peers like his former employer Columbia, arguing that a system organized entirely around rejection rates cannot serve a country of 350 million people and inevitably produces a second tier of institutions that get treated as lesser.

Much of the conversation covers how Crow operationalized that philosophy structurally. A.S.U.'s 2013 charter commits the university to being measured by whom it includes rather than excludes, and Crow eliminated 85 legacy schools and departments in favor of more than 40 new ones organized around what faculty and fields actually are, not inherited naming conventions; a shrinking geology department became a School of Earth and Space Exploration and grew from roughly 20 majors to more than 500. Engineering scaled from about 900 graduates a year to over 7,000 by admitting broadly and then building instructional infrastructure to bring students up to speed, rather than screening for pre-existing preparation. Crow also discusses his leadership approach: giving faculty a sense of institutional purpose, empowering them as designers of their own programs, staying directly and personally engaged (he walks campus most mornings and talks to whoever approaches), and constantly re-articulating the university's societal role to counter the negativity he sees plaguing many institutions.

A significant stretch of the interview covers A.S.U.'s use of technology in teaching, especially "Dreamscape Learn," a virtual-reality curriculum Crow built with former DreamWorks producer Walter Parkes that replaces abstract lecture-based intro biology and chemistry with immersive, story-driven simulations (an orbiting alien zoo for biology, a Himalayan rescue mission for chemistry). Crow reports roughly a 40 percent improvement in learning outcomes, crediting the emotional engagement storytelling provides that dry lab instruction never did. The conversation turns to AI in education, where Levitt presses Crow on a real tension: AI tools are transformative for an already-engaged learner but can just as easily let a disengaged student skip learning entirely by having the model do the work. Crow's answer is that assignments and pedagogy need to be redesigned around this reality, and that A.S.U. should be a producer of bounded, curated AI tools (illustrated by its fyi.ai project with Will.i.am) rather than a passive consumer of general-purpose models.

A personal thread runs through the back half: Crow's father was a career military man and his sole parent after his mother's early death, instilling a work ethic and an adaptability Crow says came from moving 21 times and attending 17 schools before college. He also discusses turning down MIT to throw the javelin at Iowa State, and reflects candidly on succession, arguing his eventual successor inherits an already-transformed institution and therefore has a different job than he did: evolving the model rather than building it.

The episode closes with a significant announcement: after five years and more than 170 episodes, People I (Mostly) Admire is ending. Producer Morgan Levey and Levitt explain the decision, with Levitt saying he wants to step away before the show starts to feel like an obligation, and that he's drawn toward Freakonomics Radio's more policy-oriented format as a better vehicle for pushing concrete change than long-form interviews. Levitt will become a regular guest host of Freakonomics Radio starting in 2026, and the final PIMA episode two weeks later flips the format, with Stephen Dubner interviewing Levitt. In the closing producer segment, Levitt and Levey also revisit a prior Levitt episode on paying Brazil to preserve the Amazon rainforest, discussing the newly launched Tropical Forest Forever Facility (TFFF) and Levitt's critique that its per-hectare destruction penalty is likely still priced too low to outcompete illegal ranching.

Notable Quotes

"We compressed a couple hundred years of university evolution into 20 years... rather than taking 200 years or 300 years to build Harvard, we're taking 20 years to build this new A.S.U." - Michael Crow

"The notion of selectivity being the functional measurement of the quality of the university is literally absurd. You're measuring the quality of the institution on its inputs, not on its outputs." - Michael Crow

"If the questions are so simple or whatever that the A.I. becomes the substitute for the student, then we failed as the teacher." - Michael Crow

"It's really hard to do... universities are culturally driven. They're very filiopietistic, they adore tradition." - Michael Crow

"I certainly think it makes sense for me to be a role model and, at least for this one time in my life, to quit a little bit early instead of way too late like I usually do and like most people do." - Steve Levitt