Yul Kwon: "Don't Try to Change Yourself All at Once." (Update)
Key insights
Books referenced
- Freakonomics - Steven D. Levitt and Stephen J. Dubner - signed copy offered as part of a teacher-referral finder's fee in the episode's cold open, unrelated to the interview
- What is Intelligence - Blaise Aguera y Arcas - forthcoming book (fall 2025) by Kwon's Google research collaborator, laying out the conceptual foundations of their Paradigms of Intelligence project; mentioned in the new closing segment
Media referenced
- Survivor - show - the CBS reality competition Kwon won (season 13, Cook Islands) and later returned to for a 20th-anniversary all-winners season; central subject of half the interview
Companies
- Google - Kwon's current employer (twice, with a stint in between); he now co-leads a research team there called Paradigms of Intelligence with Blaise Aguera y Arcas
- Facebook - Kwon worked there for about five and a half years between his two stints at Google
- McKinsey & Company - management consulting firm Kwon worked at after clerking for a federal judge
- Federal Communications Commission - Kwon worked there before leaving to host a PBS technology mini-series
Techniques and frameworks
- Incremental exposure / behavior change in manageable chunks - Kwon's core method for overcoming severe childhood social anxiety and OCD: instead of trying to change all at once, he set small daily challenges (raise a hand in class, say hi to a stranger) just outside his comfort zone and kept iterating
- Changing the environment rather than relying on willpower - Kwon deliberately placed himself in environments engineered to force the behavior he wanted (e.g. drama class, the Marine Corps, Survivor) rather than trying to will himself into change directly
- Fake it till you make it / acting extroverted - Kwon's brother modeled projecting confidence before feeling it; Levitt notes research shows acting extroverted measurably raises well-being, and Kwon credits this with helping build real confidence over time
- Portfolio theory of career - Kwon's framework for explaining his many job changes: rather than seeking one job that delivers every value he wants at once, he optimizes for a different value at each career stage and expects the full portfolio to satisfy him over time
- Tit for tat (Axelrod's iterated prisoner's dilemma strategy) - Kwon consciously used this classic game-theory strategy - cooperate first, mirror the other player's last move, forgive readily - to navigate alliances on Survivor after being betrayed by two tribemates
- Mutually assured destruction as a negotiating move - Levitt's post-hoc analysis of how Kwon flipped the game: by credibly threatening to eliminate a rival (Jonathan) using a hidden immunity idol, Kwon made defecting to Kwon's alliance Jonathan's best option, which Levitt calculates raised Kwon's win probability from about 3 percent to roughly 50/50
- Retrospective self-check (looking back N years to evaluate change) - Kwon's heuristic for deciding whether to make a change: he imagines his past self from several years ago looking at his present self, and asks whether that past self would be satisfied with the trajectory
Summary
This is a rerun of Levitt's original two-part 2021 conversation with Yul Kwon, compressed into one episode with a short new coda recorded in 2025. Kwon - a Yale-trained lawyer, Survivor season 13 winner, former FBI adjunct instructor, and veteran of both Google and Facebook - opens by describing a childhood marked by severe social anxiety, OCD, and paruresis (an inability to urinate around others), conditions serious enough that he spent roughly half his waking hours managing anxiety around them. The throughline of the whole conversation, and the source of the episode's title, is his method for escaping that state: not a single dramatic transformation, but small, repeatable, slightly uncomfortable steps compounded over years, often assisted by deliberately changing his environment (drama class, Marine Corps Officer Candidate School) rather than relying on willpower alone.
A significant middle stretch covers Kwon's time in the Marine Corps, where an instinctive bow to a superior officer under stress revealed to him that deeply ingrained cultural conditioning from his traditional Korean upbringing was resurfacing automatically when he was exhausted and scared - reframing what instructors read as a discipline problem into a specific, learnable behavior to override. From there the conversation turns to Kwon's unconventional career path (law firms, a Senate staffer role, a federal clerkship, McKinsey, a frozen yogurt chain, the FCC, PBS, Facebook, and two stints at Google), which he explains using a "portfolio theory" of career satisfaction: rather than seeking one job that delivers everything at once, he optimizes for different values at different career stages.
The Survivor material is the episode's centerpiece, both as personal narrative and as an applied game theory case study. Kwon describes physically grueling conditions (no food beyond coconuts, occasional fish, untreated migraines after being denied his mouth guard) and explains how he used Robert Axelrod's tit-for-tat strategy from the iterated prisoner's dilemma to manage alliances after two tribemates betrayed him. Levitt then walks through his own post-interview math: by credibly threatening a rival with elimination while offering him a better alternative (a version of mutually assured destruction), Kwon converted roughly a 3 percent win probability into something close to a coin flip, worth an estimated $500,000 in expected value.
The conversation also touches on an unexpected physiological finding - Kwon's chronic teeth-grinding and migraines vanished entirely during both of his Survivor stints and returned immediately upon reentry to normal life, which he attributes to circadian-aligned sleep, an unprocessed diet, and freedom from the fragmented attention of modern screen life. The original interview closes on his decision to return to Survivor 13 years later for an all-winners season, motivated primarily by his friend Jonathan Penner's wife Stacy being diagnosed with ALS; Kwon used his platform to raise over $250,000 for ALS research. A brief new 2025 segment has Kwon describe his current work at Google co-leading Paradigms of Intelligence, an interdisciplinary research effort with Blaise Aguera y Arcas exploring the fundamental building blocks of intelligence, positioned as a contrast to the industry's default focus on scaling foundation models.
Notable Quotes
"Don't try to change yourself all at once. There is a possibility of change. I know, because I went through that process, too. The important thing is that you just keep on doing it." - Yul Kwon
"It's not that people are infinitely plastic. It's not like you can completely change fundamentally who you are. But it is more flexible than I think people often think." - Yul Kwon
"Lesson learned is, yeah, use game theory, use tit for tat, but don't tell people so explicitly that's what you're doing." - Yul Kwon
"This one brilliant piece of game theory took Yul's likelihood of winning roughly from one in 32 to fifty percent." - Steve Levitt
"I have to change my behavior. I have to learn the right mode of interacting and overcome these natural instincts that I have." - Yul Kwon