Move 37 Moments Are Going Mainstream: Goldman Sachs, KPMG, Dropbox, and More
TL;DR
“Move 37 moments” are spreading beyond Go into white-collar work — the hosts define these as the instant a professional realizes AI can match or beat their expertise, borrowing the term from AlphaGo’s famous move against Lee Sedol in 2016.
The evidence now spans coding, finance, auditing, science, math, and writing — they cite Sam Altman on Codex proposing better features than OpenAI’s own team, Goldman Sachs deploying Claude for trade accounting, KPMG pushing audit fee cuts, and an astrophysicist estimating AI already shows about 90% of the intellectual capability in his field.
Paul Roetzer’s core point is the human shock, not just the technical milestone — he says people fixate on AlphaGo as a breakthrough in machine capability, but the real story is what happened to Lee Sedol emotionally when he realized the machine was simply better.
Denial is the dangerous default reaction — Roetzer says he keeps hearing versions of “it could never do what I do,” and argues that pretending this shift isn’t coming “isn’t going to do anybody any good.”
The emotional split is becoming part of AI literacy — Roetzer describes sitting in a bar after a dads’ basketball tournament, talking AI with one informed friend while looking around at hundreds of people who were “blissfully unaware,” living in what felt like a parallel universe.
Once someone has their own Move 37 moment, their worldview changes — his argument is that after the realization that AI can do your job, “everything is different from that moment on,” including how you think about careers, business, education, and the future.
The Breakdown
From AlphaGo to the office
The segment opens with the hosts framing a bigger pattern: more professionals are having “Move 37 moments,” the point where AI demonstrates it can meet or surpass expert human judgment. They anchor it in AlphaGo’s move 37 against Lee Sedol in 2016 — not just as a technical feat, but as the moment the world’s best player saw the machine had gone somewhere he hadn’t.
The receipts: coding, accounting, audits, math, writing
They rattle off examples fast, and that speed is part of the point: this is no longer one weird anecdote. Sam Altman said OpenAI’s Codex suggested features better than his own team’s ideas; Dropbox’s former CTO said he’ll never write code by hand again; Goldman Sachs is deploying Claude for trade accounting; KPMG is using AI pressure to cut audit fees; and David Kipping said AI is already at roughly 90% of the intellectual capability he sees in astrophysics. They also mention a Polish mathematician getting help from GPT-5.4 on a previously resistant problem, Boris Cherny saying coding is “effectively solved,” and a New York Times quiz where 54% of 86,000 participants preferred AI-written passages to famous authors.
Paul’s keynote was really about Lee Sedol’s inner experience
Paul says this exact idea was the premise of his 2025 MAICON keynote: everyone will eventually have their own Lee Sedol moment. His twist was that people talk about AlphaGo as a model milestone, but he wanted the audience to sit with the human side — what it feels like when the thing you’ve mastered is suddenly done better by a machine.
A hard talk because he didn’t want to leave people defeated
He says it was one of the hardest talks he’s ever given, partly because he didn’t even know the ending until about 24 hours before going on stage. He wanted to give people that “gut punch” through clips from the AlphaGo documentary, but not end in despair; the landing was that people still have agency and can choose to use these systems as tools that create new abilities and new ways to think about business and careers.
The real-world problem: most people still think this can’t touch them
Roetzer says he constantly hears some version of, “Yeah, it’s good, but it could never do what I do.” His response is blunt: that line of thinking is probably not going to end well, and acting like this isn’t coming won’t help anyone.
The bar after the dads’ basketball tournament
The most human moment in the clip is a story from a weekend dads’ basketball tournament, where Paul and a friend who experiments with OpenClaw are talking AI at a bar. They look around at hundreds of couples and have that surreal feeling that they’re living in a parallel universe — everyone else has careers, kids, college bills, and no idea what’s about to hit.
Knowing too much, and not being able to turn it off
Paul says there’s even a part of him that envies the ignorance, because once you understand where this is headed, you can’t unsee it. That’s why these moments matter so much: after your own Move 37 realization, you don’t just update your opinion of a tool — you start seeing jobs, education, and the future differently from that point on.