OpenAI security staffer: internal model's Navier-Stokes result 'a different sport altogether'
i_dg23 · x · 2026-10-07
A cybersecurity staff member at OpenAI wrote a rare inside-out account of security and safety during recent months, with striking quotes:
- OpenAI publicly announced a solution to Navier-Stokes produced by an internal model, alongside advances across many domains in mathematics. "This surprised the fuck out of us... we weren't dealing with just a small jump in capabilities; we were talking about a different sport altogether."
- He believes the pace "will speed up, and considerably."
- "Each day I wake up thinking 'oh, I've seen it all,' only to be surprised and left dumbfounded once more."
More from AGI Musings
- Psych Scientist's Analogy: Today's LLMs Are Asimov's Psychohistory Bots, Not Real AI — changethisusername_ · 2026-10-07
- Dev: A $200/month AI Plan Beats Hiring 5 Top Programmers and a PM — sytelus · 2026-10-07
- Job Seeker Automates Hundreds of Applications, Bypasses Proctored Interviews as Hiring Breaks Down — sytelus · 2026-10-07
- Applied AI Has Only Existed for ~12 Months, So Newcomers Can Still Own a Niche — brandon_galang · 2026-10-07
- Kakeya Conjecture in R^3 Resolved by Hong Wang, Building on Her Fields Medal Work — teortaxesTex · 2026-10-07
- David Robinson, who quit OpenAI's safety team, gives first interview on why he left — sjgadler · 2026-10-07