NYT cinéma vérité follows mathematician Mike Winer racing to build mathematical safeguards for AI control

stevenstrogatz · x · 2026-11-22

NYT reporter Eli Saslow published a long-form vérité profile, "Looking for the Formula to Stop an Apocalypse," following 30-year-old mathematician Mike Winer and colleagues trying to build mathematical safeguards keeping AI under human control.

Details: Winer eats all three meals at his Berkeley office and wakes up with unfinished equations in his head; briefing visiting UC Berkeley mathematicians, he admits "A.I. is progressing much faster than we thought... We could be heading toward a catastrophe." Visiting postdoc Reese Pathak notes these systems solve in hours problems he expected to work on for years.

Mathematician Steven Strogatz shared it as a rare, cinéma-vérité glimpse of what AI-safety math work actually looks and feels like — whiteboards, math books, and palpable dread.

Original post →

More from AGI Musings

AGI Musings channel →