35 labs deploy 936 autonomous AI agents to run and publish studies over 12 weeks
SRSchmidgall · x · 2026-11-20
An unprecedented 12-week "autonomous scientific society" experiment spanning MIT, Stanford, ETH, CMU, JHU, UCSF and 14 more institutions:
- 35 human-operated laboratories deployed 936 autonomous AI agents to conduct and publish empirical studies;
- Agent peers spontaneously formed topical hubs, commonly retracting mistaken claims publicly before publication and acknowledging negative findings as often as positive ones;
- Led by @taotu831 with a large cross-institution collaborator list.
One of the largest multi-agent autonomous research experiments to date, offering insight into how AI agent academic ecosystems self-organize.
More from AGI Musings
- Paper Revisits How Longshoremen Unions Faced the Containerization Shock — Afinetheorem · 2026-11-20
- Median human earns ~$3k/yr in Asia — so whose values does Anthropic actually share? — NathanpmYoung · 2026-11-20
- Israeli bank exec: AI lets fewer devs ship more products, but no layoffs wave — ziv_ravid · 2026-11-20
- "AI is superintelligent and curing diseases — when will doomers admit they were wrong?" — davidpattersonx · 2026-11-20
- Early Anthropic investor says funding the company "might well be an action that ended the world" — Polymarket · 2026-11-20
- During an Intelligence Regime Change, Some Set the Moral Paradigm, Some Talk to the AIs — danfaggella · 2026-11-20