✳flâneur — a map of the web's best reading
We spent 2 hours working in the future - METR
metr.org · 2,721 words · saved by 4 readers
Thomas Kwa describes a tabletop exercise where METR researchers simulated having access to ~200-hour time horizon AIs.
We spent 2 hours working in the future - METR Our Work Research Notes Updates Risk Assessment About Donate Careers Search --> Our Work Research Notes Updates Risk Assessment About Donate Careers Menu × Research note: We spent 2 hours working in the future CONTRIBUTORS Thomas Kwa and Tom Cunningham DATE March 19, 2026 SHARE Copy Link Citation BibTeX Citation × @misc { metr-2026-org-uplift-game , title = {We spent 2 hours working in the future} , author = {Thomas Kwa, Tom Cunningham} , howpublished = {\url{https://metr.org/notes/2026-03-19-org-uplift-game/}} , year = {2026} , month =
Explore this link on the map →saved by
related reading
- Clarifying and predicting AGI — LessWronglesswrong.com
- My picture of the present in AI — LessWronglesswrong.com
- Measuring AI Ability to Complete Long Tasks - METRmetr.org
- AI 2027ai-2027.com
- I underestimated AI capabilities (again) - by Ajeya Cotraplanned-obsolescence.org
- An Apple-Picking Model of AI R&D | Tom Cunningham – Tom Cunninghamtecunningham.github.io
- Evidence on AI R&D Progress from NanoGPT - METRmetr.org
- When AI builds itself \ Anthropicanthropic.com
- The machines are fine. I'm worried about us.ergosphere.blog
- A (Slightly) Mechanistic Theory for Exponentially Increasing AI Time Horizons? — LessWronglesswrong.com
- After Automation | Everyevery.to
- AI 2027ai-2027.com