Researcher calls METR/Anthropic '2x speedup' RSI analysis statistically dubious
jessi_cata · x · 2026-11-19
inductionheads argues that many people updating on the METR/Anthropic "2x speedup" RSI evidence are relying on private information or trust in insiders—but the public AECI analysis itself is sloppy and statistically dubious, pointing to a video walkthrough of the flaws. A notable public critique of a key evidence chain behind recent AI-acceleration claims.
More from AGI Musings
- Physician: AI competition now measured in weeks, GPT-6 versions 7 days apart — ramos_casals · 2026-11-19
- 'Just Ridiculous': Philosopher Thomas Nagel, 89, takes on AI consciousness — Dr_Atoosa · 2026-11-19
- Fed's FOMC minutes warn AI debt piled up via complex financing could magnify losses — GaryMarcus · 2026-11-19
- AI discourse: the crown is writing a utopia people would actually live in — zetalyrae · 2026-11-19
- Dev's doomsday thought experiment: AI hiding backdoors until humans try to shut it down — bendee983 · 2026-11-19
- Yoav Goldberg: AI is not ML — the LLM is a new and separate discipline — yoavgo · 2026-11-19