Greenblatt: RSI report is a big update toward ASI in 1-2 years; Anthropic internal Model 4 rumored at Agent-3 level
haider1 · x · 2026-11-07
Former Anthropic researcher Ryan Greenblatt says a report on AI R&D automation (recursive self-improvement) is a big update toward ASI within 1-2 years: RSI is currently speeding up AI progress by 2x, close to the AI-2027 scenario, and he admits he was seriously wrong to expect only 1.15-1.3x in 2026. He cautions that much of the underlying evidence is redacted. Haider adds an unverified claim that Anthropic's internal Model 4 (finished training Sept-Oct 2026) is roughly equivalent to Agent-3 from the AI-2027 report. The attached images include an AI-2027 excerpt ('July 2027: The Cheap Remote Worker') and an Epoch AI chart showing LLM inference prices falling 9x-900x per year.
More from AGI Musings
- Engineer warns: sacrificing the present entirely for the future is a dangerous game — KevinNaughtonJr · 2026-11-07
- Amazon reportedly trying to rehire employees laid off in AI-driven cuts — DavidLinthicum · 2026-11-07
- Peter Norvig on what's still worth learning when AI takes over the work — Diyi_Yang · 2026-11-07
- RareDiseaseBench debuts: 35 reconstructed rare disease cases test AI diagnostic agents — kenbwork · 2026-11-07
- Signal president debates AI policy head: should we stop trusting AI agents with our data? — neil_chilson · 2026-11-07
- Survey claims 68% of Americans support pausing AI and banning superintelligence — notkilleveryoneist · 2026-11-07