ResearchNovember 10, 2026
-17
āļø
AI Researchers: Superintelligence by 2027?
AI researchers now expect superhuman AI researchers could arrive as early as 2027.
#AI timelines#AGI#AI safety#forecasting

š„ What happened
A team led by Daniel Kokotajlo just slashed its timelines: Superhuman AI Researcher (SAR) likely arrives in 2027, not 2029. The trigger is a METR report showing Anthropic now advances AI progress twice as fast as its historical rate, with ~80% of that acceleration driven by AI itself.
š” Why it matters
A 1.7x AI-attributable acceleration isn't a productivity boost ā it means AI is speeding up its own development. The authors now put 35ā65% odds on Top-Expert-Dominating AI (TED-AI) before 2028. Three months ago, they gave less than 4% to the current acceleration being real.
ā” Our take
The authors call for an industry-wide pause ā but whoever pauses first loses. This is no longer an alignment problem; it's a geopolitical race nobody will voluntarily brake.
The title, summary and analysis of this item were produced automatically by an AI system and have not been editorially reviewed. They may contain errors, bias or omissions ā when in doubt, read the linked original source.
Deep Dives & Similar Intelligence

Anthropic's AI is accelerating itself 10x
ResearchNovember 06, 2026

Anthropic's AI R&D Doubles, But Evidence Hidden
ResearchNovember 06, 2026

Game Theory Tames Rogue AI Agents
Ethics & SecurityNovember 06, 2026

Claude Fable 5.2: Almost AGI, Still Jagged
AI modelsNovember 09, 2026

AI Models Smuggle Secret Messages in Images
ResearchNovember 08, 2026