Francois Chollet: no 'AGI' declaration until AI can make inventions and conceptual breakthroughs
Neurogence · reddit · 2026-09-08
Francois Chollet, creator of Keras and the ARC benchmark, says he would not "declare AGI" until AI is capable of invention — conceptual breakthroughs, novel insights, and new real-world technology.
He argues the core narrative for building AGI has always been about invention: "it will cure cancer, it will solve fusion energy." By that yardstick, benchmark scores and fluent conversation don't qualify — only producing genuinely new knowledge and technology does.
Related event: Chollet: Don't declare AGI until AI can truly invent(3 posts)→
More from AGI Musings
- Engineer Says He Automated Himself Out of One of the World's Most Technical Jobs — MickeySteamboat · 2026-09-08
- Was the Golden Gate Claude Era the Peak of AI Alignment? — sprooos · 2026-09-08
- Mathematician Andrew Snowden's new preprint used ChatGPT for many arguments — littmath · 2026-09-08
- OpenAI chief scientist Jakub Pachocki: models could be much better at math, but RSI and automated alignment come first — Dr_Singularity · 2026-09-08
- The Superintelligent Dog Analogy: Can High-Dimensional AI Thought Survive Lossy Compression — MajmudarAdam · 2026-09-08
- scaling01 mocks OpenAI's attempt to claim it has achieved AGI — scaling01 · 2026-09-08