Keras Creator Reassesses LLMs: Underestimated Long-Term Value, But Scaling Isn't AGI
fchollet · x · 2026-08-07
Keras founder François Chollet tweeted a reassessment of his views on Large Language Models (LLMs). He admitted to underestimating their long-term importance in early 2023 and 2024, citing the December 2024 o3 test-time compute breakthrough as the moment he changed his mind.
He also clarified several standing convictions:
- Scaling deep learning will continue to pay off and unlock new applications.
- However, scaling alone is not the path to AGI.
- We are still not particularly close to AGI.
- Existing deep learning techniques are far from fully deployed, and a massive amount of value remains to be created with current tech.
More from AGI Musings
- A missing parameter in scaling laws: H, the human effort behind learning signals — lateinteraction · 2026-10-16
- First-ever paper on multi-agent swarm safety lands NeurIPS spotlight — LauraRuis · 2026-10-16
- Yoav Artzi on peer review: paper flooding still pays off as review system strains under load — yoavartzi · 2026-10-16
- Letta founder: work in fields defined by problems, not methods — lateinteraction · 2026-10-16
- Why pure mathematicians will soon flood into AI x-risk debates — Dr_Atoosa · 2026-10-16
- RSI began when model outputs augmented next-gen training data, agents just escalated it — akbirthko · 2026-10-16