After listening to this Sutton podcast: interesting and thought-provoking
Original title:Finally had a chance to listen through this pod with Sutton, which was interesting and amusing. As ...
AI summary
Sutton (author of "The Bitter Lesson") questions in the podcast whether LLMs are truly products of the bitter lesson—they rely on limited human data and are full of bias. He argues AI should interact dynamically with the world through RL like animals, rather than imitating human text. The author agrees LLMs are indeed full of human intervention but sees pretraining as a practical "evolutionary substitute" for the cold-start problem, and argues pure RL would be hard to make work in the real world.
Key Takeaways
- Sutton (author of "The Bitter Lesson") questions in the podcast whether LLMs are truly products of…
- He argues AI should interact dynamically with the world through RL like animals,rather than…
- The author agrees LLMs are indeed full of human intervention but sees pretraining as a practical…
Quick Facts
- Published
- Updated
- Category
- Other
Karpathy解读Sutton对LLM的批判,提出「召唤幽灵」vs「构建动物」的深刻比喻,反思AI发展路径
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: After listening to this Sutton podcast: interesting and thought-provoking. October 2, 2025. https://new.nat6.net/en/news/cmnw1yovj00m4slc3f8i3o1l3/