# Jeff Dean says recursive self-improvement cannot run infinitely fast

> Original title: Jeff Dean李飞飞最新对谈：全是RSI非共识

- Source: QbitAI
- Published: 2026-11-07T03:23:29.000Z
- Bad Influence: https://badinfluence.news/items/fgxkoy8i703ivd88rfg5fphrr
- Original: https://www.qbitai.com/2026/11/508196.html

## Summary

Jeff Dean says automation can speed up AI research, but training, computation, data storage and evaluation still take real time, so the loop cannot run infinitely fast. In a Stanford HAI conversation with Fei-Fei Li, he also described his new company Discovery Loop, which runs about 20 automated improvement loops and aims to raise domain experts' productivity 20 to 50 times.