Event ·Multi-source

Andrej Karpathy：先掌握模型再谈智能体，否则代价惨重

Anthropic工程师Andrej Karpathy指出当前AI领域最大错误是强迫智能体工作而不先理解底层模型。他回忆2016年在OpenAI犯过同样错误，因此浪费了5年时间。Karpathy强调演示容易但产品化需十年，自动驾驶已验证这一点。他认为智能体不是产品，模型基础才是，打好基础后智能体会自然涌现。

Current read

Karpathy用自己的失败经历点醒你：别急着搞智能体，先把模型弄透。少走5年弯路。

11 sources65° AI heatLast updated 2026/7/21 10:02:31

Evidence

11 sources

Primary sourceAI Will

2026/7/21 10:02:31View source

Related source 1IT之家

2026/7/20 11:40:14View source

Related source 2arXiv: Anthropic

2026/7/21 15:45:47View source

Related source 3宝玉

2026/7/21 22:40:38View source

Related source 4The Rundown AI

2026/7/20 10:30:11View source

Related source 5Decoder

2026/7/20 14:03:53View source

Related source 6techcrunch

2026/7/20 15:15:00View source

Related source 7@OpenAIDevs

2026/7/20 21:36:41View source

Related source 8elvis

2026/7/20 22:09:00View source

Related source 9pandaily

2026/7/21 01:57:10View source

Related source 10歸藏(guizang.ai)

2026/7/21 02:49:46View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story