Event ·Multi-source

Claude Code 作者 Boris Cherny：别微调，赌通用模型

Anthropic Claude Code 创作者 Boris Cherny 在 Lenny's Podcast 上分享工程观点：通用模型几乎总是胜过针对特定任务的小模型或微调模型。他给出两个理由：脚手架类优化通常只能带来 10%-20% 的性能提升，且这些收益常被下一代模型直接抹平。他的建议是，如果有选择余地，直接押注更通用的模型，甚至等下一个版本比做定制优化更划算。

Current read

Claude Code 作者本尊发话：别折腾微调和小模型，脚手架只提 10%-20%，下一代模型一发全白干，直接等新模型就行。

11 sources73° AI heatLast updated 2026/9/28 06:37:51

Evidence

11 sources

Primary sourcerohanpaul_ai

2026/9/28 06:37:51View source

Related source 1shao__meng

2026/9/29 02:36:47View source

Related source 2IT之家

2026/9/26 07:11:18View source

Related source 3Claude Code: GitHub Releases

2026/9/28 18:02:03View source

Related source 4Boris Cherny

2026/9/28 18:25:49View source

Related source 5claudedevs

2026/9/28 21:40:50View source

Related source 6Latent.Space

2026/9/29 01:30:00View source

Related source 7Latent Space (swyx)

2026/9/29 01:48:18View source

Related source 8AI产品黄叔

2026/9/29 02:51:29View source

Related source 9Gergely Orosz

2026/9/29 20:33:37View source

Related source 10宝玉

2026/9/26 07:23:53View 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