# Nathan Lambert：开源模型恐被过早封禁

- 来源：Nathan Lambert (@natolambert)
- 发布时间：2026-09-13 19:35
- AIHOT 分数：38
- AIHOT 链接：https://aihot.news/items/cmtzrec6i02o7ro1bfi91erqw
- 原文链接：https://x.com/natolambert/status/2099099697449193596

## AI 摘要

Nathan Lambert 认为开源模型更可能因普遍恐慌而被过早封禁，全面禁止开源将导致更多反乌托邦后果。他承认某些能力级别的开放权重模型或可被"暂停"，但强调这需要非常专业的判断。此番讨论源于 @tszzl 预测开源将在某次重大灾难后被禁，并希望 Kimi、DeepSeek 等继续做模型但通过 API 受监控。

## 正文

Open-source being fully banned will lead to more dystopia. There could be times where I see open weight models of a certain capability being *paused*, but we need very careful expertise to make that call.

It's more likely open models are banned prematurely by general alarmism.

### 引用推文

> roon：@sean_from_earth i won't lie to you, i think open source will be banned before too long after some major disaster. and when the day comes, you'll agree with me....