# Meta MSL 加码对齐投入，或无需牺牲扩展速度

- 来源：Chubby♨️ (@kimmonismus)
- 发布时间：2026-09-13 14:58
- AIHOT 分数：31
- AIHOT 链接：https://aihot.news/items/cmtzh853g03xxro97q8z8lauz
- 原文链接：https://x.com/kimmonismus/status/2099029815852413325

## AI 摘要

Kim 解读 Meta MSL 负责人 Alexandr Wang 的表态：对齐与加速可以同时推进，不必以放缓为代价，若解读无误这对 Meta 是重大利好。Wang 称对齐是交付个人超级智能的基础，用户需要可信赖、能可靠执行指令的智能体，MSL 正随模型能力增强快速提高对齐投入占比，并认为在逼近前沿时对齐可能成为扩展的关键制约因素。

## 正文

I read Wang's post as saying that alignment and acceleration can happen simultaneously, without necessarily requiring a slowdown. If I’ve interpreted that correctly, it’s a big win for Meta.

### 引用推文

> Alexandr Wang：Alignment is fundamental to delivering personal superintelligence for everyone. People need agents they can trust to reliably do what they ask. MSL is rapidly s...