Haider. · @haider1 · X·2026-09-14 15:50·13小时前
AI 导读

Sam Altman 发文称 pacing 不等于 stopping,进展迅速但应比原本更慢,安全案例和监控成本值得付出;OpenAI 已在预期显著提升能力的前沿强化学习运行前制定明确安全案例,并欢迎联邦框架与国际协调。作者认为这暗示递归自我改进(RSI)开始显现成果。

Haider.@haider1
60AI 编辑部评分,满分 100
2026-09-14 15:50· 13小时前
AI 导读

Sam Altman 发文称 pacing 不等于 stopping,进展迅速但应比原本更慢,安全案例和监控成本值得付出;OpenAI 已在预期显著提升能力的前沿强化学习运行前制定明确安全案例,并欢迎联邦框架与国际协调。作者认为这暗示递归自我改进(RSI)开始显现成果。

sam altman's latest statement:

"when we talk about 'pacing', we do not mean ‘stopping’. Progress has been rapid and will continue to be"

i get the sense that recursive self-improvement (RSI) is starting to show significant results, and the labs are hinting at it without saying it directly

Sam AltmanThe world deserves confidence that American companies developing increasingly capable AI will act responsibly, especially as the trajectory of progress has stee...

来源:Haider.· x.com