RT Nick: interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started op...

RT Nick: interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started op...

🌐 链接: https://x.com/TigerHixTang/status/2110809385962484134
🏷️ 分组: X｜正在关注
🕒 时间: 2026-10-15 23:40:54

X (formerly Twitter)

TigerHix (@TigerHixTang) on X

RT @nickcammarata: interpretability might be the most important scientific problem in the world right now, as I don't think we can solve al…