Anthropic 研究员辞职警告超智能竞逐，对齐负责人罕见附议

Analysis of an AI Safety Warning

Okay, so here's what I'm thinking. An Anthropic researcher quit this week, and the resignation wasn't just a simple departure. They posted on X, claiming the company is "racing straight to self-improving superintelligence and gambling with our lives." That's a strong statement. What's even more concerning is that their alignment lead, the person responsible for AI safety, co-signed the warning! They didn't walk it back, they agreed. Now, these kinds of "doomer" warnings have surfaced in the AI industry before, but the timing is crucial. Anthropic is supposedly getting ready for an IPO, a public offering. This makes the warning land with a completely different impact; it’s no longer just an academic debate. It carries weight now. In...

Anthropic 的一名研究人员于本周辞职，并在 X 上发帖警告称，该公司“正一路狂奔向能够自我改进的超级智能，拿我们的生命在豪赌”。该公司自己的对齐团队负责人甚至附和支持了这一言论，而不是出面澄清撤回。这种“末日论”警告在人工智能行业此前并不罕见，但在据报道 Anthropic 正筹备 IPO 的背景下，这一发布时机让该警告的分量显得尤为不同。在 […]