Nathan Lambert · @natolambert · X·2026-08-20 00:02·5天前
AI 导读

Nathan Lambert 呼吁建立独立组织,获取前沿模型训练全程细节以进行监督,而非仅在事故发生后介入。他肯定 OpenAI 主动分享信息,但认为安全的关键在于更多信任与更多监督。此前 OpenAI CEO Sam Altman 宣布暂停部分前沿 RL 训练,以确保对齐、安全与监控标准跟上能力快速提升。

Nathan Lambert@natolambert
65AI 编辑部评分,满分 100
2026-08-20 00:02· 5天前
AI 导读

Nathan Lambert 呼吁建立独立组织,获取前沿模型训练全程细节以进行监督,而非仅在事故发生后介入。他肯定 OpenAI 主动分享信息,但认为安全的关键在于更多信任与更多监督。此前 OpenAI CEO Sam Altman 宣布暂停部分前沿 RL 训练,以确保对齐、安全与监控标准跟上能力快速提升。

We should have independent organizations that can access the full details of these training runs for monitoring, not just after various accidents occur.

It's good that OpenAI is sharing this, but the best way for safety is more trust and more eyes on the hard problems to solve.

Sam AltmanWe have paused some frontier RL training to ensure that we can meet the appropriate alignment, security and monitoring standards for the new level of capabiliti...

来源:Nathan Lambert· x.com