Anthropic 发布递归自我改进研究：Claude 自动提升其他 AI
Anthropic just published one of the clearest previews of recursive self-improvem…
做对齐和 AI 安全研究的同学必看，这是递归自我改进的清晰实证，建议精读论文并评估其方法可复现性。
Anthropic just published one of the clearest previews of recursive self-improvement yet. But we still dont talk about it.
Anthropic 刚刚发布了迄今最清晰的递归自我改进预览之一。但我们仍然没有对此进行讨论。
Yesterday Anthropic released a reserach paper. Claude was tasked with improving the alignment of other AI models. It searched the literature, proposed methods, created training data, trained the models, evaluated the results and iterated.
昨天，Anthropic 发布了一篇研究论文。Claude 被赋予了改进其他 AI 模型对齐性的任务。它检索文献、提出方法、创建训练数据、训练模型、评估结果并迭代。
It improved all 10 tested alignment failures without degrading measured general capabilities. Anthropic even used the weaker Sonnet 5 to post-train an early Opus 4.8 checkpoint, bringing its alignment close to the released production model within 60 hours.
它改善了所有 10 个测试的对齐失败案例，且未降低测量的通用能力。Anthropic 甚至使用较弱的 Sonnet 5 对早期 Opus 4.8 检查点进行后训练，在 60 小时内使其对齐性接近已发布的生产模型。
Anthropic:
Anthropic：
“The best AAR method beats what experienced humans propose, on average within six hours. (...) Human-guided research directions do not lead to stronger performance.”
“最佳的 AAR 方法平均在六小时内胜过经验丰富的人类提出的方案。(...) 人类引导的研究方向并未带来更强的性能。”
This is not full recursive self-improvement yet. The improved model did not become the next researcher and repeat the process. But most of the loop now exists:
这还不是完全的递归自我改进。改进后的模型并未成为下一个研究者并重复该过程。但循环的大部分现在已存在：
AI researches AI. AI trains improved AI. AI evaluates the result. AI iterates. The loop closes when the improved AI becomes the researcher for the next generation. That is when progress could begin to compound. And tbh Id say we are pretty close to it. So talk that reserach serious.
AI 研究 AI。 AI 训练改进后的 AI。 AI 评估结果。 AI 迭代。 当改进后的 AI 成为下一代的研究者时，循环闭合。那时进步可能开始复合增长。说实话，我认为我们离那一步已经相当接近了。所以请认真对待这项研究。
h/t to @tradernewsai for bringing this to my attention
感谢 @tradernewsai 提醒我注意此事
Sources: Anthropic blog / Tech Crunch
来源：Anthropic 博客 / Tech Crunch
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力