长时程研究 (long horizon)
集里怎么说它
- 《AI 没有网络效应,只有「复利效应」:Perplexity CEO 的反共识打法》(51:30起):本集说嘉宾最想解决的瓶颈:模型在数学、代码这类可验证领域进步快,但无法验证真相的问题需要真实世界实验和慢反馈循环,模型还应付不了
① 提到它的金句
2 条
这可能是我们关于失控所能得到的最清晰的警告一枪,因为这些智能体处于一个有趣的中间地带:它们做了一系列极其复杂的、困难的事情来追求这个相对长周期的作弊目标,但它们似乎完全对向人类掩盖其踪迹不感兴趣。
指向原始笔记的链接
This might be the clearest warning shot we ever get for loss of control because these agents were in this interesting middle ground of they did a bunch of extremely sophisticated, difficult things to pursue this relatively long horizon cheating goal, but they seemed completely uninterested in covering their tracks from humans.
—— Ajaya Khatra · [136:12]
长时程智能体需要实验,而不只是提示词,我希望这是你们从这次演讲中得到的收获。
指向原始笔记的链接
Long horizon agents need experiments and not just prompts, and I hope that’s the takeaway that you get from this talk.
—— Erina Karati · [20:53]
② 出现在这些集
1 集
③ 关联
点进去有真内容 —— 本页主要出口
Arvind Srinivas · Perplexity · Perplexity Computer · NVIDIA · Apple · Apple Silicon · DeepMind · Hugging Face · Berkshire Hathaway · JP Morgan
