我们都希望学得更聪明，但到底怎样才算“聪明”？本期我们就从几篇最新论文出发，看看AI是如何被教导着实现真正的“开窍”：它要如何学会看透不同知识表象下的本质，如何为自己打造一个用于自我提升的“进度条”，又是如何从只追求唯一的最优解，到学会欣赏整个“高分区”的所有好答案。这些AI的“内功心法”，或许正是我们自我成长的关键钥匙，让我们一探究竟！00:00:32 AI 学习的“升维”之路：从“对答案”到“懂原理”00:05:32 机器人的“开窍”秘诀：从抄作业到上补习班00:11:18 AI训练的“内功心法”：当数据成了稀缺品00:16:59 AI的“开窍”心法：从单打冠军到全能高手00:21:36 从一锅粥里，尝出每一粒米的味道本期介绍的几篇论文：[CL] LLM-JEPA: Large Language Models Meet Joint Embedding Predictive Architectures [Atlassian & NYU & Brown University] https://arxiv.org/abs/2509.142 ---[LG] Self-Improving Embodied Foundation Models [Google DeepMind & Generalist AI] https://arxiv.org/abs/2509.15155 ---[LG] Pre-training under infinite compute [Stanford University] https://arxiv.org/abs/2509.14786 ---[LG] FlowRL: Matching Reward Distributions for LLM Reasoning [Shanghai Jiao Tong University & Renmin University of China & Microsoft Research] https://arxiv.org/abs/2509.15207 ---[LG] Optimal Learning from Label Proportions with General Loss Functions [Google] https://arxiv.org/abs/2509.15145

Ask this episode

“What did the council decide on item 42?” — answered with the speaker and the moment.

Transcript

Related