[人人能懂] 从数据炼金到思维融合

00:00:25 AI的“炼金术”：如何把“废料”变成“燃料”？00:04:50 AI替你搞科研，是馅饼还是陷阱？00:09:44 人多不一定力量大，但“脑子”多可以00:14:17 给AI当老板，你只需要做好一件事00:18:20 AI的“减肥”秘籍：高手做事，不在于“多”，而在于“准”本期介绍的几篇论文：[LG] Generative Data Refinement: Just Ask for Better Data  [Google DeepMind]  https://arxiv.org/abs/2509.08653  ---[LG] The More You Automate, the Less You See: Hidden Pitfalls of AI Scientist Systems  [Carnegie Mellon University]  https://arxiv.org/abs/2509.08713  ---[LG] Merge-of-Thought Distillation  [Zhejiang University & Inclusion AI]  https://arxiv.org/abs/2509.08814  ---[CL] Verbalized Algorithms  [MIT CSAIL & Marist University & UC Irvine]  https://arxiv.org/abs/2509.08150  ---[LG] EvolKV: Evolutionary KV Cache Compression for LLM Inference  [University of Chinese Academy of Sciences & ETH Zurich]  https://arxiv.org/abs/2509.08315  在小宇宙查看该单集文稿