[人人能懂] 从“一根筋”到“立体思维”

00:00:33 AI的“一根筋”难题：如何跳出思维的隧道？00:05:08 AI的“高情商”陷阱：当它说得越流利，你可能越要小心00:08:53 如何让AI学会“开窍”？00:13:49 给你的AI，装上一个“立体思维”引擎00:18:41 AI偷懒的智慧：如何用更少的力气，办更聪明的事？本期介绍的几篇论文：[LG] ParaThinker: Native Parallel Thinking as a New Paradigm to Scale LLM Test-time Compute  [Tsinghua University]  https://arxiv.org/abs/2509.04475 ---[CL] Knowledge Collapse in LLMs: When Fluency Survives but Facts Fail under Recursive Synthetic Training  [Holistic AI & University College London]  https://arxiv.org/abs/2509.04796  ---[LG] Bootstrapping Task Spaces for Self-Improvement  [Meta Superintelligence Labs]  https://arxiv.org/abs/2509.04575  ---[CL] Enhancing Diversity in Large Language Models via Determinantal Point Processes  [Boston University & University of Maryland & University College London]  https://arxiv.org/abs/2509.04784  ---[LG] Recurrent State Encoders for Efficient Neural Combinatorial Optimization  [University of Hildesheim]  https://arxiv.org/abs/2509.05084  在小宇宙查看该单集文稿