[人人能懂] 从乐高蓝图、视觉思考到决策梦之队

你有没有想过，AI的“聪明”和我们的“聪明”，到底有什么不一样？本期节目，我们将一起探索AI如何用乐高一样的蓝图搭建软件帝国，如何识破只会考试的“高分低能”陷阱，又是如何扔掉专家地图、让“眼睛”学会思考，并最终用“精兵策略”做出更聪明的决策。准备好了吗？让我们从五篇最新的论文出发，一探AI智慧的边界。00:00:31 软件世界的“乐高”说明书：从一句话到一个帝国 00:05:50 AI医生的“高分低能”陷阱：别被排行榜骗了00:10:51 扔掉“专家地图”，AI也能走出一条新路00:15:51 AI的下一场革命：当“眼睛”开始像“大脑”一样思考00:21:18 从“人海战术”到“精兵策略”：让AI的每一次计算都花在刀刃上本期介绍的几篇论文：[CL] RPG: A Repository Planning Graph for Unified and Scalable Codebase Generation  [Microsoft]  https://arxiv.org/abs/2509.16198  ---[LG] The Illusion of Readiness: Stress Testing Large Frontier Models on Multimodal Medical Benchmarks  [Microsoft Research]  https://arxiv.org/abs/2509.18234  ---[LG] SimpleFold: Folding Proteins is Simpler than You Think  [Apple]  https://arxiv.org/abs/2509.18480  ---[LG] Video models are zero-shot learners and reasoners  [Google DeepMind]  https://arxiv.org/abs/2509.20328  ---[LG] Best-of-∞ -- Asymptotic Performance of Test-Time Compute  [New York University & Institute of Science Tokyo & NEC Corporation]  https://arxiv.org/abs/2509.21091  在小宇宙查看该单集文稿