你有没有想过，聪明的AI不只靠堆算力，更要靠高质量的思考方式？本期我们要聊的几篇最新论文，就为我们揭示了AI正在经历一场深刻的“认知升级”。我们将看到，AI如何像一个身处江湖的开源模型，用聪明的策略追赶顶尖高手；又如何进行哲学层面的“自我觉醒”，把自己看作世界的一部分来做出更优决策。我们还会探讨，AI怎样像武林高手一样边解决难题边“涨功夫”，甚至学会给自己的思维“断舍离”，用最少的步骤直达问题核心。准备好，我们马上进入AI的思考进化之旅。00:00:41 AI江湖：开源大模型如何追赶“独孤求败”？00:06:34 AI的心智革命：当我成为世界的一部分00:12:38 AI如何像高手一样，边解题边涨功夫？00:18:14 AI思考，也需要“断舍离”00:22:58 如何让你的AI助手，思考速度提升三倍？本期介绍的几篇论文：[LG] DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models[DeepSeek-AI]https://modelscope.cn/models/deepseek-ai/DeepSeek-V3.2/resolve/master/assets/paper.pdf---[LG] Embedded Universal Predictive Intelligence: a coherent framework for multi-agent learning[Google]https://arxiv.org/abs/2511.22226---[LG] ThetaEvolve: Test-time Learning on Open Problems[Microsoft & University of Washington]https://arxiv.org/abs/2511.23473---[LG] ORION: Teaching Language Models to Reason Efficiently in the Language of Thought[Harvard University & Hippocratic AI & MIT]https://arxiv.org/abs/2511.22891---[CL] Focused Chain-of-Thought: Efficient LLM Reasoning via Structured Input Information[FAR.AI & German Research Center for Artificial Intelligence (DFKI) & University of Kassel]https://arxiv.org/abs/2511.22176

Ask this episode

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

Transcript

Related