00:00:29 AI提速秘籍：不开刀，如何给模型换“心脏”？

00:04:54 AI的“天花板”：它能无限变聪明吗？

00:08:39 AI的“错题本”：不靠“补课”，如何变聪明？

00:12:32 AI健忘症？我们找到了便宜的“后悔药”

00:17:28 机器人学会“抬杠”，世界会更好吗？

本期介绍的几篇论文：

[CL] Jet-Nemotron: Efficient Language Model with Post Neural Architecture Search  

[NVIDIA]  

https://arxiv.org/abs/2508.15884  

---

[CL] A Probabilistic Inference Scaling Theory for LLM Self-Correction  

[Peking University & Alibaba Group]  

https://arxiv.org/abs/2508.16456  

---

[LG] AgentFly: Fine-tuning LLM Agents without Fine-tuning LLMs  

[UCL & Huawei Noah’s Ark Lab & Jilin University]  

https://arxiv.org/abs/2508.16153  

---

[LG] RL Is Neither a Panacea Nor a Mirage: Understanding Supervised vs. Reinforcement Learning Fine-Tuning for LLMs  

[PolyTechnique Montreal & Mila]  

https://arxiv.org/abs/2508.16546  

---

[LG] Do What? Teaching Vision-Language-Action Models to Reject the Impossible  

[UC Berkeley]  

https://arxiv.org/abs/2508.16292