想知道为什么教机器人玩最“笨”的玩具，反而能让它学会抓取任何东西吗？本期节目，我们将一起探索如何将神秘的AI“炼金术”变成一门严谨的科学，看看怎样让AI大神学会“说人话”并带得动AI小白，并最终揭示，那些五花八门的调教秘籍背后，其实藏着同一个简单的目标。让我们马上进入今天的前沿速递！

00:00:28 AI大模型调教指南：从玄学到科学

00:05:39 返璞归真：最笨的方法，可能就是最好的方法

00:11:25 想让机器人变聪明？先教它玩“笨”玩具

00:16:41 如何让AI大神，带得动AI小白？

00:00 大模型调教秘籍：条条大路通罗马？

本期介绍的几篇论文：

[LG] The Art of Scaling Reinforcement Learning Compute for LLMs

[Meta & UT Austin & UC Berkeley]

https://arxiv.org/abs/2510.13786

---

[RO] VLA-0: Building State-of-the-Art VLAs with Zero Modification

[NVIDIA]

https://arxiv.org/abs/2510.13054

---

[RO] Learning to Grasp Anything by Playing with Random Toys

[UC Berkeley]

https://arxiv.org/abs/2510.12866

---

[LG] Tandem Training for Language Models

[Microsoft & EPFL & University of Toronto]

https://arxiv.org/abs/2510.13551

---

[LG] What is the objective of reasoning with reinforcement learning?

[University of Pennsylvania & UC Berkeley]

https://arxiv.org/abs/2510.13651