[人人能懂AI前沿] 一个大脑、一本秘籍、一次指点：AI进化新思路

你有没有想过，我们能不能像给人指路一样，只对机器人“指一下”就让它心领神会？怎样才能给AI一本“武功秘籍”，让它告别“瞎忙”，拥有真正高手的“手感”？本期节目，我们将通过几篇最新论文，揭示AI如何抛开事物的表象、看见动作的“骨骼”，并探索如何用一个更统一、不“精神分裂”的大脑，来更高效地理解这个世界。
00:00:28 让机器人认路，只需要教它“指一下”？
00:05:23 让AI告别“瞎忙”，给它一本“武功秘籍”
00:10:35 大模型提速的“第三条路”
00:15:35 抛开皮囊，看见骨骼，机器人怎么学“手艺”
00:20:43 AI的大脑，怎样才能不精神分裂
本期介绍的几篇论文：
[RO] LightNav-0: Eliciting VLM Spatial Intelligence for Generalist Embodied Navigation
[Light Origins Team]
https://arxiv.org/abs/2608.30935
---
[AI] Repo-To-Skill: Distilling GitHub Repositories Into AI4AI Skills
[Beijing Academy of Artificial Intelligence]
https://arxiv.org/abs/2609.02749
---
[LG] Unlocking Lossless Speedups in LLMs via Discrete Diffusion
[Institue of Foundation Models]
https://arxiv.org/abs/2609.04010
---
[CV] RoboTok: An Internet-Scale Data Engine for Human Demonstration Retrieval and Dexterous Manipulation Learning
[Rice University]
https://arxiv.org/abs/2609.03199
---
[IR] NeoMME: A Single-Tower Multimodal-Native Multilingual Foundation Encoder for Efficient Fine-Tuning and Inference
[H Company]
https://arxiv.org/abs/2609.01657