00:00:25 如果AI自己就能当科学家，我们该怎么办？

00:04:46 AI天气预报，猜得到开头，猜不到结局？

00:08:41 AI变聪明的秘密：死磕难题，还是广见世面？

00:12:41 像画家一样思考：AI学会了打草稿

00:17:03 AI刷题也“内卷”？换个姿势，自己出题自己做

本期介绍的五篇论文：

[LG] Virtuous Machines: Towards Artificial General Science  

[Explore Science & The University of Queensland]  

https://arxiv.org/abs/2508.13421  

---

[LG] Numerical models outperform AI weather forecasts of record-breaking extremes  

[University of Geneva & ETH Zurich]  

https://arxiv.org/abs/2508.15724  

---

[LG] Depth-Breadth Synergy in RLVR: Unlocking LLM Reasoning Gains with Adaptive Exploration  

[The Hong Kong University of Science and Technology (Guangzhou) & ETH Zurich]  

https://arxiv.org/abs/2508.13755  

---

[CV] Next Visual Granularity Generation  

[Nanyang Technological University & SenseTime Research]  

https://arxiv.org/abs/2508.12811  

---

[CL] Beyond Pass@1: Self-Play with Variational Problem Synthesis Sustains RLVR  

[Microsoft]  

https://arxiv.org/abs/2508.14029