Promote fine motor development and cognitive training for younger students through origami lessons." Benefits of Origami - An Origami Lesson to Train Finger Muscle Development This talk was given at a TEDx event using the TED conference format but independently organized by a local community. Learn more at https://www.ted.com/tedx ...

两个 编译：haozhen 编辑：siqi 海外独角兽原创编译 转载请注明 去年以来，随着 OpenAI 在 o1 模型中提出 RL 叙事 ，以及 DeepSeek 发布的 R1 模型 解开了 RL 谜 题，AI 行业进入了新范式，智能的下半场也真正开启。 如果说过去 LLM 主要依赖于模式匹配与数据记忆，如今，推理能力的兴起让模型能力从表层关联跃 升到复杂认知。推理不仅仅是参数数量或训练数据的增加，而是能充分利用算力进行深度探索。因 此，推理能力既是涌现智能的重要催化剂，也是未来模型在科学发现、复杂决策与 multi-agent 协作 中的关键。 本篇内容是 OpenAI 研究员 Noam Brown 的最新播客。Noam 是全球最顶尖的推理研究员之一，他最 知名的两个项目分别是在德扑中击败顶尖人类玩家的 AI 系统 Libratus 和 Pluribus，2022 年他又开发 了首个在复杂多人策略游戏 Diplomacy 中达到人类水平的 AI，名为 Cicero。 这次播客中，他详细分享了自己在 scaling test time compute 上的前沿观点： • 推理（reasoning）是模型涌现 ...

Vests used for decades in military-type training are now popular with middle-aged women and other power walkers. What does the research say? https://t.co/uz3vHJOqMJ ...

Vests used for decades in military-type training are now popular with middle-aged women and other power walkers. What does the research say? https://t.co/FIhPcTP2ru ...

5 Surprising Ways Too Much Screen Time Impacts Your Brain https://t.co/NKlkUWA5cP https://t.co/NKlkUWA5cP ...

Neuralink is working to restore vision for the blind by bypassing damaged eyes and directly connecting to the brain’s visual cortex turning blindness into sight. https://t.co/u1VQWS3x2c ...

As tools and tests that gauge brain health become more accessible, a growing body of research suggests we can actually do something about it. 🧠 https://t.co/GVRxFDytxh https://t.co/tWcqsxSWpR ...

India is expected to receive above-normal rainfall in July, the wettest month of the monsoon season, boosting prospects for the planting of key crops such as rice, soybeans, and corn https://t.co/9A6R3cjeAE ...

5 Surprising Ways Too Much Screen Time Impacts Your Brain https://t.co/aQ18q4DcAy https://t.co/aQ18q4DcAy ...

论文链接：https://arxiv.org/abs/2506.20512 代码仓库：https://github.com/GAIR-NLP/OctoThinker 近期，一份来自上海创智学院、上海交通大学的前沿研究论文吸引了人工智能领域的广泛关注。该论文深入探讨了不同基础语言模型家族（如 Llama 和 Qwen）在 强化学习（RL）训练中迥异表现的背后原因，并提出创新性的中期训练（mid-training）策略，成功地将 Llama 模型改造成高度适配强化学习的推理基础模型，显 著缩小了其与天生擅长 RL 扩展的 Qwen 模型之间的性能差距，为下一代 reasoning 能力 AI 系统的开发提供了关键的科学基础和技术路径。 论文发布后在社交媒体引发广泛关注，Meta AI 研究科学家、即将赴 UMass Amherst 任助理教授的 Wenting Zhao 率先盛赞："Truly impressed by how an academic lab just figured out a lot of mysteries in mid-training to close the RL gap betwee ...