githubchy
793 posts


OpenClaw 让我们看到了一个更好的、适用于未来长期的架构。 这套架构能保证未来每个人都有一个独属于自己的、高度可定制的、数据相对本地化且安全的 AI Agent。虽然现在的小毛病还很多,但是这套架构是值得长期投资的。(使用它本身就是一种投资) 即使未来有一个更好的 Agent 出现,现有的这些架构上记录的大量数据对于未来也都是长期非常有价值的。这些数据的可用性,要比之前离散在 ChatGPT、Claude 等等各种产品里积累的东西更强的多。

Recent neuroscience suggests learning may not be “weights → activity,” but rather activity inferred first → weights updated to stabilize it. Inference before plasticity. That’s interesting. Most AI systems mutate parameters directly (backprop). But biological systems may settle into a stable internal configuration first ,then consolidate it. In ARIA, we’ve taken a similar approach at the systems level: • Infer candidate state • Constrain through governance • Only then persist structural change Stability before mutation. If brains work this way, it reinforces something important: Learning isn’t just optimization. It’s controlled state transition inside bounded dynamics. Backprop made AI powerful. Constraint-first learning may make it stable. We believe next-generation AI systems will need governance layers that function like biological constraint dynamics, not just gradient descent. Paper: nature.com/articles/s4159…

科沃斯推出了一个新产品,叫“八界”! 内置OpenClaw的家庭机器人管家!! 可以远程发消息,让它收拾屋子。 可以帮你整理家里的桌面、收纳玩具、取东西、整理凌乱的鞋子、或者把衣服放进洗衣机等等。 虽然长得不咋样,也还是个雏儿。 但这个能力确实是很多人需要的,这才是以后家庭机器人应该干的事情! 而不是拿着遥控器,让机器人后空翻!





We’re making robots more capable than ever in the physical world. 🤖 Gemini Robotics 1.5 is a levelled up agentic system that can reason better, plan ahead, use digital tools such as @Google Search, interact with humans and much more. Here’s how it works 🧵

Today, I made the decision to leave our Collaboration Agreement with OpenAI Figure made a major breakthrough on fully end-to-end robot AI, built entirely in-house We're excited to show you in the next 30 days something no one has ever seen on a humanoid




Very happy to start sharing our work at Pi 🤖❤️ - a 3B pre-trained generalist model trained on 8+ robot platforms - a post-training recipe that allows robots to do dexterous, long-horizon tasks physicalintelligence.company/blog/pi0 What's exciting isn't laundry, but the recipe- a short 🧵


Where do I start with robotics? Wanna code and build my own robots at home for fun and potential world domination









