Ning Ding

Building upon SimpleVLA-RL, we have implemented real-world RL on long-horizon dexterous tasks and witnessed a non-trivial (~relatively 300%) performance improvement over the SFT model, along with surprising capabilities on auto-recovery. Blog coming soon. The entire process uses very little data and training compute—basically costing no more than a single robotic arm—hinting that real-world generality for machines is actually within sight.

Open source the recipe

working on distillation

✨ [ICLR 2026] How Far Can Unsupervised RLVR Scale LLM Training? The dream: models can improve themselves without human supervision. The reality: sometimes they can only sharpen what they already believe. Intrinsic rewards struggle to scale LLM training because they follow a rise-then-fall pattern that makes collapse mathematically inevitable. But that's not the end of the story. We find unsupervised RLVR (URLVR) is particularly well-suited for test-time training and quantifying model priors. The full picture 👇 📄 Paper: arxiv.org/abs/2603.08660 🧪 GitHub: github.com/PRIME-RL/TTRL (1/n)

@junxian_he 😂😂

@stingning Haha me from that era

Annotated Transformer: a legend that anyone from that era has studied. github.com/harvardnlp/ann…

Okay Grok, that’s terrifyingly good.

@goodhunt true

@stingning github.com/Hmbown/DeepSee… If this time is fake I’m going to be so sad

🐳🐳🐳

it is coming.

@sciencegirl Open claw!

I increasingly believe that all the things — AI, robotics, controlled fusion, space, quantum computing, and more — will converge into one thing.

GPT-5.3-Codex-Spark on @cerebras is something very different.

Today I heard a line that stuck with me: "the real moat is the organizational structure."

I’m not sure how many people have noticed this, but there are already very, very few humans left in @openclaw's PR queue: AIs open the pull requests, other AIs review and score them, and (partially maybe) decide what gets merged. If you want to see “self-evolving AI” in the wild, it’s right here.

Is OpenClaw actually the fastest-shipping projects on GitHub history? (by release cadence maybe)

@geekbb 能，相信ds

When your project gets too good, your next PR is either a legal letter or a payroll.