Robotics has spent years scaling demonstration data and model size. What happens when we scale EXPLORATION for robot agents? We're incredibly excited to share RAX, an agentic robotics system that learns by scaling online exploration. A coding agent connects directly to the robot’s perception and control APIs, learns from feedback, and builds programs that perceive, plan, and act. The model stays frozen. As exploration scales, the agent improves its programs and builds up reusable skills. The results: • ZERO teleoperation or action demonstrations on real robots—from insertion with 0.1 mm clearance and air-fryer manipulation to deformable objects and long-horizon mobile whole-body control. • +37 percentage points on average over leading learned policies across five simulation benchmarks: RoboDojo, BEHAVIOR-1K, HumanoidArena, DexJoCo’s bimanual tasks, and LIBERO-Pro. • One framework spanning single-arm, bimanual, dexterous, humanoid, and mobile robots, with new embodiments connected through APIs. The craziest part: on each real-world embodiment, going from scratch to mastering the tasks took roughly two hours. No teleop. No model finetuning. We’re open-sourcing the full framework AND every experiment session. We can’t wait to see what people build with it! Project + sessions: robot-agent-exploration.com/ Paper: assets.robot-agent-explorati… GitHub: N Project fully led by @RunhanH and @xczhangxx who are absolutely awesome. How it works, experiments, and robot videos below 🧵 (1/N)

Leonard Read's pencil story: no one knows how to make a pencil from scratch, yet millions are made without a central manager. Each worker knows only their part. Incentives, not central control, coordinate their efforts. Current multi-agent AI seems to work the opposite way. A central manager does everything: assigns roles, writes the workflow, and decides who acts next. As agents increase and tasks get harder, that manager might become the bottleneck: it should understand every agent and every task, and adaptation becomes inefficient. So how can a group of machine agents organize themselves into something none of them could achieve alone? Further, how can that society of agents keep improving itself over time? In Economy of Minds (EoM), we show that simple local incentives can drive global self-coordination and self-improvement for modern LLM agents. We let agents live in a minimal economy: they compete for work, earn by helping others succeed, and go broke when they don't. Workflows, specialists, and improvement naturally emerge, with no one in charge. Across 5 domains (math reasoning, financial research, scientific research, accelerator design, and distributed-system optimization), limited agents grow into strong societies. The whole idea in a 90-second cartoon👇

Today, we introduce Priming Models, a new paradigm for training one-step generative models from scratch. No diffusion teacher. No flow map. No CFG. No JVP. We believe this work will challenge how we think about building one-step models. Joint work with @sitanch, @du_yilun, @parkwbin03, Bohyung Han. 🔥 1.19 FID one-step on ImageNet 256×256 while using substantially less compute, e.g., 12X less than Drifting-B, 6X less than Drifting-L 🎨 Scales to text-to-image generation (GenEval 0.82) 🤖 World modeling on RT-1 / Bridge-V2 Priming Model is built on one core message: [Reconstruction pretraining --> Distributional post-training].

How much representational diversity exists around a trained network? A lot! Moving a ViT’s weights by 1.3% while preserving its predictions, its representations end up less similar to the original than a random-init ViT’s ‼️ New paper: Hessian Null Space Continuation (HNC) 💫🧵 (1/12)

New Preprint! 🚨 Looks can be deceptive! Language models store numbers internally in pretty shapes like curves and helices. Do they compute with these shapes? Not always! To compare two numbers, we discover a neat algorithm: use lines per number and combine them geometrically!

SFT is not dead! 🥳 We found a way to make SFT rival current prevailing posttraining methods, often generalizing better and forgetting less than RL and OPSD. 🤯 Following our prior work on reasoning with sampling, we now introduce sampling to the posttraining stack. 1/n

When do biological reasoning models use their biological inputs? High performance can create a mirage of reliance on biological representations. Reasoning traces don’t always reflect the evidence. Our paper examines these limits and paths to better biological reasoning models.

AI titan Anthropic has launched a biology ‘wet lab’ where human scientists and AI agents will work together to design and conduct experiments. The company has released one of the team’s first finds: a peculiar pattern of DNA in the genomes of several giant viruses. go.nature.com/46K4qmC

Check out our new work, to be presented at @COLM_conf! We find mechanistic evidence that creativity and hallucinations are tied together. Working on this made me realize that some of the same mechanisms that make LLMs useful may also create failures.

🚨 We finally have a better understanding of how the different paradigms for discrete diffusion compare! In new work with Liye Wang, we prove a quadratic separation between the parallel generation capabilities of uniform and Gaussian diffusion relative to unmasking! 1/

Biology does not need models that just describe the world. It needs models that predict what happens when we intervene. In @CellCellPress Focus on AI in Biology, we lay out how to build biomedical world models, from the data they need to how they should be trained and evaluated.