What's different between these two BC policies? It's the... same architecture, training budget, and data collection setup — the only difference is the controller gains! Controller gains are an understudied design parameter in robot learning. In our new work (w/ Antonia Bronars*, Pulkit Agrawal), we show how they act as an inductive bias across BC, RL, and Sim2Real transfer, with real consequences on performance. Here's what we found 🧵 * Equal Contribution 📄arxiv: 🔗website:show more

Here’s a pretty weird and surprising result - retrieval-augmented... generation works unreasonably well for robot learning – but only when parameterized using difference vectors! We introduce Difference-Aware Retrieval Policies for Imitation Learning (DARP), a simple, semi-parametric RAG architecture for imitation learning that achieves gains of up to 200% over standard behavior cloning. No additional assumptions beyond BC, just a little architecture switch! The theory backing it up is pretty cool too and it works on real robots! :) Play with our website to understand better: 🧵(1/7)show more

How can we more effectively leverage robot data from... different embodiments for skill transfer? Excited to share that our new work, RoVi-Aug, has been accepted to Conference on Robot Learning as an oral paper！ WIth RoVi-Aug, you can augment an existing robot dataset into a different robot and different viewpoints. A policy trained on the augmented dataset can zero-shot deploy on the unseen target robot with significantly different camera angles! 🧵👇 🔗 Check out our paper:show more

Controlling a CM6 robot with a LEAP motion controller.... J2 of the robot is a bit underpowered but we mainly tested our algorithm here that we will implement on our other robots. Data is shared at 125Hz between robot and Leap over local network.show more

Inkling-small is out today! With SGLang, you can get... 648 tok/s decode with DSpark (simulated acc len=4) and 288 tok/s w/o DSpark, under the same setup (8x NVIDIA AI B200, TP 8, NVFP4, bs=1). What makes this model different is the size. 276B total with 12B active is a sweet spot for RL, and both LoRA and full-parameter training become well within reach. Miles is ready and verified for multimodal RL on Inkling-small, so you can turn your multimodal data into real capability gains. At ~1/4 the size, Inkling-small matches the bigger version in capability and even wins on some benchmarks. Run Inkling-small with SGLang, and customize it with Miles.show more

I FOUND AN INPUT FOR RAGING DEMON (I cant... believe I'm actually labbing in MH...) When there's a wound > ⏹️+⏹️ +🔼 +⏺️ Here's also my camera focusing on the controller to show how it works THIS IS CRAZY, ARE WE PLAYING SF6?!show more

Robora Sim: A PyBullet-Powered Environment for Learning Robotic Physical... Intelligence We are currently building our Robora simulation environment setup for our sim based learning, leveraging PyBullet, an industry-standard physics engine widely used in AI-driven robotics research and development. The environment is optimized with GPU-accelerated learning algorithms, enabling high-speed imitation learning and reinforcement learning within a safe and controlled virtual setup before shipping out to real world. This simulation platform allows our models to learn, adapt, and generalize across different robot morphologies, terrain types and task objectives - all before deployment to the real world. At it's core, the system combines a VLA-powered high-level planner with low-level motion control algorithms, working cohesively to produce emergent, physically intelligent behaviors. This synergy between simulation, learning, and real-world transfer marks a major step forward in our pursuit of adaptive and intelligent robotic systems. Through advanced domain randomization and synthetic data generation, the Robora Simulation Environment ensures that policies trained in simulation transfer effectively to real-world robots, minimizing the sim-to-real gap. Moreover, users will be able to test and integrate their own hardware kits within selected simulation environments in the Robora Dapp, ensuring seamless compatibility and safer real-world implementation.show more

𝗥𝗼𝗯𝗼𝘁𝘀 𝗱𝗼𝗻’𝘁 𝗻𝗲𝗲𝗱 𝗺𝗼𝗿𝗲 𝗱𝗲𝗺𝗼𝗻𝘀𝘁𝗿𝗮𝘁𝗶𝗼𝗻𝘀. 𝗧𝗵𝗲𝘆 𝗻𝗲𝗲𝗱 𝘁𝗼 𝗹𝗲𝗮𝗿𝗻... 𝗳𝗿𝗼𝗺 𝗳𝗮𝗶𝗹𝘂𝗿𝗲 — 𝗮𝗳𝘁𝗲𝗿 𝘄𝗮𝘁𝗰𝗵𝗶𝗻𝗴 𝗵𝘂𝗺𝗮𝗻𝘀. Most robot learning systems assume failure is the end of learning. In our new work, we study whether robots can improve after deployment by learning from their own failures, without any human intervention, teleoperation, or corrective labels. The key idea is simple: human videos contain structure about how the world works. We use them to learn cross-embodiment representations of action, dynamics, and value, enabling a shared predictive space between human behavior and robot experience. This allows a new learning loop: 👉 pretrain on human videos 👉 deploy robot policy 👉 observe failures 👉 reinterpret failures using human priors 👉 improve autonomously We evaluate this across 7 real-world manipulation tasks, showing: 📈 40% → 81% success rate 🏆 Strong improvements over π0.6 RECAP and RISE ✔️ Zero human intervention during post-deployment improvement 🧬 Generalizes across robot embodiments and policy backbones A key finding is that explicit failure repair significantly outperforms failure reweighting, yielding substantially larger gains under identical data conditions (+25 pts vs +5 pts on the same π0.5 base policy). Overall, the results suggest a shift in how we think about robot learning: Human videos are not only for pretraining policies. They can provide the structure needed for continual self-improvement after deployment. 📄 Paper: 🌐 Project: I am grateful for working with the fantastic leads Hanzhi Chen and Anran Zhang, and our collaborators Simon Schaefer, Kejia Chen, Shi Chen, Daniel Cremers. Special thanks to Stefan Leutenegger for co-advising this project with me. ETH Zürich TU München Microsoft Check out Hanzhi's 🧵 for more detailsshow more

Mia's in another two-week intensive, and she's blowing us... away. With lots of supports, she's been able to hold an independent sit AND an independent wall stand. As a whole, we are very, very, very tired. But her gains show how worth it this all is.show more

New work on Training Foundation Models with a Single... Algorithm: Paper: Code, Models: This work introduces Unified Training, a framework that unifies pretraining, midtraining, supervised fine-tuning and RL under a single maximum-likelihood objective. Building on MaxRL, it enables static data and RL-generated data to be trained together within the same system. Empirically, Unified Training improves accuracy and compute efficiency, generalizes better to harder reasoning problems, and reduces forgetting. Interestingly, reusing data across training stages can improve performance even when conventional staged training fails. It also produces models that adapt more effectively to new tasks, leading to stronger continual learning. See an excellent thread by Yuda Song.show more

The visuals and sound design in Gears of War:... E-Day are actually insane! Playing it on PC with the DualSense controller and having an absolute blast. This is exactly the kind of intense, cinematic experience I love.show more

📍 At the border between Poland and Belarus. Poland... is facing hybrid attacks from across the border. We stand in solidarity with our Member State. In our new EU budget, we have proposed to triple border management funding. We are all responsible for Europe’s external bordersshow more

Can we make Transformers better and more efficient for... robot learning? Excited to introduce Body Transformer (BoT), an architecture that leverages robot embodiment in the attention mechanism, by treating it as a graph of sensors and actuators.show more

So I hear that behavior cloning is all the... rage now. What if we could do better, but with the same data? :) In CCIL, we show that imitation via BC is improved by synthesizing corrective labels to account for compounding error, without interactive oracles. Lets you do 👇! 🧵(1/9)show more

It's Christmas morning: OpenAI and Anthropic shipped new models... on the same day! We tested GPT 5.3 Codex vs. Opus 4.6 head-to-head. Verdict: the models are converging. Here’s what we found 🧵show more

Disappointed with your ICLR paper being rejected? Ten years... ago today, Sergey and I finished training some of the first end-to-end neutral nets for robot control 🤖 We submitted the paper to RSS on January 23, 2015. It was rejected for being "incremental" and "unlikely to have much impact" Our resubmission to NeurIPS was also rejected It now has >4,000 citations (and more importantly, end-to-end training is widely accepted!) It's also cool to think about what's changed and what's the same -- - The network was 92k parameters and trained on ~15 minutes of data - The code was a combination of matlab, caffe, ROS, a custom CUDA kernel for speed, and a low-level 20 Hz controller in C++, all talking to each other. ROS+matlab was as bad as it sounds. - We pre-trained the encoder and did inference off-board on a workstation with a larger GPU. - We were paranoid about varying lighting messing up the network, so we did all the experiments after sunset (so long nights running experiments on the robot past 3 am) Now, we have manipulation policies that are far more dextrous, far more generalizable, and maybe on the cusp of breaking into the real world. :) (the paper:show more

NVIDIA's Jim Fan has verified from reliable sources that... the video by EngineAI is a real robot demo and not CGI. The controller is a neural net trained in the NVIDIA Isaac simulator using reinforcement learning. There’s also another video from a different angle.show more

I joined Google DeepMind to develop a research program... on the economics of AGI. Today, together with Andrew Koh (Andrew Koh) and Shane Legg (Shane Legg), we are incredibly excited to release The Economics of Intelligence, the first in a series of projects from our group on how AGI will change the economy. How do capability improvements transform the economy, and when? In this project, we define thresholds of economic capabilities, e.g., minimal AGI, based on how they transform the economy. We then track how the economy–market structure, employment, and wages–changes during the transition across capability regimes. We have seen exponential increases in what AI agents can do but the broader economy seems almost remarkably unchanged. Part of the reason is that the solo AI agent is the wrong economic object; economic transformation is an equilibrium outcome where organizations compete. Looking at how intelligence functions within organizations, transformation will depend on when two key capability thresholds are crossed: competence (gains on individual skills, e.g., coding, math) and coordination (gains on ability of agents to cooperate and work together). We show that what economic transformation looks like depends on whether AI becomes supercompetent, supercooperative, or both—the regime of artificial superintelligence (ASI). Tracking this transition shows why the real visible transformation may come from improvements in coordination—a capability most benchmarks don’t yet track. Past a certain threshold, AI capabilities can break the Coasian ceiling on firm size and lead to an economy that looks drastically different than the one we see today. Two main takeaways: 1. To understand the path of transformation, we should be tracking agentic coordination more closely. 2. The impact of AI will likely show up in wages before market structure or employment. We are really excited about this paper. Excited to hear people’s thoughts. Link:show more

System ID for legged robots is hard: (1) Discontinuous... dynamics and (2) many parameters to identify and hard to "excite" them. SPI-Active is a general tool for legged robot system ID. Key ideas: (1) massively parallel sampling-based optimization, (2) structured parameter space, and (3) active exploration based on Fisher Information to collect the most informative data in real. SPI-Active provides an accurate robot model and effectively reduces the sim2real gap. In sim2real policy learning setting, it outperforms baselines by 42-63% in various quadruped & humanoid tasks. Led by Nikhil Sobanbabu Guanqi Heshow more

THESE ROBOT “DOGS” ARE ARMED & EQUIPPED WITH BIG... DATA FACIAL RECOGNITION AND SURVEILLANCE CAPABILITY. Word is they are being deployed in LA now THEY WILL BE TURNED ON WE THE PEOPLE MARK. MY. WORDS.show more