
Here’s the book to read for MPC by Profs Bryson “The Father of Modern Control Theory” & Ho @stanford @StanfordEng Applied Optimal Control a.co/d/9VsSpmu
Yann Lecun favours MPC over RL? Interesting! What about data-augmented MPC? @ylecun #nmpc2024

Co-founder&General Partner: The Venture Reality Fund/VRF; investing in Spatial AI & Physical AI 🤖||| @stanford @carnegiemellon @SJSU @NPS_Monterey

Here’s the book to read for MPC by Profs Bryson “The Father of Modern Control Theory” & Ho @stanford @StanfordEng Applied Optimal Control a.co/d/9VsSpmu
Yann Lecun favours MPC over RL? Interesting! What about data-augmented MPC? @ylecun #nmpc2024

Introducing Scene Generation on Palatial. We’ve gone from generating individual simulation-ready assets to reconstructing entire environments that robots can interact with. Our multimodal Scene Agent brings the real world into simulation, at 1:1 scale. Request early beta access at dashboard.palatial.cloud (1/3)

early experiments with ray tracing for our 4d reconstructions my favourite is shadow from semi-transparent part of the dress

Train robots in any game at massive scale. Introducing Melty Gym!

SuperMimic has won the CORL’s best paper award. Thanks for everyone’s support!
SuperMimic is open-sourced during my talk at CORL. Have fun with it: github.com/HybridRobotics/su…

When I started my PhD, quantum computing gatherings were small. By the time I shifted to AI, they filled conference halls. CoRL this year feels the same. Here are some thoughts on what comes next for robot learning.

Dexterity is the last frontier for robotics. Egocentric, dex-umi, and tele-op have so far failed to deliver the quality and scope of data required to solve manipulation. Simulations will. For the past few months, we’ve been working with @origami_robo on the first Dexterity Factory, a real2sim2real pipeline that we believe will solve dexterous manipulation. Objects go in, dexterous policies come out. By combining Palatial’s new real2sim capture system and Mad Max agent with Origami’s S1 Skill-capture gloves and robot hand, we’ve been able to close the sim2real gap on hundreds of complex tasks. We want to contribute this research to the robotics community. Next week, we’ll open source the first 100 hand object interaction dataset, and we’re just getting started.

if you could rewrite history, what would you do? bring the army home from Dunkirk as Admiral Ramsay? keep Britain in the fight as Churchill? WorldOS lets you actually do it. every move has consequences. play it free → worldos.cc/?utm_source=x&utm…


Two months ago, we introduced Atlas, an autoregressive world model with pixel-perfect camera control. Today, we're releasing Frameboy, a demo allowing you to explore worlds in real-time with a highly optimized version of Atlas, streaming from cloud GPUs at up to 15 fps.

Been comparing notes with researchers at labs and data vendors, The new data money at frontier labs is going off-screen. Coding data they can mostly make in-house now. Wet-lab runs, robot hours and a real company's inbox they still have to buy Anthropic alone paid data vendors $757M in the first 9 months of this year. Here's where it's going: 1/Bio is the hardest to fake Anthropic built its own wet lab and bought Coefficient Bio for ~$400M in stock OpenAI ran GPT-5 through 36,000 experiments in Ginkgo's cloud lab, and got 173 scientists to write a bio benchmark. Best model: 36% What they're really buying is bench time 2/Robots are growing fastest OpenAI reportedly has ~100 contract teleoperators working in shifts around the clock @xdofai sells to 20 customers, several of them frontier labs. Mecka hit a $100M run rate in June video is easy now. Touch and force are the gap 3/Then the white-collar stuff @HalluminateAI builds PE due diligence environments for 4 of the top 5 US closed labs. Best model gets 51% Google paid $10M for bankrupt Spirit Airlines' emails and Teams messages, and outbid Mercor A dead company's Slack has a price now 4/Even evals turned into a purchase. Anthropic and Accenture each plan $1B+ over 5 years for evaluators embedded at Anthropic So the scarce thing is access: a bench, a robot, a real company's history

Introducing RoboColosseum An open robot eval initiative by @huggingface @BitRobotNetwork @eaic_sg @DJiafei Access: - 100+ robots (@UnitreeRobotics G1s, @LeRobotHF SO-101s, YAMs, Franka and more) - 140+ in-the-wild scenes - Open rollout data - 10,000 FREE eval hrs Thread ↓

🤖 BREAKING: China is investing billions in the next generation of humanoid robots. Chinese companies are shifting their focus from robots that can simply run, dance and perform stunts to machines capable of understanding their surroundings, learning new tasks and adapting with less human guidance. According to Reuters, China's embodied AI sector attracted 93.5 billion yuan ($13 billion) across 322 deals in the first half of 2026 — five times the investment recorded a year earlier. The race is no longer just about building robots. It's about giving them intelligence.

LLMs accumulate a lot of declarative knowledge. That's how they can answer a large number of questions. That's also why they require so many parameters and hence so much memory. But one can replace the accumulation of knowledge by *understanding* That what JEPA world models do. That's one reason they can be useful with considerably less memory and training data than LLMs. This paper merely shows that the underlying similarity structure of the data, whether text or image (e.g. co-occurence of words in text or objects in images) uniquely determines the shape of the representation point cloud, when the encoder is trained trained with a common anti-collapse criterion like SIGReg or VCReg (not by reconstruction).

Possibly you suspected a couple of days could go by without a new release from us. Here we developed and then open sourced ultra fast weight updates between trainers and samplers for RL, even when trainers and samplers cross datacenters.
Large-scale RL has a problem: weight updates from trainers to samplers are slow and expensive We’re sharing weight teleportation: faster weight updates than existing intra-cluster techniques with internet speeds, enabling samplers anywhere (maybe even space!) Weight teleportation is so efficient (<0.23s sampler pause time, ~3s non-blocking compression + prepare), the bottleneck ends up being the GPU hardware’s decompression ASIC We unlock these gains by using GPUs for massively parallel compression of model weight changes, engineering a heterogeneous decompression and patching pipeline, along with other optimizations and overlaps. These techniques are co-designed with hardware-native features in NVIDIA’s Blackwell generation of chips We’ve open-sourced weight teleportation, so everyone can use it today

Download and have fun ⭐️⭐️⭐️⭐️⭐️
Astrocade is now on iOS Play, make, and share games right from your phone > Scroll through thousands of games > Make your own from just a few words > Post it and watch it go viral Download now for free: apps.apple.com/ca/app/astroc…