11 demonstrations, then the robot taught itself the rest! 🎯 @Gobanorobotics, out of Nantes and London, published Toutatis v1 this month. It's a reinforcement learning engine aimed at the part of robotics that doesn't make it into demo videos: getting from "it works" to "it works 999 times out of 1,000." A 95% policy fails 50 times per 1,000 cycles. A strong lab result at 99% still fails 10 times. Production wants roughly one. Human demonstrations can't close it. They teach a robot what a person did, not what maximises success rate or removes the rare failure that stops a system running unsupervised. The engineering choice worth noting is where the RL actually runs. Gobano trains a task-specific world model from the same demonstrations, freezes the visual encoder, and caches the embeddings. RL then operates on that compact latent state instead of backpropagating through the full vision stack on every update. Around 10x faster policy training than their own end-to-end visual RL setup. → Ball picking: 11 human demos, 24/100 on the initial imitation policy, 108/108 after 854 real-world rollouts → Zip-tie insertion: 107 demos, 3,375 rollouts, 31/31 on the final evaluation → Towel folding: 99/100 → Simulation: 100% on Square and 99.8% on Tool Hang over trailing windows of 1,000 rollouts → Training runs on a gaming PC, and deployed policies run locally on a gaming laptop at roughly 3 ms per control step They also built a separate "Reliability" training mode that optimises repeated success directly, since maximising average task reward lets rare failures hide behind routine wins. Europe keeps showing up in physical AI 🇫🇷🇬🇧 🔗 Read the blog here: ~~ ♻️ Join the weekly robotics newsletter, and never miss any news →

11 Paverstone Lane. 11th overall pick in the 2011 NBA Draft. K is the 11th letter in the alphabet. @KlayThompson explains the significance of wearing No. 11 for Miami 🔥 (via @MiamiHEAT)

11 Great Christopher Hitchens Quotes. “I’ll try anything once except incest and folk dancing.” “To whom do you award the right to decide which speech is harmful or who is the harmful speaker? To whom would you delegate the task of deciding for you what you could read?” “I learned that very often the most intolerant and narrow-minded people are the ones who congratulate themselves on their tolerance and open-mindedness.” “Owners of dogs will have noticed that, if you provide them with food and water and shelter and affection, they will think you are god. Whereas owners of cats are compelled to realise that, if you provide them with food and water and shelter and affection, they draw the conclusion that they are gods.” “If someone tells me that I’ve hurt their feelings, I say, ‘I’m still waiting to hear what your point is.’ I’m very depressed how in this country you can be told, ‘That’s offensive!’ as if those two words constitute an argument.” “Those who are determined to be ‘offended’ will discover a provocation somewhere. We cannot possibly adjust enough to please the fanatics, and it is degrading to make the attempt.” “I have one consistency, which is being against the totalitarian – on the left and on the right. The totalitarian, to me, is the enemy; the one that’s absolute, the one that wants control over the inside of your head, not just your actions and your taxes.” “When you hear people demanding that the Ten Commandments be displayed in courtrooms and schoolrooms, always be sure to ask which set. It works every time.” “The four most overrated things in the world are champagne, lobster, anal sex and picnics.” “Everybody does have a book in them, but in most cases that’s where it should stay.” — h/t @CosmicSkeptic and @SteveStuWill