Robotics Has Been Stuck for 70 Years — Deepak Pathak, Skild AI
Source
youtube.com
Author
AI Engineer
Date
Context
Skild AI, a robotics company, is trying to reduce the amount of robot-specific demonstration data needed to teach new skills. Human videos offer abundant examples, but they omit forces and touch feedback, and human limbs move differently from robot hardware. Skild says its model bridges these differences and can learn new skills from video with less than an hour of robot data. The aim is to reuse human demonstrations across different bodies, reducing dependence on teleoperated robot data. That is the company's reported learning approach; it does not establish reliable performance on every task.
Key takeaways · AI-distilled
Pathak estimates that collecting teleoperation data at one example a minute would take the entire US population more than a century to reach GPT-3 scale, which is why he says robotics has no internet of data.
Skild AI's 'omni-bodied' recipe trains one model for any robot and task: pre-train on scalable data such as simulation and human video, post-train on teleoperation, then improve through a deployment flywheel.
Demos include inserting AirPods with a simple gripper, learning from human video with under an hour of robot data, and cooking omelets on $4,000 arms using only a camera.
Pathak frames adaptation as a safety property: after its other legs are disabled, a robot learns to walk on two legs within three tries.
Terms in this piece · Glossary
pretraining — The first, biggest phase of building a model: training it on enormous amounts of text so it learns language, facts, and reasoning in general.
fine-tuning — Taking a trained model and training it a bit more on your own examples so it gets better at one specific job.
videoYour KV Cache Can Spot Attacks on Your AI Agents - Ari Azarafrooz, Intrinsec AI
videoWill AI Agents Pretend to Be Human to Get Past a Check? - Maty Bohacek, Stanford
videoThe One Thing Every AI Agent Framework Agrees On - Mislav Ivanda, Daytona
videoAI Writes Code Faster Than Uber Can Merge It - Preetam Dwivedi & Manjari Akella, Uber
