Weโve published our first technical report on Sprout.
Our paper outlines the design choices behind a lightweight, compliant humanoid platform built to operate safely in human environments and the system architecture that makes experimentation possible.
https://t.co/qCJaYQSUZ2
My favorite part of the Sprout launch video is when he demonstrates soft compliance of the robot by lifting his little leg.
Tells me they're perhaps the first ones to implement soft compliance in a humanoid product, similar to FACET, SoftMimic, etc.
Overall, the policies demonstrated, even though the behavior was basic, impressed me even more than the robot (though it is a very cuddly robot)
Meet Sprout.
Today, weโre releasing a new kind of robotics platform. One designed to move out of the lab and into the real world, closer to the people who will shape what robots become next. [1/4]
Dishoom's quality at scale is so impressive. The King's Cross branch has 340 seats; it has 15 branches across seven cities; it has lots of Deliveroo dark kitchens. As far as I can tell, it maintains a high quality bar across all of this. *And* it is hugely popular โ the King's Cross branch was heaving last night, in mid-January, and Americans love it. Incredible success!
Rolling out some awesome updates to Meta Ray-Ban Display, like teleprompter (clutch for public speaking) and EMG handwriting in early access. EMG handwriting is one of the best examples of how you and AI can co-evolve to develop over time a personalized input scheme that makes it faster and more reliable. Excited at the progress we continue to make โ๏ธ
That's the problem. People who are responsible for the issues are not the people who got laid off๐ In January, our team put down all the research we are currently doing, was (forced?) to move to GenAI <2 months before the llama 4 release deadline to help with all the post-training dirty works, figures out and fixes lots of problems after the release, and then get laid off later.
I have looked up to Yuandongโs work from afar since ELF-OpenGo. Axing fundamental research in an era where even a single unexpected breakthrough could determine winners/losers feels shortsighted.
I always thought the decline in fundamental AI research funding would happen because AI didnโt generate enough value to be worth the cost.
But it seems like itโs happening because it generated too much value. And the race to capture that value is taking priority.
Just remembering that a lot of this started in curiosity driven industry research labs.