Robots that feel right.
San Francisco, CA
Joined April 2025
- Tweets87
- Following2
- Followers5.6K
- Likes857
Enigma retweeted
What happens if you connect a hundred AI-powered robots to the internet and just let people give them any instructions they want? @Enigma_AI founder @j0nathanj joins us to talk about his experience deploying a fleet of robots on the internet, keeping them operational, and watching people truly test the generality and robustness of frontier robotics models.
Watch Episode 109 of RoboPapers with @chris_j_paxton and @ruijie_sg today!
But the race condition was only one part of the rabbit hole.
Running a large fleet of these arms pushed us into understanding the STS3215 way beyond what we originally intended:
Failure modes, calibration, bus behavior, recovery, timing quirks, and a bunch of undocumented behavior.
So we wrote it all down.
4/5
And now we’re sharing the rabbit hole.
The race condition, what we learned about the STS3215, the failures, the debugging techniques, and the tooling we built around all of it.
Full research:
enigma.inc/posts/sts3215-ser…
Tools:
github.com/Enigma-Incorporat…
Have fun breaking your servos.
5/5
We have partners across different industries: retail, logistics, security, health, and more – building towards intelligence in the physical world.
One particularly exciting use of Enigma’s models is through early collaboration efforts with researchers at @CedarsSinai aiming to explore the future of physical AI in their fully autonomous robotic ultrasound solutions.
We're thrilled for the future.
10/10
Enigma retweeted
Real footage of someone online using @Enigma_AI’s Scientist AI robot
(Please don’t blow up our warehouse 🙏)
Drop #3 is here: 2 new types of robots are now LIVE online. Those are real robots.
1. Scientist robot: A robot that does real chemistry experiments.
2. Bomb-Defusing robot: A robot on a Special Ops mission to defuse a bomb.
Try them out for yourself at robots dot online.
Today is Drop #2: Research.
Robotics foundation models need a lot of data, and a lot of engineering even for tiny adjustments.
This feels wrong.
Today, we share with the world a sneak peek to our research - training more efficient robotic foundation models.
1/3
Today is Drop #2: Research.
Robotics foundation models need a lot of data, and a lot of engineering even for tiny adjustments.
This feels wrong.
Today, we share with the world a sneak peek to our research - training more efficient robotic foundation models.
1/3
2/3
Research post: enigma.inc/posts/obsessed-en…
We share "The Obsessed Encoder", a phenomena that describes a latent space misallocation that leads to inefficiency.
The phenomena occurs in state-of-the-art self-supervised models, such as LeJEPA, DINOv3, and LeWorldModel.
It limits how efficiently they use their latent space.