Pinned Tweet
After operating in stealth for the last 18 months @rhodaai , we’re excited today to finally show the world what we’ve been working on. We believe we’re on a path to physical AGI with the launch of our brand new foundation model, the Direct Video Action (DVA) model.
Jagdeep Singh retweeted
What can a robot do without vision?
We gave our robot a sense of touch and showed it can recognize and retrieve objects with its cameras blindfolded.
Introducing Rhoda’s new tactile sensor—it feels pressure, texture, vibration, slip, and temperature, just like your fingertips.
Jagdeep Singh retweeted
More industries. More real customer work.
Introducing RhodaAtWork, a series on the customer workflows our robots are learning to handle across manufacturing and logistics.
First up: unpacking, scanning and stacking component reels for electronics manufacturing.
Jagdeep Singh retweeted
At Rhoda, we care deeply about the science of pre-training for robotics.
In one of the most rigorous studies of its kind, over thousands of trials and hundreds of hours of robot evaluations, we show how scaling web-video pre-training leads to better real-world robot performance.
Jagdeep Singh retweeted
Does scaling pre-training on general web video improve a complex manipulation task in real deployment?
We scale model size and pre-training compute, and test on one industrial task.
Yes. The better a pre-trained model predicts web video, the better its post-trained policy. 🧵
ALT Seven pre-trained video-model checkpoints on one chart. Task performance on a real robot (at-speed completion rate: every requirement met within 100 s, 0 to 100%) rises with pre-training quality (DINO FD on held-out web video, log scale, axis reversed so better is to the right). XS, S, M, L are model sizes. The four M points are one model at 0.08×, 0.18×, 0.37× and 1× of its pre-training compute.
Jagdeep Singh retweeted
If you are like @RhodaAI robots and like watching YouTube, the full video is here: youtube.com/watch?v=QTRbW1YP…
Jagdeep Singh retweeted
Teaching a robot a new task typically means stopping operations, collecting teleoperated demonstrations, and retraining. That process takes hours at a minimum. We wanted to know if we could collapse it to seconds — from a single human demo, on the fly, no retraining required.
Early research preview: we can.
Jagdeep Singh retweeted
Here’s something we’ve never seen done before.
Real-world tasks are long and ambiguous. Solving them requires visual memory and state tracking. Most robot policies only see the last few frames. Ours doesn't.
We put our DVA, FutureVision, to the perfect testbed: the shell game 🐚. The DVA nails it.
More curated training data does not produce better generalist robot models. Less curated internet video does, because what you're actually teaching is intuitive physics, how things move. Once a model has that, it generalizes with very little task data. Full conversation with @BHeater at @a3automate below. bit.ly/4s4cbvD
Jagdeep Singh retweeted
1/ We are speed running industrial robotics.
It took us just 19 days from the first day of data collection to filming a 2.5-hour continuous run of our model autonomously breaking down industrial containers — zero human intervention.
The data efficiency of our DVA model is fundamentally changing how fast we bring robots out of the lab and into the factory.
Autonomous operation with 3 hours of data collection at a customer factory.
Every company I've built comes back to the same four things: a big unsolved problem, a differentiated solution, a world-class team, and early customer validation.
With @rhodaai , the problem is robots that fail outside the lab. The differentiator is internet-scale video pretraining. The team is exceptional. And the traction is real — strong customer interest and evals with 100% autonomous operation at a major global automaker.
I talked through the full journey on the Dealmakers podcast.
Link in comments →
Our research team at @rhodaai comes from a bit of a different background than traditional robot companies. We're surrounded by an amazing team of computer vision and generative AI experts. Not surprising then, that this team took an unconventional -- and extremely promising -- approach to addressing what we believe is the biggest unsolved problem in robotics today: how to get robots to work in the real world, not just in controlled lab settings. The result: the Direct Video Action Model, or DVA. To learn more, check out the deeper technical blog post our team put together here: rhoda.ai/research/direct-vid…
Jagdeep Singh retweeted
Happy Tuesday. On today's show:
- @omooretweets (a16z)
- @adamgoldstein13 (Archer)
- @MaxJunestrand (Legora)
- @startupjag (Rhoda AI)
- @davepaffenholz (Juicebox)
- Allan McLennan (PADEM)
See you on the stream.
Deals Deals Deals, Meta Acquires Moltbook, BYD F1, Thinking Machines, Project Hail Mary, Goldman's Short xitter.cf/i/broadcasts/1DGLdvWnW…
When it comes to robotics, the gap between working in the lab and working in the real world is larger than most people think
#robots #embodiedAI #physicalAI