Pinned Tweet
some of my favorite imagery from our recent launch video is made even cooler when you know how it's made.
not ai. not photoshop. pure analog circuit bending goodness.
if you like making cool stuff, come join us at Ultra 😉
not ai. not photoshop. pure analog circuit bending goodness.
if you like making cool stuff, come join us at Ultra 😉
If you want to build robots that ship, we're hiring: [ultra.tech/careers]
2
1
30
1,917
i love open releases, but also "largest open robot dataset" seems just wrong
@AGIBOTofficial agibot world and @xdofai abc.bot are both larger iirc. probably more out there...
too many claims in robotics are overstated for the headline
@AGIBOTofficial agibot world and @xdofai abc.bot are both larger iirc. probably more out there...
too many claims in robotics are overstated for the headline
In collaboration with @huggingface, we’ve released the largest open robot dataset, alongside their first open System 1-System 2 model that can reason over step-by-step actions.
FineART is built with 40K+ demonstrations across 151 tasks and 500K subtask labels.
Together, the designed for each other, and both are available on @LeRobotHF where robotics developers work today.
0
0
18
751
not bad, but i want the robots to be faster lol 🚀
pretty impressive what you can do with not that much data using the newer @physical_int backbones
pretty impressive what you can do with not that much data using the newer @physical_int backbones
work in progress. zero humans, 1x speed, you know the drill
1
2
43
3,921
Really cool to see people pushing the limits of agentic robots. What a dark horse path to robot agi — i wouldnt have put it as a contender until pretty recently 🤔
Frontier LLMs are like the stereotype of a genius who struggles to change a lightbulb or the oil in their car.
They have great visual and spatial reasoning but they struggle to control a robot body and get it to perform simple tasks. We used tens of thousands of hours of videos to train a specialised model that’s a great handyman, not a math professor
0
0
10
657
Congrats @jang_yoel and team. Big show of confidence to make the model so easily accessible!
Today, Dream Labs is coming out of stealth with Dream-1, a robot foundation model that controls bimanual robots it has never seen, on tasks it was never shown, in places it has never been, from $3/hour.
Dream API is live for everyone, today. 🧵
0
0
20
664
Love what @JasonMa2020 says about deploying here
A pilot is one thing. Making a robot people love enough to want more is a much higher bar.
Existing customers doubling down on their fleet is the #1 validation that the product works
A pilot is one thing. Making a robot people love enough to want more is a much higher bar.
Existing customers doubling down on their fleet is the #1 validation that the product works
Robots That Predict the Future: Dyna-2 & Physical AI — Jason Ma, Dyna Ro... youtu.be/BHiYomkfoGQ?si… via @YouTube
3
1
14
1,617
really excited for this to become more available.
pi07 is much much more powerful than pi05 and will really make it way more accesible to build great robot products 🚀
pi07 is much much more powerful than pi05 and will really make it way more accesible to build great robot products 🚀
Announcing the beta release of the Pi API. We’ve partnered with companies across diverse industries to adapt our models to their robots and applications, from harvesting strawberries to assembling servers. Now we’re opening access to more teams.
4
3
40
3,053
Chetan retweeted
Congrats to @Ultraroboticsco on the raise! 🚀
Ultra and Pi have been working together to bring our models into real warehouses. We've gone from training ~1 policy per week to 8+, and improved autonomous performance from ~60% to ~95%.
Really cool to see how quickly things can improve when you're iterating on models with real-world robot data.
Ultra and Pi have been working together to bring our models into real warehouses. We've gone from training ~1 policy per week to 8+, and improved autonomous performance from ~60% to ~95%.
Really cool to see how quickly things can improve when you're iterating on models with real-world robot data.
Robotics has a distribution problem. Today, Ultra is announcing $62M in funding to solve it, including a $50M Series A led by @hiFramework.
We believe that deployment is the bottleneck in robotics.
It’s relatively easy to make a robot look impressive in a controlled harder to put a robot into the open world, make it useful, keep it running, earn the trust of the people working alongside it, and do that over and over and over again.
So that’s what we’re focusing on.
3
12
48
3,118
Chetan retweeted
Now, our robots are running in warehouses across America, doing valuable work for customers every day.
Ultra’s mission is to build the world’s most useful and deployable robots. And we’re just getting started.
Ultra’s mission is to build the world’s most useful and deployable robots. And we’re just getting started.
1
4
38
8,264
We've seen incredible progress in robotics over the last few years -- a cambrian explosion of low cost hardware options and the birth of scaling laws in robot intelligence -- yet deployed robots doing real work are still rare. At Ultra, we are laser focused on changing that.
Robotics has a distribution problem. Today, Ultra is announcing $62M in funding to solve it, including a $50M Series A led by @hiFramework.
We believe that deployment is the bottleneck in robotics.
It’s relatively easy to make a robot look impressive in a controlled harder to put a robot into the open world, make it useful, keep it running, earn the trust of the people working alongside it, and do that over and over and over again.
So that’s what we’re focusing on.
20
4
122
8,613
beautiful
Today we give Atlas dexterous hands, with the strength to take on real work, and the simplicity for mass manufacturing. Like the body, the hands are designed with high fidelity simulation in mind, to enable sim-to-real RL. If you give dexterous hands to a robot, make them do thoughts here
Blog: bostondynamics.com/blog/robot-han…
Video: youtube.com/watch?v=4whgw2…
0
0
10
581
Pretty robot 😍
We are releasing Dyna-2.1, the first Physical Agent that achieves reliable super long-horizon whole-body autonomy. It combines our brand-new semi-humanoid hardware with an agentic system built around Dyna-2 to handle ultra-long real-world workflows.
Here is an uncut footage of entire hour-long laundry room workflow, just like a human does.
0
0
9
522
Cool launch… but anyone already running the comparison against astra?
Mk1.5 excels at understanding first-person (egocentric) video. For smart glasses, this means reasoning about scenes and events over time. On EgoSchema’s hard split, a public benchmark for long-term temporal reasoning, Mk1.5 leads by 12%.
For robotics, egocentric video can also On EgoChores, our internal benchmark of household tasks, Mk1.5 localizes hands 50% better than the strongest Gemini model we tested, Gemini 3.8 Flash. High-quality hand and action annotations can turn human demonstrations into training data for robot manipulators.
0
0
2
463