Travis Armstrong

Head of Robot Operations at Rhoda AI.
Former Google DeepMind, Google X, Jabil. Harvest Automation, Adept Technology

[Travis Armstrong @RandomRobotics]

@RhodaAI Humans don't rely solely on vision, why should robots? Enabling our robots to learn and use touch unlocks even more opportunities.

[Travis Armstrong @RandomRobotics]

Humans don't rely solely on vision, why should robots? Enabling our robots to learn and use touch unlocks even more opportunities. Really exciting!

[Rhoda AI @RhodaAI]

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.

[Rhoda AI @RhodaAI]

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.

[Rhoda AI @RhodaAI]

Internet video is diverse, passively collected, and virtually inexhaustible.

Previously, we showed that pre-training purely on web video can enable robots to solve complex real-world tasks.

Now, we show how that approach scales: with both model size and training duration.

https://t.co/fv7cbcoEpD

[Travis Armstrong @RandomRobotics]

*sigh*

[Travis Armstrong @RandomRobotics]

@rhodaai Sometimes skill in games unlocks skills in the factory. Having a visual memory and object permanence is important for completing tasks when your view of the workpiece is obscured.

[Rhoda AI @RhodaAI]

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.

[Rhoda AI @RhodaAI]

To bring generalist intelligent robots to the real world, we have to overcome the data scarcity problem.

At Rhoda, we are solving it by reformulating robot policies as video generation.

Today, we introduce the Direct Video-Action Model (DVA)

[Rhoda AI @RhodaAI]

Observe. Predict. Act. Repeat. In closed loop. We call this the Direct Video-Action (DVA) model, and it’s why Rhoda can operate beyond lab conditions in real production environments.

[Travis Armstrong @RandomRobotics]

@startupjag @rhodaai Seeing AI robotics move out of the lab and actually handle those edge cases autonomously is a massive step forward for the automation world!

[Rhoda AI @RhodaAI]

The gap between robotics in the lab and robotics in the real world has been one of the hardest unsolved problems in the industry. We’re excited to come out of stealth and show the research community how we’re tackling the issue.

Bloomberg article in comment.

[Travis Armstrong @RandomRobotics]

@rhodaai Robots shouldn't just work in a lab, they need to work in the real world. I've deployed them in factories, warehouses, and farms! Excited to be part of the Rhoda AI team as we announce today, bridging the gap between research and the real world.

[Rhoda AI @RhodaAI]

03.10.26

[Travis Armstrong @RandomRobotics]

Too many robots end up in the “Automation Graveyard” but with AI we can build intelligent automation that lasts. 
#Automation
#EmbodiedAI #Robotics

[Travis Armstrong @RandomRobotics]

Humanoids in the home make for viral demos, but factories are the real proving ground. The path to a humanoid in your house starts on the industrial floor, where environments are engineered for success and the ROI is proven. #robotics #automation #AI

[Travis Armstrong @RandomRobotics]

The real business case for robotics: cheaper, faster, better. The rest is just a demo.
#Robotics #AI #Innovation #Automation #Tech #Operations

[Travis Armstrong @RandomRobotics]

Great tour of the @GoogleDeepMind robotics lab! 🤖🦾
Proud to have led the team that built the ALOHA workcells and scaled up data collection efforts for the first Gemini Robotics models. Excited to see the fleet in action! 🚀#Robotics #AI https://t.co/K1ABL3o7Zp

[Travis Armstrong @RandomRobotics]

@TripIt You guys need an AI Image generator that I can use to create trip images. It would also be nice to save a library of images I've used in the past so when I go to an annual event I can re-use the image, rather than uploading it again.

[Travis Armstrong @RandomRobotics]

@coreylynch curious: "never seen these items"- is it trained on similar bottles of ketchup & condiments, does the policy expect one, so it asks a vision model "where is the condiment?" & then it's applying "pick up the condiment" that would work on any 20oz bottle?