PokeBot Launches RoboAPI, Using API Calls to Poke-Aurora to Control Robotic Arms
PokeBot has launched RoboAPI, driven by the world-action model Poke-Aurora. It supports five compatible robotic arms in executing tasks via natural language instructions, without the need to collect data and train separately for each task. It works similarly to a text LLM API: camera images from the robotic arm are uploaded to PokeBot's servers, Poke-Aurora returns actions, and the robotic arm executes them at 30Hz. Rohan Paul believes this is progress toward zero-shot robotics, where only the instructions need to change between tasks.
Why read this
AI score 62
The original text explains how RoboAPI transfers the pattern of calling models to robotic arms, so readers can understand the implementation path for zero-shot robotic tasks.
Topics
View original postDon't trust demos. Try it yourself. Introducing RoboAPI, powered by Poke-Aurora, our world-action model. Give compatible robots natural-language instructions through a simple API. No task-specific training. More information: https://t.co/BHzTOo4I8u https://t.co/3AVNh040Vf
View original
Don't trust demos. Try it yourself. Introducing RoboAPI, powered by Poke-Aurora, our world-action model. Give compatible robots natural-language instructions through a simple API. No task-specific training. More information: https://t.co/BHzTOo4I8u https://t.co/3AVNh040Vf
Body
The call-a-model pattern that made LLMs easy to build on is reaching physical robots.
> Until now, getting a robot model working on your own arm meant collecting data and training for each task. PokeBot’s RoboAPI lets owners of 5 supported arms skip that and send typed commands instead.
> It’s an API for robot arms, much like an LLM API for text: the arm’s camera image goes to PokeBot’s servers, and Poke-Aurora sends back movements the arm executes at 30Hz.
This is a great progress towards zero-shot robotics, the instruction is the only thing that changes between tasks.
https://t.co/7JnuCiJ59p