February 20, 2025: Figure AI introduces Helix, a new generalist VLA model for humanoid robots.

🇬🇧 Hello! 🙂

FIGURE INTRODUCES HELIX, A VLA MODEL FOR CONTROLLING HUMANOID ROBOTS.

This video was released by Figure AI, the company behind the Figure 01/02 humanoid robots. Here we can see how two second generation units (Figure 02), interact with a human who brings them a shopping bag to put partly in the pantry and partly in the fridge.

For us humans, this seems a minor action, but the logical process is complex for machines. As you can see, the androids are a bit slow, especially in interacting with each other, but think about what is behind what we are seeing.
The new HELIX model allows the two androids to interact, share tasks, catalogue objects (these objects are first seen by the androids), all on the fly. Helix is a generalist VLA (Vision-Language-Action) model, designed for Figure 02 type androids.

A greeting and see you soon!
Pipplo 🙂

Credits: original video courtesy of Figure AI Inc. Full AI processing by PipploIMP.
Credit for background audio: "Humanoid Robots" by PipploIMP, Copyright 2024 all right reserved.

Documentation:
https://www.figure.ai/
https://youtu.be/Z3yQHYNXPws

Tags:
#HELIX #VLA #VisionLanguageAction #Figure02 #FigureAI #HumanoidRobot #Android #ArtificialIntelligence