Figure AI’s Helix Lets Robots Think, Move, and Adapt Like Humans

Helix operates using two main components. One processes language and vision, helping the robot understand commands and plan actions. The other translates those plans into physical movements, which ensure smooth and precise control.

This combination allows robots to recognize objects and figure out how to handle them without prior knowledge.

Did you know?

Want to get smarter & wealthier with crypto?

Subscribe - We publish new crypto explainer videos every week!

Crypto Token VS Coin (Animated Explainer & Examples)

SUBSCRIBE

ON YOUTUBE

On February 20, Figure AI shared a video of two robots sorting groceries. One passed items to the other, which placed them in either a fridge or a cabinet. Despite having no prior information about these specific objects, they correctly identified which items needed refrigeration and which could be stored elsewhere.

Brett Adcock, founder of Figure AI, stated on X:

Like a human, Helix understands speech, reasons through problems, and can grasp any object - all without needing training or code.

According to Figure AI, this ability comes from its Vision-Language-Action (VLA) model. This system connects visual recognition, language processing, and movement control, allowing robots to work together and adjust to new situations in real-time.

Figure AI trained Helix on about 500 hours of robot demonstrations and used an automated system to generate natural language instructions for each action. The final model runs on built-in processors, making it ready for commercial use without external computing power.

On February 18, Elon Musk's xAI launched Grok 3, its latest AI model. What are Grok 3's key highlights? Read the full story.