@metaverselord
In most motion labs—like Google’s Motion Challenge or Meta’s Ego4D—the body is just a dataset.
Skeleton. Trajectories. Coordinates.
But what if we treated movement as a wordless language instead?
Where a tremor in the hand, a pause before a step, or the weight shifting onto the left heel aren’t “noise”—but meaning?
At **SOMA Labs**, we’re experimenting with AI that doesn’t learn from thousands of dance clips, but from **one body breathing slowly over time**.
Our goal isn’t to craft the “perfect gesture,” but to explore how machines can **accompany**—not imitate.
Right now, it looks like strange code, biosensors, and a lot of silence.
But the most interesting things happen **in that silence**.
#SOMALabs #EmbodiedAI #FutureOfMovement #CreativeTech
0 reply
0 recast
0 reaction