๐—ฃ๐—ต๐˜†๐˜€๐—ถ๐—ฐ๐—ฎ๐—น ๐—”๐—œ ๐—ก๐—ผ๐˜„ ๐—›๐—ฎ๐˜€ ๐—ฆ๐—ฐ๐—ฎ๐—น๐—ถ๐—ป๐—ด ๐—Ÿ๐—ฎ๐˜„๐˜€

Scaling laws mean more data equals better results. NVIDIA proved this for robot dexterity.

Training data grew from 1,000 to 20,000 hours. Success rates doubled. Labs now know the exact data needed for results.

NVIDIA released GR00T N1.7 with an Apache 2.0 license. Companies now use this for free. This mirrors the move in LLMs from 2023. A factory engineer with one GPU now runs a top model.

Money is moving fast:

Testing is also changing. The ManipArena competition scores real robots. It ignores simulators. Physical skill now determines the winners.

Source: https://dev.to/xberry-tech/physical-ai-has-scaling-laws-now-the-race-just-became-something-else-1p3d Optional learning community: https://t.me/GyaanSetuAi