every time Memo uses its hands to smooth out a nicely folded garment, my heart skips a beat!
Replying to @sundayrobotics
The core breakthrough behind ACT-2 is a general training recipe.
Just like LLMs, scaling data and compute gives us predictable improvements. Unlocking one Solve accelerates the timeline of the next Solve.
The same ACT-2 model is now learning many new tasks:
