AIGPT-6 Astra reportedly improves dexterity through simulation training
A user testing Astra points to pen spinning, Rubik's Cube tasks and hammer use as examples of its learning in simulation.
TLDR
A user experimenting with GPT-6 Astra says it learns and improves at dexterous manipulation through simulation training, citing pen spinning, Rubik's Cube tasks and hammer use. They describe developing that dexterity in the real world as the goal.
Combined views
15
1 Source, first seen 4h ago
GPT-6 Astra reportedly improves dexterity through simulation training
A user testing Astra points to pen spinning, Rubik's Cube tasks and hammer use as examples of its learning in simulation.
TLDR
A user experimenting with GPT-6 Astra says it learns and improves at dexterous manipulation through simulation training, citing pen spinning, Rubik's Cube tasks and hammer use. They describe developing that dexterity in the real world as the goal.