Commentary on X
@chris_j_paxton The intelligence is already showing promise. The bottleneck is speed. If future models can make the same decisions 10–100× faster, this could work on real robots instead of only in slowed-down tests.
9:28 AM · Jul 14, 2026@chris_j_paxton 100x time dilation and "almost". Obviously this isn't the way to go. Interesting experiment to verify what we kind of already knew anyway.
9:56 AM · Jul 14, 2026@chris_j_paxton It's impressive how LLMs are proving to be capable in fields we never even thought of before.
9:48 AM · Jul 14, 2026Honestly its kind of impressive that opus 4.6 can almost keep a humanoid robot standing on its own (given massive time dilation)
9:20 AM · Jul 14, 2026Combined views
47.3K
2 posts, first seen 14h ago
244 likes10 comments