Deep learning and computational science’s theoretical efficiency limits
A user argues that today’s best human solutions fall far short of theoretical efficiency limits and believes deep learning can efficiently simulate the universe.
TLDR
A user argues that today’s best human solutions in computational science are much less efficient than theoretical limits allow. They believe the universe can be simulated efficiently and that deep learning can do it all. They also expect few “verifier bits” to be needed, though more may make things faster.
Combined views
—
1 Source, first seen ago
