Article Deep Learning works because it taps physical laws

Its a neat article covering some findings by Tegmark.

Its interesting that some have said quantum computers are not a good paring for deep learning. D-Wave has said the exact opposite and said its actual approach is optimal but more recently that a classical part of the machine won’t be going away. More specifically deep learning is a tuning problem where a buch of knobs need to be turned to precise positions and the a Q machine of D Wave sort is actually a optimal knob turner.

I suspect if you pair this with the Tegmark paper and digital physics you get something like the following especially with Tegmark’s magical locks and keys (his crypto remarks are a Q machine lead in): Its all a tuning or resonance problem, its like a system of tunable forks or tunable lenses. There aren’t really classical computer also no time, space or causality or even chance/not chance. Its really all tuned parings or synchronized parings. Its all quite contrived or synchronized in way the pixels on a screen are. Its like trying to study the relationship between two pixels on a screen. If you look at enough of pixel paring on the same screen you get a window. We are just looking at on going processes. Our computers especially Q Deep Learning are just tunable data lenses or maybe even simpler more like dimmable led lights.