Machine Learning

This was a terrible article from a serial plagiarist who refuses to do work or cite sources.

But at a fundamental level we still don’t really know why neural nets “work”—and we don’t have any kind of “scientific big picture” of what’s going on inside them.

Neural networks are Turing-complete just like any other spreadsheet-style formalism which evolves in time with loops. We've had several theories; the best framework is still PAC learning, which generalizes beyond neural networks.

And in a sense, therefore, the possibility of machine learning is ultimately yet another consequence of the phenomenon of computational irreducibility.

This is masturbatory; he just wants credit for Valiant's work and is willing to use his bullshit claims about computation as a springboard.

Instead, the story will be much closer to the fundamentally computational “new kind of science” that I’ve explored for so long, and that has brought us our Physics Project and the ruliad.

The NKoS programme is dead in the water because — as has been known since the late 1960s — no discrete cellular automaton can possibly model quantum mechanics. Multiple experts in the field, including Aaronson in quantum computing and Shalizi in machine learning, have pointed out the utter futility of this line of research.