And Nathaniel, you’re making the point that Rule 30 is not in fact known to be universally computing. I write about this issue in *The Lifebox, the Seashell, and Soul*…you can read the section here. It goes back to Post’s Problem: “Is there a computation M such that M has an unsolvable halting problem, but M is not universal?” And the answer turns out to be yes. So we want to be careful about saying all naturally occuring computations are universal. So Wolfram waters it down a bit.

• Wolfram’s Principle of Computational Equivalence (PCE). Almost all processes that are not obviously simple can be viewed as computations of equivalent sophistication.

If you want to play it even safer, you can go for what I call the NUH.

• Natural Unsolvability Hypothesis (NUH). Most naturally occurring complex computations have unsolvable halting problems (that is, are unpredictable.)

]]>I’ve skimmed his paper, and as for founding physics, he’s got less than the string theorists. I suspect that, like them, he’s not even wrong. For one thing, physics is known to be nonlocal. That’s OK if computation time is not the same as cosmos time. That’s how I write my books: nonlocally in book time.

I do like how he derives both the Lorentz transformation and sum-over-histories from confluence; that is, computing the same thing two different ways yields the same answer. Basic physics from basic math; neat.

I’ve noticed that an age’s cosmology sometimes reflects its coolest technology. When clocks were new, the universe was a clockwork; some of the same physicists who built the Bomb also built the Big Bang; and now that computers are cool, the universe is turning into a computation. I call it the Law of Cosmic Coolness.

]]>