From: Joscha Bach cfl. To: Jeffrey Epstein <[email protected]> Subject: Re: Date: Mon, 19 Feb 2018 21:25:12 +0000 Yes, our universe is reversible, i.e. each state seems to have exactly one possible preceding state, and history is being preserved. This means that it is probably deterministic, because it is hard to recover reversibility from probabilistic computation. (Conversely, there are many ways in which we can get apparent stochasticity from a deterministic universe.) We can implement reversible computation on an irreversible computer, but each operation that deletes a bit will have to store it somewhere: there is going to be an accumulation of entropy, of garbage bits. We can also implement reversibility on an irreversible computer by simply not performing any of the operations that delete bits (if we want to, we can keep an undo history). It is possible to build reversible Turing complete automata. On the other hand, if you take an irreversible finite deterministic automaton (like Game of Life) and let it run for long enough, it will always become periodic. In the worst case, the period is I, i.e. every dynamic structure is dead, but the period can also be extremely long. Once the automaton has entered a loop, it is reversible. I don't hold a strong ontological belief in computationalism. I am just surprised that it seems to work so well as a possible candidate theory of everyhing, and it seems to work much better than all the competitors, like string theory. Both are stubs that cannot yet recover our observables, but as far as I can see they hold the promise of doing so. Theories based on computation have one big burden: I don't know how they could explain what process leads to the emergence of the primary automaton. This first automaton can be extremely simple, as long as it is Turing complete, but it is not nothing, and it cannot bring itself into existence. Theories based on classical math are much more liberal when it comes to c