There's also the possibility that any function that can be learned (most cannot, too high kolmogorov complexity) can also be effectively coarse grained and presented composably. The AI would, like Euler, just write up its discoveries in dozen book series.
-
-
Consider that we may be living in a universe where the functions that govern its dynamics can be decomposed into simple ones (that still yield observables). Even if not, the remainder will have to be treated as noise.
1 reply 0 retweets 0 likes -
That would imply a universe of non-Turing machines.
1 reply 0 retweets 0 likes -
There are only finite state machines.
2 replies 0 retweets 0 likes -
Replying to @Plinz @IntuitMachine and
I personally don't think of computers as Turing machines, although in practice it isn't a wrong statement. In a strict sense, we live in a universe of finite-state machines. It's the only model of computation that exists in the material world.
1 reply 0 retweets 0 likes -
Replying to @FieryPhoenix7 @Plinz and
So what's the argument for this conjecture?
1 reply 0 retweets 0 likes -
Replying to @IntuitMachine @Plinz and
For all practical purposes, computers are equivalent to Turing machines. But this isn't technically accurate due to the limited amount of memory. E.g. No physical computer can recognize the language L = {a^nb^n | n >= 0} because n will, at some point, exceed its memory capacity.
2 replies 0 retweets 0 likes -
Replying to @FieryPhoenix7 @Plinz and
Well finite state machines are not equivalent to finite Turing machines. Furthermore, when people talk about Turing machines, they expect finite computation. So this is nitpicking at best.
1 reply 0 retweets 0 likes -
Replying to @IntuitMachine @Plinz and
You are correct that they are not equivalent: that is the point I was making. Also, a Turing machine is guaranteed to halt for ALL regular and context-free languages, but it may or may not halt for an arbitrary recursively enumerable language.
1 reply 0 retweets 0 likes -
Replying to @FieryPhoenix7 @Plinz and
Finite state machine then be the wrong term you seek. What you mean is machines that have finite computation or Turing machines that halt. Even for FSM you can have non-halting behavior (i.e. a* ).
2 replies 0 retweets 0 likes
The halting problem is only relevant wrt to a particular description. Some system behavior is not well compressible via deterministic algorithmization.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.