We will observe ourselves creating AGI, and see the AGI creating a simulation to understand its own nature and its past, and that simulation will contain us, observing ourselves creating AGI; there is no way of knowing how deep into the loop we are
-
-
Replying to @Plinz
computational limits mean that can't possibly be true - simulated worlds are either much smaller or much simpler than outer worlds. First AGI likely also small, potentially smaller than humans who created it. AGI != Laplace's demon, could create paradox if it was.
2 replies 0 retweets 3 likes -
Replying to @gallabytes @Plinz
I gave a talk about this at Princeton 3 years ago. Presumably the demon is tricking us about the existing of a quantum OS. I think u need reversibility to avoid those limits (in practice that's what quantum is for) & there's the berkenstein bound as an 'empirical' comp limit.
2 replies 0 retweets 1 like -
Replying to @boyphysiker @Plinz
hmm is it actually reversible in the relevant sense? I know QM is reversible if you have the entire joint state, but is there enough info to reconstruct that in every decohered Everett branch?
1 reply 0 retweets 0 likes -
Replying to @gallabytes @Plinz
The way to make an irreversible computation reversible, is you embed it in reversible gates, like toffoli gates. You have to have a bunch of extra bits to make them perform AND OR NOT and COPY gates. Then you generate a bunch of junk along the way, extra debt.
2 replies 0 retweets 0 likes -
Charlie Bennet calls it garbage. then you have a garbage collection routine. At the end of the day, you get your output. Then you put the output in the output register and simply reverse the rest of the computation. So you’re left with the input program and a worktape which is
1 reply 0 retweets 0 likes -
now blank and the output in the output register. It does the same thing as a classical irreversible machine but with a bunch of extra stuff in the middle.
1 reply 0 retweets 0 likes -
Even classically if you just do this in a naive fashion where every classically irreversible logical gate you perform is mapped to a reversible gate but with some extra bits, then you have this funny feature, where the amount of memory space you use is proportional to the numb
1 reply 0 retweets 0 likes -
of operations you perform. This is kind of interesting because it seems to eliminate the distinction between temporal complexity and spatial complexity. Normally in ordinary classical measures of complexity PSPACE contains PTIME because you can erase stuff all along the way in
2 replies 0 retweets 0 likes
it's proportial to the number of bits you 'delete'
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.