[cthalupa]

>Its pretty easy to argue that this is impossible.

Yep! It is very likely that this is impossible.

>you would need a computer running a model of the world at the sub-atomic level and somehow running it faster than the world itself unfolds. I'm 100% sure that would be impossible.

I'm not 100% sure.

The universe? Sure. We know that's impossible because of our understanding of physics - we cannot store the amount of information in the universe in anything smaller than the universe, much less compute on that information. Every single state we compute would again require the entirety of the universe to be stored, plus any transitional states in between. But I can't say with 100% certainty we'll never be able to compute everything necessary to predict how humans behave.

It might be, at some point in the future.

But my point was less about the theoretical possibility of this and more of how the fact that it currently is not, and is unlikely to be so at any point that is particularly relevant.

[codeulike]

Chaos theory is what seals the deal on this one.

[seekingcharlie]

This is called Laplace's Demon. It has been disproven by chaos theory.

[cthalupa]

>This is called Laplace's Demon.

Pretty close!

Laplace's Demon also talks about seeing the past - which my hypothetical does not. Nor am I concerned with calculating the entire universe, only the portions that would effect human cognition - just that which will be observable. Much of the universe is beyond our light cone, and unless we are in the center of the universe, almost certainly the vast majority is, and none of that needs to be calculated. Severely restricts the requirements vs. Laplace's demon.

>It has been disproven by chaos theory.

Not quite! Chaos theory relies on imperfect information. It does nothing to disprove Laplace's Demon, which involves perfect information. For chaos theory to be applicable, there has to be minor variations between the initial conditions. Chaos theory is not a theory that we cannot predict things because they are unpredictable, it is a theory that says that minor variations at the onset can result in huge differences in the eventual outcome. These problems only exist if you do not have perfect information.

Thermodynamic irreversibility disproves Laplace's Demon's seeing into the past portion. The Copenhagan interpretation of quantum mechanics also puts a nail in the Demon.

Things get interesting if we are ever able to create a 300 qubit quantum computer - at this point, you can perform more calculations in an instant than there are atoms in the known universe. Is that granular enough to accurately determine human behavior?

The whole computer thing is more an aside than anything, though - just providing a hypothetical in which case free will not existing would actually have practical application on how humans live. Without it, I can't think of any reason why we should behave in any other way than we do, regardless of the underlying truth.

[codeulike]

Lets say you have that 300 qubit quantum computer, and you only need to simulate your own light cone. Trouble is, that quantum computer is also in your light cone. So to accurately simulate your light cone, your simulation has to include itself, recursively. Hence: not going to happen.

[cthalupa]

Not if you're only attempting to prove or disprove free will and run the simulation after the fact. If you can get perfect knowledge of the starting conditions for everything in the light cone, run it, and then compare it against what actually occurred.

It might not work for telling the future, but if you limit the computation to see if it accurately models what is now history, you don't need it to simulate itself.