Will computers ever be able to accurately predict the economy?

[Spotty J]

No, they might come reasonably close most of the time, but truly accurate prediction is an impossibility. For one thing, chaos theory applies. Small things happen in unpredictable ways, and accumulate, interact, amplify, and cause feedback loops that simply cannot be predicted. It's the butterfly effect (wherein supposedly a butterfly flapping it's wings in South America can ultimately cause a hurricane in North America ... though that's an exaggeration.)

Secondly, the computer doing the simulation cannot be as smart as the whole economy because it's PART of the economy itself. The economy as a whole is a kind of supercomputer working on a complex algorithm using all the human brains and computers that are within it -- a massive multi-processor supercomputer. When you decide whether or not to buy an item at the store, you are executing part of this program yourself with your own brain, which is networked with the rest of the brains in the economy through the monetary system -- your decisions become information that is factored into the whole program, so to speak.

So the problem there is that any computer being used is PART of this system, and provides feedback to this whole system. It can never be more than a small part of the whole. Using a bigger more complex new supercomputer just makes the whole economy bigger and more complex in its processing, making the simulation hard to do accurately. Plus if an economist has access to a new supercomputer, then everyone else also has access to similar supercomputers for their own purposes, making the processing problem vastly more complex.