Saturday, August 16, 2008

What difference will quantum computing make to me, anyway?

I think, broadly speaking, a lot of what's written about quantum computers is somewhat wrong-headed. It will make a difference to cyptography (in that it will be easier to have "unsnoopable"** communications). It will speed up some algorithms dramatically.

**of course all that means is that the "snooping" shifts to other parts of the process.

It's usually this last thing that gets the attention. What quantum computing essentially does is give you the potential to "halve the exponent". If something was going to take something on the order of 2^60 calculations, then if you're lucky, you may be able to reduce that to on the order of 2^30 quantum calculations. And if the exponents happen to be of that order, that could be enormusly useful. It might be possible, for example, to reduce a computation from centuries to maybe days or weeks (and ordinary parallelism could reduce that further, of course).

But there aren't all that many calculations that matter to us right now that have exponents of that order (well, actually there are a lot, but as a fraction of calculations we want to do, not so much). Most calculations are either much, much smaller or much, much bigger.

There's not much to be gained in taking a calculation from 2^400 to 2^200 - it's still going to take longer than you likely have.

So in the region past the outer limits of what we can practically do with ordinary parallelism, quantum computing will make a big difference - it will revolutionize what we can do with certain kinds of calculation. Most of those kinds of calculation won't impact you directly (indirectly, sure - in things like design of drugs and such) in the sense that you probably won't be doing much in the way of calculation you weren't before. At least not to begin with.

I think the big impacts on our personal lives will come in ways we can't even anticipate right now. The software that will change everything - maybe even decades after we have quantum computers as things we can buy - we don't know what that's going to be like. We won't until we have quantum computers, until we have coding environments and programming paradigms that let us think about things in ways we don't right now. Until we've had a generation of people who grew up with quantum computers.

Because when that exponent-halving is available, we're going to invent new problems that lie in that space above what we can do right now with parallelism, but that quantum computers can do. And we won't know what they are until we're there.


Dana Hunter said...

Indeed. Which, I might add, is one of the most frustrating things about trying to write SF: having to try to think in ways that we can't currently think, because our technology is a stone axe compared to what these folks've got.

One thing is beyond doubt, though: quantum computers will be used to make more interesting porn. ;-)

valhar2000 said...

I predict that video games will be among the first consumer applications that will make use of this technology.

They do seem to be the most computationally-intensive pieces of software in general use nowadays.