You never know what’s going to shake out from pure math. Still, hunting for extremely large primes might not be efficient, even by the standards of pure math.
This is only tangentially my field, but I’d expect the numbers themselves to be much less potentially useful than the algorithms needed to find them. Since GIMPS is just throwing FLOPs at the problem through established math, it doesn’t look like an especially good approach to me.
What could be learned by getting to know more of those numbers? What’s the benefit of knowing them now over waiting, e.g., 100 years when computing power is cheaper and better algorithms might exist?
And what else could be done with the computing power?
Although you can indeed never know the outcome of research, I think we can estimate whether particular research is worthwhile.
Depends on your relative valuation of entertainment and the other things that could be done with the resources. (And, in the case of creating rather than consuming, what you expect the people playing your game would be doing if you didn’t make it.)
If playing a computer game causes (aside from the entertainment value which is in fact its main point) a little harm and no good to speak of, that’s sad but many people will deem the entertainment worth that cost.
But in the case of Folding@home there isn’t much entertainment involved; rather, people run the Folding@home client because they think they’re thereby doing something useful. So if in fact they’re doing a little harm and no good to speak of, that’s a problem.
On this topic, you may be interested in Gwern’s brutal takedown of Folding@home.
Should The Great Internet Mersenne Prime Search be considered a bad idea by that measure?
Are there any benefits to knowing prime numbers so large they can’t even be used in cryptography?
No?
Then I guess it’s a bad idea.
You never know what’s going to shake out from pure math. Still, hunting for extremely large primes might not be efficient, even by the standards of pure math.
This is only tangentially my field, but I’d expect the numbers themselves to be much less potentially useful than the algorithms needed to find them. Since GIMPS is just throwing FLOPs at the problem through established math, it doesn’t look like an especially good approach to me.
What could be learned by getting to know more of those numbers? What’s the benefit of knowing them now over waiting, e.g., 100 years when computing power is cheaper and better algorithms might exist?
And what else could be done with the computing power?
Although you can indeed never know the outcome of research, I think we can estimate whether particular research is worthwhile.
Is it a bad thing to invent or play resource-intensive computer games?
Depends on your relative valuation of entertainment and the other things that could be done with the resources. (And, in the case of creating rather than consuming, what you expect the people playing your game would be doing if you didn’t make it.)
If playing a computer game causes (aside from the entertainment value which is in fact its main point) a little harm and no good to speak of, that’s sad but many people will deem the entertainment worth that cost.
But in the case of Folding@home there isn’t much entertainment involved; rather, people run the Folding@home client because they think they’re thereby doing something useful. So if in fact they’re doing a little harm and no good to speak of, that’s a problem.