The problem with Prime Numbers

[osprey]

Not sure if this is the right section (so I've posted it on Science as well).
A multi million reward is available for the person who finds a surefire way to predict primenumbers.

This would crack everything related to encryption.

Does this make any sense? Wouldn't the discovery of such a prime-number-predicting algorithm be rather catastrophic in terms of internet business (and a bunch of other things?).
What would happen? Would they take the solution and lock it up, or would they release it to the public, thereby taking down all encryption, since they'd be much easier to crack?

[Llamatron]

How would it help crack encrypted stuff?

My understanding of encryption techniques is very limited but I don't think knowing a prime number would help to crack encrypted anything.

[fo3nix]

It wouldn't totally, no.

Public key algorithms often depend on having the product of two (rather large) prime numbers. It's assumed that it'll be hard to find those two original prime numbers from their product, and hence it works quite well. If a faster method of doing this is found, then we're doomed for now.

Bring on quantum encryption!