hypothetical discovery (what to do with it?)

[Metric]

It is an interesting fact that factoring large numbers is a comutationally difficult task. So difficult, in fact, that the cryptography of the world hinges crucially on this being the case.

Interestingly, though, this situation makes people (people who need cryptography) nervous for a few reasons:

1) Computing power has been increasing exponentially for quite some time, meaning secrets "now" probably won't be secrets ten years from now.

2) If quantum computing ever takes off, the current system is bound to collapse immediately since there are efficient algorithms for factoring large numbers using quantum computers (which don't exist yet).

3) It has not even been proven that factoring large numbers (using classical computers) has to be computationally hard. That is, for all we know, someone could come up with (or has already come up with) an efficient classical factoring algorithm, meaning that there are no secrets.

Point 3 raises an interesting and surprisingly non-trivial question. Suppose you discover an efficient classical factoring algorithm. What do you do with your discovery that maximizes your profit, minimizes your chances of getting killed, minimizes the chance of theft by the government (i.e. it becomes classified government info), minimizes (or maximizes, depending on your personality) meltdown of the world's economies, or otherwise takes into account your particular priorities?

I have been told that the best strategy is simply to "post it online immediately." But I have serious doubts that this is really anywhere close to the optimal course of action for just about anyone.

[Borodog]

1. Publish.
2. ????
3. Profit!

[thylacine]

FWIW a few years ago a paper came out with a polynomial time algorithm (input size ~ number of digits) for deciding whether or not a number is prime. (I don't remember hearing if it was ultimately deemed correct.) But while it can tell you a number is composite, it does not find the factors.

Interesting question you raise. I'll cross that bridge if/when I come to it.

[r3vbr]

1.Break sbrugby's Fulltilt password and transfer all monies to my account.
2.Use proceeds to buy a small country or some other uber-balla stuff.
3.Profit!

[arahant]

I'd just sell it to the US government. If I'm going to profit significantly from it, I'd obviously have to control distribution EXTREMELY tightly. So I need a motivated buyer with a lot of money.

If one had moral qualms about that buyer being the government, you could always release it later.

[AWoodside]

[ QUOTE ]
I'd just sell it to the US government. If I'm going to profit significantly from it, I'd obviously have to control distribution EXTREMELY tightly. So I need a motivated buyer with a lot of money.

If one had moral qualms about that buyer being the government, you could always release it later.

[/ QUOTE ]

Unless you could manage to conduct this transaction completely anonymously somehow, you would be killed immediately afterwords. Having access to this algorithm would create such an extreme power differential between the US government and all other governments that it would be insane for them to leave a potential leak like you alive.

If I discovered this I would be extremely, extremely cautious. I would use it to slowly accumulate massive amounts of wealth over the course of years, probably syphoning it out of government agencies, and at the point where I felt I could set up a private security force that could either hide me from or protect me from the largest governments I would start using it to deconstruct all the world institutions I view as evil.

[Metric]

[ QUOTE ]
Unless you could manage to conduct this transaction completely anonymously somehow, you would be killed immediately afterwords. Having access to this algorithm would create such an extreme power differential between the US government and all other governments that it would be insane for them to leave a potential leak like you alive.

[/ QUOTE ]
Yep. Dealing directly with the government would be an efficient way to commit suicide.

[Metric]

[ QUOTE ]
1.Break sbrugby's Fulltilt password and transfer all monies to my account.
2.Use proceeds to buy a small country or some other uber-balla stuff.
3.Profit!

[/ QUOTE ]
I understand that you are just having fun with this, but people should keep in mind that personally having the algorithm does not mean that you have personally broken all codes and can manipulate all networks. It only means that an organization with sufficient resources can now break all codes, etc. if you give it to them.

[surftheiop]

Thinly veiled brag post... we know you have the algorithm, mods plz move to BBV

[soon2bepro]

1. assemble a team of trustable partners/co-workers

2. get tons of money

3. become God(s).

[Jcrew]

I would write a set of scripts uploaded at various free internet connections to free web hosting websites. The scripts would send a massive amount of emails detailing the discovery to various individuals and orgranization if their timers are not reset every 6 months. At that point, I would proceed to negotiate with the US government.

[Metric]

[ QUOTE ]
I would write a set of scripts uploaded at various free internet connections to free web hosting websites. The scripts would send a massive amount of emails detailing the discovery to various individuals and orgranization if their timers are not reset every 6 months. At that point, I would proceed to negotiate with the US government.

[/ QUOTE ]
I like this to some extent, but one major problem is that the government can step in and say "your research is now classified" the moment they realize you have the algorithm, and you have no alternative but to deal with the US government, who now has a tremendous bargaining chip. If they brand you a national security threat or terrorist (for essentially holding them hostage), they will take steps to make your live as miserable as possible to coerce you into giving up your scripts. If you do release your info to the public, you now have explicitly broken the law and are likely to be thrown into prison by a royally pissed-off government. The only upside is that you're pretty much guaranteed that the government won't bump you off personally.

So this particular course of action depends crucially on how effective you are in dealing with government pressure over the long term. I personally think there is just too much that could go wrong with this approach.

[Justin A]

http://www.impawards.com/1992/posters/sneakers.jpg

[FortunaMaximus]

[ QUOTE ]
So this particular course of action depends crucially on how effective you are in dealing with government pressure over the long term. I personally think there is just too much that could go wrong with this approach.

[/ QUOTE ]

And if you had already implemented or are capable of using the technology beforehand, you have an enormous weapon in which to counteract any threat, and a huge negotiating tool.

Me? Meh. I'm of the opinion the only secrets you have are in your head.

http://img156.imageshack.us/img156/1299/capncrunchwhistlepo5.jpg

With the right people, there certainly could be a case made to find utility in such a tool. <shrugs> POV, I guess.

[HP]

[ QUOTE ]
1) Computing power has been increasing exponentially for quite some time, meaning secrets "now" probably won't be secrets ten years from now.

2) If quantum computing ever takes off, the current system is bound to collapse immediately since there are efficient algorithms for factoring large numbers using quantum computers (which don't exist yet).

[/ QUOTE ]Taking a wild guess here..

Couldn't you use the increase in computer power to create harder to crack codes? So give it a few days or whatever, and we are back to normal?

Like in Re: to factoring numbers, if we could factor numbers faster, we could find harder to factor numbers too right?

[Metric]

[ QUOTE ]
[ QUOTE ]
1) Computing power has been increasing exponentially for quite some time, meaning secrets "now" probably won't be secrets ten years from now.

2) If quantum computing ever takes off, the current system is bound to collapse immediately since there are efficient algorithms for factoring large numbers using quantum computers (which don't exist yet).

[/ QUOTE ]Taking a wild guess here..

Couldn't you use the increase in computer power to create harder to crack codes? So give it a few days or whatever, and we are back to normal?

Like in Re: to factoring numbers, if we could factor numbers faster, we could find harder to factor numbers too right?

[/ QUOTE ]
The point is that if the problem becomes easy to solve, you are betting everything that someone doesn't have more computing power than you do, or is more patient than you are. This is almost always a bad bet. For example, we could make a cryptography protocol based on adding two numbers. But adding two numbers is easy to do, so simply making two numbers so stupendously huge that most potential hackers won't bother to try to add them it is both impractical and unsecure.

What you need for cryptography is a problem that is computationally hard to solve, but easy to generate examples. Factoring is like this -- you just multiply two big primes and so you yourself instantly know the factors. But someone who didn't personally generate the number is going to have a hell of a time factoring it. Unless, of course, they come up with a polynomial-time algorithm to do it.

[Jcrew]

Haven't been into crypto for awhile but can quantum computing break elliptical based encryption just as easy?

[Metric]

[ QUOTE ]
Haven't been into crypto for awhile but can quantum computing break elliptical based encryption just as easy?

[/ QUOTE ]
I don't know, actually. It could be one of those things that is easy to do, but nobody has bothered to write down an algorithm for it yet.