Hacker News Reader

Comment by ithinkso | Hacker News Reader

That's how maths works. Hashes are trivially reversible, all it takes is time, no miracle needed.

Oracles in complexity theory are a bread and butter concept

If two different x can produce the same y information is lost on the hash transformation. Finding _an x_ for your y is not the same as finding _the x_. That said, Laplace's Demon is, well, a demon. Presumably the devil can get the demon to unspool time to see what the input was and the math doesn't matter, collisions or not.

In cryptography either you only need 'an x', in cases that the original x does not matter, or if it does, then you can just as easily find all x of size less than large enough N and find 'the x', it is still solely restricted by time.

Laplace's demon, information loss, Landauer's principle etc. are at most tangentially related to the problem discussed

You understand that isn't the same mathematically as reversing the hash? If you farm the set of all possible x and try them to see what works you do not have an inverse function in your hands.

> If you farm the set of all possible x and try them to see what works you do not have an inverse function in your hands.

I'm not sure I follow, what do I have then? Could you give me an example?

> Finding _an x_ for your y is not the same as finding _the x_.

Not in all cases, but in the case of NFTs and other cryptocurrencies, it is. If a second private key fulfills all constraints of the original one (like size and all computed results so far), it is functionally equivalent to the original one.

If hashes were trivially reversible, there'd be no cryptography.

You're thinking of something like a rainbow table. Definitely not the same thing as rewinding a function that loses information.

Reply on news.ycombinator.com