Best progress on pre-image attacks?

William Entriken

12/12/23, 6:58 PM

What is the best progress (i.e. most bits) in crafting messages with a chosen hash image for modern hash functions?

Blockchain networks with proof of work are basically brute force machines for pre-image attack on keccak with a target image of 0x000.000.

So I was looking for some idea of what the "best hash" ever found was.

0 + 5

preimage-resistance

forest

12/13/23, 12:09 AM

It really depends on the specific hash. There are hashes which do not have any known preimage attacks more efficient than a generic one, but that doesn't mean they're _better_ or more secure.

William Entriken

12/19/23, 4:31 AM

Right. Just because of Bitcoin we have ASICs that are tuned specifically for it. Even if the ASIC was a little smart about rejecting some hashes it knew wont have zeros at the front that would be an improvement over naive algorithm. And it would be worth good money.

forest

12/19/23, 4:32 AM

All those are are enhanced brute force tools. It's not an actual preimage _attack_.

William Entriken

12/19/23, 7:14 PM

If aborting a wrong hash is notably faster than computing a potential hash then it's an attack, regardless of O-notation

forest

12/19/23, 11:51 PM

That's not quite what it does. The ASICs use a different type of adder which isn't always accurate but is _much_ faster than a typical full adder. It's nothing more than a really, really fast brute force tool that sometimes makes mistakes, but the mistakes it makes are worth the performance improvement.

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: Best progress on pre-image attacks?

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.