>Ethereum researchers contract AI psychosis, begin doubting basic mathematical principles, believe signatures can be magically decomposed into private keys using secret undiscovered math tricksAI top signal. also bearish for Ethereum
>mathematical superintelligence is upon usMay I see it?
If AI cracks encryption our crypto holdings are least of our worries
>doubting basic mathematical principlesP != NP has never been proven. There's no basic mathematical principle that says that crypto is uncrackable until it has.
>>62765416>also bearish for Ethereumwhere is my sub1500, NIGGER?!
>>62765462yeah but he's not talking about P != NP specifically, he's talking about some super secret embedded weakness of ECDSA that has somehow never been discovered in the past 30 years that would allow accelerated computation of a private key. They've also expressed concern over lattice cryptography. They're going whole hog into hash-based cryptography (using AI to assist them, no less)
>>62765416>Quantumclucks and cryptobaggies linked upLmao. This is the saddest case of "NOTICE ME SENPAI" I've ever seen. Face it, AI is the cool kid on the block and no one cares about your stupid blocks and 0 degree kelvin black boxes anymore.
>>62765416Having used AI for coding and we've all seen the edge cases where AI is incredibly powerful (navier-stokes), there may be something to what Drake is saying. It very well could be that all this shit goes to zero and the underlying math is broken. I say this as someone that has 95% of my net worth in crypto and hope it's not the case, but AI has the tremendous ability to go through what would take mathematicians dozens of lifetimes in just a few days. I'm probably a fool for having this much of my net worth in crypto. It's totally possible he's right.
>>62765416>Ethereum researchersthis retard is not a researcher he's a grifter
>>62765487lmao you're fucking delusionalnobody linked up, fiatcel nigger, the retard OP posted is having a mental breakdown from people ignoring his magical hypothetical super computer and being mocked on twitter
>>62765488>there may be something to what Drake is sayingthere is literally nothing to what justin is sayinghe's essentially trying to make people freak out over divination magic that he wasted a good chunk of his year on quantumkikes are desperate for relevance now that pharma bros stopped talking to them
Unironically, this brings up an interesting point. He says even in worst case scenario it'd take weeks on a GPU cluster. It seems the best thing you could do is break everything into multiple wallets. If you're coinbase with 10,000 BTC in a wallet, you're a target. If you're some guy with .5 BTC in a wallet, it's not economically feasible to break your wallet even in the doomsday scenario.
>>62765553I disagree.
>>62765556That is kinda a cool idea. Maybe like a second layer solution where you can pay from multiple wallets into multiple different wallets
call me what you want, but this shit really weakens my hands. gee, ethereum or nvidia stock next paycheck? heehee might get your funds wiped because I didn't play musical chairs with where I keep everything. fuck this.
>>62765835The problem with this is AI has been tremendous at finding implementation bugs. Second layer solutions/sidechains/extension blocks have been shredded by current frontier AI models. Look at the liquid hack, look at ZEC inflation bug, look at LTC MWEB exploit. You wouldn't want a second layer solution for more security, you'd be getting the opposite. It'd be best to just send them to multiple wallets, maybe a script can generate wallets on the fly if you're lazy about it.
>>62765852Thing is, who knows if that shit is any better? It's all just a database stored on a server. You can imagine someone hacking into the main servers and going in there and manually changing things without letting anyone know. Ultimately even stocks are just a ledger. A sufficiently advanced AI capable of covering its tracks could potentially be a nightmare scenario. Maybe gold or silver is what you're looking for where idk. At least a percentage. They have their own issues though
>>62765855Yeah but there's no point using it as a currency if you pay into a single wallet and suddenly the wallet makes financial sense to attack
>>62765485It's very unlikely that AI will find a way to un-multiply very large integers. For all intents and purposes it's a random hash function.
>>62766044exactly. it's like saying that an LLM will somehow find a flaw in the concept of arithmetic itself
>>62766044This technology is very early. In 5 years it's apt to be 100x better than it is now. It can be done with Shor's algorithm on a QC. Who is to say there isn't a classical way to do it that mathematicians missed? The thing to understand is that ECDSA is "broken" in a sense, we just don't have the machines to run the algorithm that breaks it. But there may be a classical way to derive the private key from a public key. Probably not, but nobody knows for sure.
>>62766078shor's algorithm works because you can find the period of the function by simulating all states at once and then extracting the period using a classical computation. that is why it requires a quantum computer, to create a physical state that otherwise is simply impossible to create with normal silicon materials. ecdsa is "broken" in the sense that already the private key can be probabilistically derived from any signature - the problem is that there exists no device that will ever probabilistically discover the solution, because the search space is so incredibly gigantic.
>>62766078Well I was overstating it a bit. The best algorithm is the GNFS which can factor 250 digit numbers. That is bounded by number theory and prime distributions so it's not likely that AI will do better.
