AI's capabilities make me wonder about Quantum Computing's impact

It feels like expert audits don’t really matter anymore. Not long ago, we relied on projects that got checked by well-known audit firms as a solid way to figure out if we could trust them or not. Quantum computing is still a looming threat, and while we haven't seen anyone successfully hack or jailbreak a Blockchain with QC yet, it seems like AI is now the one identifying weaknesses in Blockchain systems. Basically, AI can hack Blockchains now. It's pretty disheartening, and I’m still trying to wrap my head around how we got to this point. Take Zcash as an example. This privacy coin, which we considered trustworthy, has had a flaw for four years, and AI found it out in just a day. That really bums me out. What’s the point of having Devs and teams if it turns out it could be better if they just let AI find these issues? If AI is capable of this, a lot of other coins could be riddled with pretty serious flaws right now. 1. I really hope audit firms start using advanced AI too, so they can catch these kinds of problems. 2. Or Blockchain teams and developers should start utilizing strong AI tech to review their systems. For the first time, I gotta admit I see the potential in AI; I never realized it was this powerful.

It could have been worst. At least AI was used in this case to find out the bug and fix it, instead of being used to steal funds from a wallet. As you can see, AI is being used in a positive way by developers to find errors and improve security of their systems. And that is the most likely scenario you are going to see from now on: AI being used to build and to improve, instead of being used to destroy and to steal.

I will let this reply from NotFuzzyWarm do the talking and also give you a good reply It’s now becoming too much, there is no Quantum computer apocalypse going to happen, the Qc thread is now a Qc panic obsession that can be cured doing your research on existing Qc topics.

The problem is always with us humans. AI is just a tool but how and what we use it for depend on us entirely. Ai is not attacker but the bad guys are using it negatively. Some devs now use AI tools to run a check on their code internally but attackers also use it from different angle. Now it's a matter of who know how to use it the most.

You seem a little confused about what these two things are and how they work. Sure, AI can analyze code and find weaknesses in it to exploit - this is not limited to the blockchain, but similar vulnerabilities have been found in operating systems, websites and plenty of places that we otherwise thought were secure because a human couldn't think or program enough variants to throw at it. The most cutting edge forms of AI are starting to break code like that, but they still sit on an underlying computer system. Quantum computing would be a massive speed increase to current computer systems so instead of AI taking days or months to trawl through code, it could do it in seconds. That is the distinguishing factor here, the speed of processing data.

The auditors works with very limited budgets and limited human attention span and limited time and the audits were never the guarantee that the projects were free from any sort of vulnerability and even though we have very high end projects they still can have bugs that can remain hidden for many years,  what AI is actually doing is it is increasing the speed of analysis and it can easily review large amount of codes it can compare patterns acrose repositories and test scenario's much faster than human beings that does not mean that AI is hacking blockchain it is just helping the researcher discover their weaknesses that they often overlook. Example of Zcash shows that even very trusted project can contain flaws for a long period of time without even being noticed by anyone , the developers of it should take it as another security tool and there auditors security researchers the all should use AI to strengthen their system before attackers can exploit weaknesses and as for quantum computing it is a long term

If there are vulnerabilities within any system, there is a possibility that it would be exploited. And those who have the skill and resources would employ tools that would make the job easier and that includes the use of AI. But this does not mean the same thing as "AI can now hack blockchains" like you've said, so you're wrong. The "threat" of quantum computers is not real yet. And yes, AI can be an important tool in identifying vulnerabilities because of how fast it can analyze data, but the vulnerability has to exist first.

We won't always be that lucky. What about next time? Thanks to AI everybody is expert at everything now. Some the US government recently banned some AI model because they said it was way too capable. China will close the gap and let everybody use them highly capable AI models eventually. Then the real shitshow will begin. So havent seen anything so far.

Why these projects have dev teams if dev teams are not using these models on their own code? And I do not mean that as disrespect. The quantum computing angle, however. In March, Google released a paper that claimed one had to install less than 500,000 physical qubits to break ECC. Previous estimates were in the millions. It is not quite there yet, but the distance is closing and. Despite NIST finalizing the standards in 2024, not many chains have actually made the switch to post-quantum cryptography. So we have AI finding bugs that humans cannot find. And quantum computing on the verge of cracking the cryptography system that everything is based on. And most projects are taking no action on either. Your two ideas at the end are pretty much correct. AI-enhanced auditing must be an ongoing process, not a tick-the-box exercise. The old way of paying CertiK or anyone else one time and sleeping tight at night? That is gone. Zcash just proved that it is gone. And the less funded altcoins, the smaller team, the less eyes on the code? Yeah. I just do not want to think about what is written in those contracts these days.