soon, you will find out that *all software* handling funds is prone to AI finding vulnerabilities. all of it
that's why you also saw tons of defi exploits already this year
open source, L1, L2, DeFi, mutable, immutable, private, transparent, cross-chain. if it's software, it is at risk. in fact, it has always been at risk, but now AI changes the math.
so what does one do about this? is the idea of crypto over?
in theory, the main difference for privacy contracts vs regular contracts is detectability
in practice, we've seen countless times in defi where the hackers got away fast even after detection (happened at least 7 times this year that I can count)
it is not an accident that you've seen so many in recent times
the safest way against this is self-custodying the native asset itself on the native chain. because even in the worst case, the validators could roll it back if the exploit is large and core enough. so there is some lower bound there
but as vitalik, toly, and many others have pointed out, the main improvement to defend against this going forward will be formal verification
this is a rigorous, mathematically-based method to formally prove that software behaves exactly as intended
this is also why open source will be critical. because you will have countless others also trying to help you collectively improve security (for rewards) whereas with closed source the math is fundamentally skewed because only the core team can defend
the next zcash upgrade, some defi protocols, and some chains are already in the process of doing this, aggressively
the result will be that crypto emerges stronger than ever before and in fact safer than centralized counterparts. there will be no stopping crypto.
seems hyperliquid:native has flipped solana:So11111111111111111111111111111111111111112 in price
kudos to them. well deserved.
now, we commence the Solana comeback, 3rd edition