VITALIK BUTERIN: AI COULD PUSH BLOCKCHAIN SECURITY TOWARD A DIFFERENT APPROACH
Vitalik Buterin rejected the view that AI-powered hackers will cause cybersecurity to collapse. According to PANews, he argues that the focus should be building systems that resist attacks and can be verified, rather than racing to find vulnerabilities.
Buterin compared it with mathematics: if AI can prove Navier-Stokes and Fermat’s Last Theorem, then in principle it could also prove a program secure like a theorem.
He noted that “security” covers forgery, replay attacks, server compromise, key leakage and side-channel attacks. This scope remains smaller than the attack surface of source code and may be easier to verify.
Buterin said earlier efforts to verify only parts of security systems had failed. With AI, the entire program would need verification, including databases, networks, caches and other components.
He does not frame this simply as a race between good and bad actors. The goal is more attack-resilient code, and Buterin said this will be part of Ethereum’s development.
This view matters because many blockchain software and infrastructure layers interact. Buterin said that without strong verification, blockchains combining scalability and privacy will struggle long term.
Buterin also said he continues to hold crypto, around 90% of his net worth, as a bet on improving security through design and verification.
AI is increasing offensive capabilities, but Buterin’s argument shifts the question from “who finds the bug first?” to “how well can a system prove that it is secure?”
Could full-program verification become a new security standard for Ethereum and blockchains combining scalability with privacy?
Please do your own research carefully before making any transactions (DYOR). $ETH $SOL $ZEC
Vitalik Buterin rejected the view that AI-powered hackers will cause cybersecurity to collapse. According to PANews, he argues that the focus should be building systems that resist attacks and can be verified, rather than racing to find vulnerabilities.
Buterin compared it with mathematics: if AI can prove Navier-Stokes and Fermat’s Last Theorem, then in principle it could also prove a program secure like a theorem.
He noted that “security” covers forgery, replay attacks, server compromise, key leakage and side-channel attacks. This scope remains smaller than the attack surface of source code and may be easier to verify.
Buterin said earlier efforts to verify only parts of security systems had failed. With AI, the entire program would need verification, including databases, networks, caches and other components.
He does not frame this simply as a race between good and bad actors. The goal is more attack-resilient code, and Buterin said this will be part of Ethereum’s development.
This view matters because many blockchain software and infrastructure layers interact. Buterin said that without strong verification, blockchains combining scalability and privacy will struggle long term.
Buterin also said he continues to hold crypto, around 90% of his net worth, as a bet on improving security through design and verification.
AI is increasing offensive capabilities, but Buterin’s argument shifts the question from “who finds the bug first?” to “how well can a system prove that it is secure?”
Could full-program verification become a new security standard for Ethereum and blockchains combining scalability with privacy?
Please do your own research carefully before making any transactions (DYOR). $ETH $SOL $ZEC
