Ethereum Founder Vitalik Refutes the View that "AI Hacking Will End Cybersecurity"
Vitalik Buterin posted on X platform to refute the view that "AI hacking will end cybersecurity," clearly stating his disagreement with this judgment and asserting that cybersecurity is essentially a mechanism favorable to the defenders.
He specifically mentioned the binding of interests to his own position: including himself, anyone currently holding cryptocurrency is implicitly betting on the judgment of "defenders being favored," and since about 90% of his net worth is held in cryptocurrency, this bet carries significant weight for him, directly relating to the safety boundary of his personal wealth.
Vitalik pointed out the specific direction of AI technology: AI can transform "program security" into verifiable mathematical theorems, meaning security no longer relies on manual audits or empirical judgments, but can achieve deterministic conclusions through formal proofs, which fundamentally differs from the traditional passive defense logic of "patching and plugging vulnerabilities."
However, he also delineated the boundaries of this method: the definition of security must cover multiple risk factors including key management, servers, end devices, operating systems, databases, network layers, and cache layers; merely verifying the code logic itself is insufficient, and true security requires establishing equally strong protections for every link in the entire chain.
For security-critical components, Vitalik proposed the solution of "holistic verification"—that is, no longer conducting local reviews of single modules or functions, but performing end-to-end formal verification of the entire program, which means that the granularity and workload of verification will significantly increase.
Vitalik stated that Ethereum's technological roadmap for the coming years will advance in this direction, emphasizing that while blockchain must achieve both "scalability" and "privacy," software security must be a fundamental prerequisite; otherwise, scaling and privacy themselves will amplify the attack surface.
From a market mechanism perspective, such statements are more narrative signals rather than direct capital movements: they target users who firmly hold ETH and Ethereum ecosystem assets long-term, essentially providing counter-narrative support against the potential sell-off sentiment triggered by the "AI threat theory"; if there were previously funds that hesitated or reduced their holdings in Ethereum assets due to AI security panic, such statements may alleviate some selling pressure, while funds hedging against AI risk narratives may come under pressure.
Source: Public Information
ABAB AI Insight
Ethereum has not been without costs for security in its history: in 2016, the DAO smart contract vulnerability led to the theft of about $60 million in assets, and the Ethereum community ultimately chose a hard fork to recover the funds, resulting in the split into Ethereum Classic (ETC), which is recognized as one of the industry's most severe smart contract security incidents; in 2017, the Parity multi-signature wallet was frozen due to a code vulnerability, affecting over 300,000 ETH, further exposing the destructive power of security flaws under the "code is law" model. These historical incidents provide the real-world context for Vitalik's emphasis on "security must cover the entire chain."
On the capital path, the Ethereum Foundation has been continuously directing resources towards formal verification and zero-knowledge proof directions in recent years: its internal Privacy and Scaling Explorations (PSE) team specifically studies the application of zero-knowledge proof technology in privacy and scalability, while the Ethereum ecosystem has long funded projects like KEVM (formal modeling of the Ethereum Virtual Machine based on the K framework), aiming to shift "security" from relying on manual audits to being a mathematically verifiable problem, which aligns perfectly with Vitalik's statement that "AI will transform program security into mathematical theorems."
In industry comparisons, formal verification is not a blockchain innovation but is derived from traditional engineering fields with high security requirements: for example, the seL4 microkernel is the world's first operating system kernel proven "bug-free" through complete mathematical proof, widely used in aviation, military, and other high-reliability scenarios; the blockchain industry is currently in the expansion phase of migrating this "critical infrastructure-level" verification methodology from small-scale academic practice to large-scale production systems, rather than a mature application phase.
From a structural perspective, this is essentially a "technological substitution"—AI-driven formal verification is replacing the traditional manual security audit model. Traditional security audits rely on the accumulated experience of auditing firms and sampling checks, which have inherent coverage blind spots and cost ceilings; as AI gains the ability to automatically generate mathematical proofs at scale, verification costs decrease with scale while coverage can approach complete chain links, transforming "verification" from a scarce expert service into a scalable infrastructure capability, thereby re-pricing the cost structure of security games.
ABAB News · Cognitive Laws
The moat of defense is to keep the cost of attacks always higher than the benefits.
Trust that can be mathematically proven will not be negotiated.
The more transparent the system, the less room there is for vulnerabilities to hide.