TheBlock
NEAR Protocol co-founder Illia Polosukhin warned that AI-assisted hacking has outpaced traditional manual code review, calling formal verification a necessary shift for securing blockchain infrastructure. Speaking to the broader security community, he framed the gap as a structural problem: defenders are auditing at human speed while attackers now operate at machine speed.
Why it matters
Polosukhin pointed to a specific acceleration in mathematical proof-generation work that used to take years and now completes in weeks when AI is applied. That asymmetry matters because formal verification is the methodology that produces machine-checkable proofs of correctness rather than relying on human reviewers catching edge cases. If proof generation speeds up by an order of magnitude, attackers get a faster pipeline for finding exploit paths, but defenders can also industrialise the audit process that today runs on small teams of specialised engineers.
Market impact
The framing positions formal verification, not bigger audit budgets, as the central security primitive for the next cycle of protocol design. For NEAR, a chain that has leaned heavily on account abstraction and sharding, the message is that the cost of a single missed vulnerability compounds faster when exploits can be generated and tested automatically. Projects still relying on conventional audit cycles face a widening exposure window.
Frequently asked questions
-
What did the NEAR co-founder say about AI and code security?
Illia Polosukhin argued that AI-assisted hacking has outpaced traditional manual code review and that formal verification is the necessary structural shift for securing blockchain code against AI-driven attacks.
-
What is formal verification and why does Polosukhin think it matters?
Formal verification produces machine-checkable mathematical proofs of code correctness, replacing reliance on human reviewers catching edge cases. Polosukhin framed it as the only methodology that scales to match machine-speed attackers.
-
How fast has AI made proof-generation work?
According to Polosukhin, mathematical proof-generation tasks that previously took years can now be completed in weeks when AI is applied, accelerating both defensive audits and offensive exploit discovery.
-
Why does this hit NEAR harder than simpler chains?
NEAR leans on account abstraction and sharded architecture, where a single missed vulnerability compounds faster when exploits can be generated and tested automatically rather than hand-crafted.
-
What should crypto projects take from this warning?
Projects still relying on conventional audit cycles face a widening exposure window as the timeline from vulnerability discovery to working exploit shortens. The argument is that formal verification must replace, not supplement, traditional review.
WuBlockchain
CoinTelegraph
CoinDesk
WatcherGuru
CryptoSlate