AI-Assisted Hacking Outpaces Traditional Code Review, NEAR Co-Founder Advocates Formal Verification

Odaily Odaily reports that Illia Polosukhin, co-founder of NEAR Protocol, stated that AI is dramatically accelerating hackers' ability to discover software vulnerabilities, making traditional code review models inadequate. He pointed out that the Balancer incident in October last year marked the beginning of AI-assisted hacking, and the old model reliant on manual review is no longer sustainable in the AI era. Polosukhin advocates for formal verification, using mathematical proofs to ensure that code behavior fully aligns with specifications. AI can generate most of these proofs, which are then verified by humans for correctness, significantly reducing proof generation time. He believes formal verification should extend beyond smart contracts to cover critical internet and network software. On the same day, NEAR released an upgrade, allowing users to switch to quantum-resistant account keys, and the network can automatically scale capacity as activity grows.