News
Crypto industry tackles AI-driven bug hunting
Crypto security teams are adapting to AI tools that can rapidly discover vulnerabilities. Bitcoin Core implemented a policy to quickly patch bugs detectable by large language models. Starknet voted to explore an independent post‑quantum L1 for faster cryptography upgrades, while Ethereum promotes minimalistic, hash‑only code to reduce attack surface. Formal verification advances, such as Anthropic’s Claude proving thousands of library entry points, illustrate costly but promising defenses. Recent incidents include a Day‑0 bug exploit costing $2.7 M and a rise in $1 M+ hacks to 112 in 2026, highlighting the accelerating threat landscape.