Relying on AI to find vulnerabilities first is not a sustainable security strategy, warns Bitcoin Core developer Niklas Gögge

Odaily reports: Bitcoin News posted on X platform that Bitcoin Core developer Niklas Gögge warned that recent AI-driven vulnerability scanning is changing the Bitcoin security landscape. Large language models have significantly reduced the cost of vulnerability discovery, and attackers may be able to find catastrophic vulnerabilities with only a few hundred dollars in computing costs. For Bitcoin Core, Project Loupe, Bitcoin Red Team, and individual contributors have generated over 1,000 reports, but so far no high-risk or critical vulnerabilities have been found. Gögge stated that relying on stronger models to find vulnerabilities before attackers do is not a sustainable security strategy. Developers should build testing infrastructure through automated testing, fuzzing, and property-based testing that can prevent entire classes of vulnerabilities in advance. Key components of Bitcoin Core have cumulatively completed over 100 years of CPU fuzzing and decades of Bitcoin node network simulation.