The recent discovery of 21 zero-days in FFmpeg by an AI agent and Google's patch of 429 bugs in Chrome has sparked discussions about the role of AI in cybersecurity. While AI has proven to be a powerful tool in identifying vulnerabilities, the challenge lies in the subsequent steps of triaging, fixing, and deploying these patches. The cost of finding these bugs has decreased, but the cost of addressing them remains high, with much of the work still relying on human volunteers and a limited layer of human triagers. This highlights the need for a comprehensive approach to cybersecurity that includes both AI-driven detection and human expertise in the critical stages of patch management and deployment.
