To write secure code, be less gullible than your AI

In this episode of The Stack Overflow Podcast, host Ryan Donovan interviews Greg Foster, CTO and co-founder at Graphite, about security breaches triggered by AI code. They discuss the shift in trust and the increase in code volume due to AI, which creates bottlenecks in code review. Greg suggests that smaller code changes and better tooling can help mitigate these issues. They also explore the gullibility of AI in the face of malicious prompts and the potential for LLMs to be used as security judges. Greg emphasizes that AI should be an additive layer to existing security practices, not a replacement, and that fundamental principles of clean code and architecture are more important than ever.