Is Vibe Coding Breaking the Internet? with Tanya Janca

The internet may be on the brink of a security crisis fueled by the unchecked rise of 'vibe coding'—where AI generates code with minimal human oversight. Ron Eddings, host of Hacker Valley Studio, opens with a stark warning: if we continue building software at lightning speed using AI without updating our security practices, the result will be catastrophic breaches, public backlash, and serious government fines. Drawing parallels to Margaret Hamilton’s meticulous, life-or-death software engineering during the Apollo missions, Ron contrasts today’s chaotic, AI-driven development with the disciplined, failure-anticipating mindset that once saved astronauts. Guest Tanya Janca, co-leader of the 2025 OWASP Top 10, reveals that even with security-focused prompts, AI can still introduce vulnerabilities—like a prompt that told a Python linter to ignore secrets and log them publicly. She argues that AI isn’t just a tool; it’s learning from the internet’s worst code, embedding OWASP Top 10 flaws into every app. The solution? A cultural shift: treat AI-generated code like a hammer—never close your eyes before swinging. Tanya’s free resources, including a secure coding prompt library and her YouTube series DevSecStation, aim to empower developers to review, verify, and secure AI output. The future of software depends not on speed alone, but on accountability, discipline, and the courage to ask: who owns the failure when the code breaks?