Interview

AI agents and the new cybersecurity crisis

Roeland Delrue

Co-Founder, Aikido Security

41 minHosted by Kim Isenberg

Watch on YouTube

About this conversation

Kim Isenberg speaks with Roeland Delrue, co-founder of Aikido Security, about how AI-generated code is reshaping software security. As AI agents and coding assistants make development faster than ever, they also bring more complex codebases, new vulnerabilities, dependency issues and supply-chain attacks.

They discuss why software security may become one of the biggest bottlenecks of the AI era, what “self-securing software” could look like, and how autonomous security agents, continuous pen testing and auto-remediation may change cybersecurity.

Topics

  • How AI-generated code changes the security landscape
  • Why code is becoming more complex and bloated
  • Autonomous AI agents in software development
  • Continuous pen testing and automated remediation
  • Open source, supply-chain security and vulnerability management
  • The future of self-securing software

Chapters

  1. 0:00AI in software security
  2. 1:57Introduction to Aikido
  3. 3:29The AI coding bottleneck
  4. 6:33Company origins and mission
  5. 9:16Vulnerabilities in AI code
  6. 12:24Self-securing software
  7. 15:50Supply chain and CVE trends
  8. 17:50Practical security workflows
  9. 21:22Model selection and strategy
  10. 26:58Attack surfaces in AI tools
  11. 30:42Automation and human roles
  12. 33:30The future of security
  13. 38:24Democratization and conclusion

Kim Isenberg is co-founder and Editor-in-Chief of Superintelligence, an AI newsletter published every weekday.

Read the Newsletter

More interviews

All interviews