From the description:

In episode 5 we sat down with Alan Zabihi Co-Founder of Superagent.

Superagent secures your code and agents with autonomous security researchers that find and patch vulnerabilities.

A Super exciting opensource product which you can try out here - superagent.sh

Chapters:

0:00 Introduction and Background

12:54 The Shift to AI Security:

23:41 AI Attacks and Vulnerabilities

33:04 The Failure of Guardrails

39:05 The Attack Surface of Autonomous Agents

47:35 Data Poisoning and The Agentic Web

57:05 The Lethal Trifecta and Risk Management

1:06:23 Introducing Brin and Future Directions

The Death of Guardrails: Why Your Enterprise AI Isn't Safe Yet