Quick Take (TL;DR)

LLMs don’t think—they predict. Keith Hoodlet shows what this means for CISOs facing bias, slopsquatting, MCP risks, and burnout.

Guest Spotlight

Keith Hoodlet is Engineering Director at Trail of Bits. He previously led at GitHub and Rapid7, co-founded Application Security Weekly, and launched the InfoSec Mentors Project.

LinkedIn | Website | Newsletter

Resources & References

Books

AI Snake Oil
Four Thousand WeeksArticles / Studies
Marine Corps Times
2025 Cloud‐Native Security and Usage Report
The Register: SlopsquattingTools / Frameworks
Model Context Protocol
NVIDIA NeMo Guardrails
Meta Llama Guard

Call to Action

If this episode reshaped how you think about AI security, share it. Connect with your hosts:

Conor Sherman — LinkedIn | Website | Sysdig;

Stuart Mitchell — LinkedIn | Website.

Subscribe to Zero Signal: Apple | Spotify | YouTube | Website

Challenging Trust in AI Systems ft Keith Hoodlet