#28: Security Taco

AI agents are becoming more capable, persistent, and creative, and that’s forcing businesses to reconsider some fundamental assumptions about cybersecurity. Dustin and Andy discuss an AI security test that escaped its intended sandbox, why AI could dramatically lower the barrier to discovering vulnerabilities, and what concepts like zero trust, least privilege, sandboxing, and active monitoring mean in an AI-powered world.

They also explore the risks of giving AI tools access to business data, how seemingly private shared conversations can become exposed, and why adopting AI responsibly doesn’t mean avoiding it altogether. The goal is understanding your risk, putting practical guardrails in place, and making sure your business is prepared when something inevitably goes wrong.

🔗 Resources and mentions from this episode:

• OpenAI
• Hugging Face
• Anthropic
• Claude Code
• xAI and Grok
• Zero trust security
• Least privilege
• AI sandboxing
• The Adolescence of P-1 by Thomas J. Ryan
• Cyberpunk
• The Angry Clean Energy Guy
• AI and renewable energy

👉 Want more practical insights on using AI in your business? Subscribe to the podcast, follow us on YouTube, or explore past episodes at modernlogic.io/podcast.

💬 Let’s Talk AI!

Using AI in your business and wondering if you’re doing it securely? Modern Logic can help you think through your AI strategy, automation, software, and security considerations. Book a free 30-minute consultation at molo.now.