Thrilling Threads - Conspiracy Theories, Strange Phenomena, Unsolved Mysteries, etc!
When AI Goes ROGUE: Stopping Catastrophic Failures in Agentic Systems
January 23, 2026
AI Summary
5 min read🎙️ The Voices & The Context
- The Format: Casual chat between two hosts dissecting a YouTube keynote talk, blending narration, banter, and technical breakdowns for a podcast audience.
- The Key Players:
- Hosts of Thrilling Threads (unnamed duo with sharp chemistry—playful back-and-forth, one sets up scenarios, the other amps up the drama).
- Guest via video: Dr. Mike Pound, computer scientist and cybersecurity heavyweight, fresh from his alarm-ringing InfoSecurity Europe keynote titled "Cybersecurity 2026 Warning: AI Makes Every System Riskier".
- The Vibe: Educational with dark humor and urgency—funny horror stories mixed with "wake-up call" intensity, like a thriller podcast pulling at AI's fragile threads.
🗝️ Key Themes & Topics
The episode unpacks AI's cybersecurity pitfalls through Dr. Pound's lens, contrasting hype with harsh realities. Main topics: the core tech clash fueling risks, sneaky attack methods, hidden supply chain dangers, and risky "agentic" AI plus fixes.
Continue reading the full summary in the app — free to try.
Read Full Summary →Free • No credit card required
What you'll learn
- 1 (01:30) **🎙️ Introduction: Dr. Mike Pound**
- 2 (05:12) **Deterministic vs. Probabilistic Computing**
- 3 (08:59) **The 99.9% Trap**
- 4 (10:33) **Culture Clash: AI Devs vs. Security Pros**
- 5 (11:50) **Prompt Injection Attacks**
- 6 (15:16) **Indirect Prompt Injection**
- 7 (19:01) **AI Supply Chain Risks**
+ Full timestamped outline available in the app
Show Notes
Is your AI assistant actually a double agent waiting to leak your bank details?
We are reacting to a terrifyingly eye-opening discussion between cybersecurity legends David Bombal and Dr. Mike Pound, and let’s just say: the future of Agentic AI is basically the Wild West.
We’ve all played with ChatGPT, but what happens when we give AI "agents" the power to read our emails, access our calendars, and execute code? In this episode, we break down the massive security gap that the industry is ignoring. We explore the concept of Non-Deterministic AI—which is a fancy way of saying "we don't know what it’s going to do next."
We dissect the mechanics of Prompt Injection (the new SQL injection) and how hackers can trick an LLM into bypassing its own safety filters to steal sensitive medical or financial data. It’s not just about a chatbot saying a bad word anymore; it’s about autonomous agents having the keys to the castle while developers treat security as an afterthought.
From Defense-in-Depth strategies to the absolute necessity of "human-in-the-loop" verification, we cover what you need to know to survive the next wave of cybersecurity threats.
Are we building tools to help us, or weapons to hack us?
Tune in as we bridge the gap between AI development and digital defense.
👉 Found this useful? Tap FOLLOW or SUBSCRIBE to stay ahead of the hackers. Share this episode with your IT friends or anyone who trusts AI a little too much—it might just save their data.
#AgenticAI #Cybersecurity #DavidBombal #PromptInjection #LLMSecurity #AIHacking #TechNews #Privacy #FutureOfTech #Podcast
Become a supporter of this podcast: https://www.spreaker.com/podcast/thrilling-threads-conspiracy-theories-strange-phenomena-unsolved-mysteries-etc--5995429/support.
You May also Like:
🤖Nudgrr.com (🗣'nudger") - Your AI Sidekick for Getting Sh*t Done
Nudgrr breaks down your biggest goals into tiny, doable steps — then nudges you to actually do them.
We are reacting to a terrifyingly eye-opening discussion between cybersecurity legends David Bombal and Dr. Mike Pound, and let’s just say: the future of Agentic AI is basically the Wild West.
We’ve all played with ChatGPT, but what happens when we give AI "agents" the power to read our emails, access our calendars, and execute code? In this episode, we break down the massive security gap that the industry is ignoring. We explore the concept of Non-Deterministic AI—which is a fancy way of saying "we don't know what it’s going to do next."
We dissect the mechanics of Prompt Injection (the new SQL injection) and how hackers can trick an LLM into bypassing its own safety filters to steal sensitive medical or financial data. It’s not just about a chatbot saying a bad word anymore; it’s about autonomous agents having the keys to the castle while developers treat security as an afterthought.
From Defense-in-Depth strategies to the absolute necessity of "human-in-the-loop" verification, we cover what you need to know to survive the next wave of cybersecurity threats.
Are we building tools to help us, or weapons to hack us?
Tune in as we bridge the gap between AI development and digital defense.
👉 Found this useful? Tap FOLLOW or SUBSCRIBE to stay ahead of the hackers. Share this episode with your IT friends or anyone who trusts AI a little too much—it might just save their data.
#AgenticAI #Cybersecurity #DavidBombal #PromptInjection #LLMSecurity #AIHacking #TechNews #Privacy #FutureOfTech #Podcast
Become a supporter of this podcast: https://www.spreaker.com/podcast/thrilling-threads-conspiracy-theories-strange-phenomena-unsolved-mysteries-etc--5995429/support.
You May also Like:
🤖Nudgrr.com (🗣'nudger") - Your AI Sidekick for Getting Sh*t Done
Nudgrr breaks down your biggest goals into tiny, doable steps — then nudges you to actually do them.
More from this podcast
Thrilling Threads - Conspiracy Theories, Strange Phenomena, Unsolved Mysteries, etc! →