Season's Savings | ₹5/mo for 3 months

₹199/month after first 3 months. Renews automatically. Terms apply.
AI Can Find Every Unlocked Door. Who's Going to Lock Them? | Daria Bahrami, Dreadnode cover art

AI Can Find Every Unlocked Door. Who's Going to Lock Them? | Daria Bahrami, Dreadnode

AI Can Find Every Unlocked Door. Who's Going to Lock Them? | Daria Bahrami, Dreadnode

Listen for free

View show details
About this listen
 AI can now find the holes in our systems faster than anyone can fix them. So who's going to lock the doors? Is your organization ready for AI? Take the free 5-minute AI Readiness Assessment from ByteWise and get your score instantly: https://assess.bytewise-llc.com My guest is Daria Bahrami, Head of Policy at Dreadnode, the offensive AI security company that CrowdStrike agreed to acquire days after this conversation, reportedly for around $100 million. Daria is one of the sharpest minds working where AI meets cybersecurity. Before Dreadnode, she was a technical advisor to DARPA's AI Cyber Challenge, leading engagement strategy for the global competition to build AI systems that find and fix software vulnerabilities on their own. She has worked as a lead cyber threat intelligence analyst at Deloitte Global and a cyber program manager at the R Street Institute, and her writing and analysis have appeared in Fortune, War on the Rocks and CyberScoop. What makes her rare in this field is how she got here: from a Philz Coffee counter to National Geographic to the front lines of AI security, bringing a storyteller's instinct to one of the most technical debates of our time. Read Daria's Substack, Machine Speed: https://newsletter.techpointe.org We get into what DARPA's AI Cyber Challenge proved about machines finding and fixing vulnerabilities, why she thinks "rogue agent" is the wrong way to describe AI that escapes a sandbox, whether the best defense really is a good offense, and how a simple question to Gemini handed her a printer's admin password. We also cover water utilities still running on factory-default passwords, banks pulling back on AI agents, and the gap nobody is talking about: finding vulnerabilities is now easy, and nobody owns the patching. Recorded in late October 2026 at the Johns Hopkins Podcast Studio. Special thanks to Blair Johnson, Media Suite & Events Production Manager at Johns Hopkins, for his behind-the-scenes production support. Chapters 0:00 Why this conversation, why now 1:06 From journalism and Nat Geo to cybersecurity 7:30 What DARPA's AI Cyber Challenge proved 11:17 Inside an offensive AI security company 15:12 Why "rogue agent" is the wrong term 16:55 The sandbox escape: breach or rogue AI? 24:27 Open-source models and the race with China 29:45 Is the best defense a good offense? 32:58 How Gemini handed over a printer's admin password 35:33 Monopoly, chess, and the AI game we're playing 40:57 Water utilities and the basics we still miss 45:46 AI agents and your bank account 51:48 Are we waiting for a disaster to regulate? 58:11 Would you trust an agent with your money? 1:00:15 Black October: who patches the internet? 1:04:42 The one change that matters most 1:09:23 What scares her most, and the fix Watch on YouTube: https://www.youtube.com/watch?v=3Q2vkCRnIRs Free AI Readiness Assessment: https://assess.bytewise-llc.com All links: https://linktr.ee/mohafez Vote for Mo for Top Host: https://tophost.org/2026/mo-hafez Stay curious, stay ethical, and lead human.
adbl_web_anon_alc_button_suppression_t1
No reviews yet