The Hidden Cost of BlackBox AI: Bridging Cloud and Code Security
Show notes
Traditional SCA and SAST tools are notorious for drowning security teams in false positives, historically flagging nine out of ten alerts incorrectly. But while generative AI seems like a magic bullet, simply wrapping an out-of-the-box LLM around your code can result in confident hallucinations and astronomical costs extrapolating to as much as $52 million a year for a large enterprise using frontier models.
In this episode, Ashish sits down with Harry Wetherald, CEO and co-founder of Maze, to discuss the evolution of AI-native AppSec and Cloud Security. Harry breaks down the critical difference between vulnerability reachability (is the code active?) and true exploitability (can an attacker actually trigger it logically?). He also explains why the historical walls between cloud security and application security teams are finally crumbling as AI acts as a perfect translator between the two domains.
If your team is debating "Build vs. Buy" for AI security tools, this episode is essential. Harry shares the biggest red flags to watch out for in AI vendors (beware the "black box"), how to intelligently route across models to optimize token costs by 100x, and how a true "security brain" orchestrates multiple investigations to provide reliable context across your entire environment.
Guest Socials - Harry's Linkedin
Podcast Twitter - @CloudSecPod
If you want to watch videos of this LIVE STREAMED episode and past episodes - Check out our other Cloud Security Social Channels:
If you are interested in AI Security, you can check out our sister podcast - AI Security Podcast
Questions asked:
(00:00) Introduction to AI in AppSec(01:50) Harry Wetherald's Background and the Founding of Maze(02:30) Reachability vs. Exploitability Explained(04:45) The "Build vs. Buy" Dilemma for AI Security Tools(08:30) Bridging the Gap Between Siloed AppSec and CloudSec Teams(11:30) Evaluating Out-of-the-Box LLMs vs. Specialized Security Tools(14:20) Solving the Historic AppSec False Positive Problem(18:50) AI Vendor Red Flags: The Danger of "Black Box" Products(20:50) How to Build a True AI-Native Security Architecture(24:45) The Hidden Cost of AI Models: Why Optimization is Crucial(28:00) When to Keep a Human in the Loop for Remediation(34:00) Building a "Security Brain" to Inform AI Coding Agents(39:20) The Launch of Maze Code for Deep Cloud and Code Investigations