Pattern Labs to Present Two Talks at BlueHat 2025

We're excited to announce our participation at BlueHat security conference this week. Join Pattern Labs as we present two talks:

  • Yoni Rozenshein will present Hey AI, how many "r" in "buffer overflow"? This talk will discuss how AI models can excel at certain security tasks while surprisingly failing at seemingly simple tasks. Drawing from Pattern Labs' work with frontier AI labs, Yoni will share examples of these unexpected limitations in vulnerability discovery and exploit development.
  • Omer Nevo will present Hack like a robot: Journey into the logic of LLM-based vulnerability hunters. This session will explore what happens when LLMs tackle vulnerability discovery, examining their successes, spectacular failures, and surprising insights. Based on Pattern Labs' evaluations of frontier AI models, Omer will reveal how these systems "think" about finding and exploiting software bugs and how their capabilities have evolved over the past two years.

We will share these insightful presentations when they become public.

To cite this article, please credit Pattern Labs with a link to this page, or click to view the BibTeX citation.
@misc{pl-bluehat2025,
  title={Pattern Labs to Present Two Talks at BlueHat 2025},
  author={Pattern Labs},
  year={2025},
  howpublished={\url{https://patternlabs.co/blog/bluehat-security-conference-2025}},
}
← Back to Blog Feed
Pattern Labs to Present Two Talks at BlueHat 2025