IRL Location: Knuckleheads Garage - 701 North Montgall Avenue, Kansas City, MO 64120 United States
Remote Location: NOPE! This month will NOT be streamed! If you aren’t at Knuckleheads, you’ll miss it. And you won’t want to do that!
If you haven’t been to Knuckleheads or want to make sure you’re not stuck on the wrong side of the train, check out the venue info page!
5:00 - 6:00 Make excuses / pre-drinking
6:00 - 6:10 SecKC 101 / announcements / DIET COKE AUCTION / Welcome to SecKC - @TBD
6:15 - 6:25 Open Source News! - @TBD
6:35 - 7:00 Against the Clock: A story of winning a DEFCON Black Badge - @_seash & @AntlerSec
Brayson and Chris walked in the SECVC booth for the first time ever this year at DEFCON with hopes of winning it all and getting that coveted black badge. Little did they know just how much work, stress, and luck it would take to get to that point.
During this talk, they'll discuss everything from team application to walking across the stage and all the chaos that happened in between. Think you can step in to the booth because it look easy? Come listen to their experience and you might just change your mind!
7:10 - 8:10 noRecognition - @hevnsnt
Facial recognition evasion research has a costume problem. Masks, infrared LEDs, real-time face swaps, adversarial makeup. Every approach either requires active electronics, makes you look like a Batman villain, or replaces your face with someone else's. None of them scale. None of them are subtle. And none of them actually attack the model.
noRecognition takes a different approach. I built a genetic algorithm that breeds adversarial textile patterns, printed on ordinary fabric, that cause cascading failures across the full facial recognition pipeline. No batteries, no software, no one staring at you on the subway. People nearby see a person wearing a scarf. The AI sees nothing.
The platform tests evolved patterns against a gauntlet of 10 models mapping directly to real-world deployments: the same YOLOv8, RetinaFace, and ArcFace architectures running inside Clearview AI, Axon body cameras, Palantir systems, and commercial surveillance infrastructure. Using a library of 61+ attack techniques (gradient-based adversarial ML, surgical landmark targeting, frequency-domain disruption) and a distributed computing network contributing GPU cycles worldwide, I have tested and evolved patterns that simultaneously defeat person detection, face detection, and identity recognition.
This talk is the SecKC insider edition, building on what was shown at BlackHat with new, unreleased research!
8:15 - ??? Mad Beatz by the amazing @j0nny54l1v3
Sponsors
Malmoset & Shelly (GOLD)
Horizon3.ai (Silver)
Cohesity (Silver)