None of My AI-Written Detections Were Right | Prompt to Prod - with Anthony Jirouschek
What if your blue-team AI didn't write detections at all, and just had to respond?
Anthony "AJ" Jirouschek builds autonomous detection engineering at Spectrum Security, after Air Force cyber and years of threat hunting and purple teaming at Capital One. We get into why AI seems to favor offense right now, the model asymmetry between attackers on unguardrailed models and defenders who can't type "attack" into theirs, and what changes when red AI fights blue AI in a lab. AJ tells the story of Gemini agreeing that Claude was right about his own detections, explains why every AI-written detection…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.