Blackhat Rat Pack Recap

There’s too much happening at Blackhat, so in the spirit of cutting through the noise suffice it to say the Empirical team had some great discussions and came away with three big learnings:

Not everything should be automated. Security is full of edge cases where human judgment is still essential, especially when the stakes are high. Automation is powerful, but knowing when not to use it is just as important as knowing how.

AI is everywhere, but initial attempts still require a ton of handholding. We saw plenty of impressive demos from folks in informal sessions, but most of them still require experts to steer the outputs and catch mistakes. The promise is real, but the “set it and forget it” moment isn’t here yet.

Ground truth matters more than ever — models are only as good as the data they’re built on. Without reliable, real-world telemetry, even the most sophisticated models will drift or mislead. The conversations that stuck with us were the ones centered on measuring reality.

More importantly, we spent some quality time with some quality people.