Post by SentinelOne

In the final video from our LABScon 2025 Replay series, Juan Andres Guerrero-Saade (JAGS) argues that the experimental era of cybersecurity is ending. Years of piling complexity onto non-standardized software stacks have produced systems that have left security unsteerable and costly to human-only management. What changed the entire equation is the rise of large language models. JAGS describes them as a new source of cheap, effectively unlimited evaluative power, a "lossy compression of human knowledge." Used well, that kind of mechanized intelligence gives defenders a scalable way to assess, prioritize, and act. It also lowers the cost of analysis and changes how defensive work can be done at scale. This argument shapes JAG-S' broader point about how security should evolve. Drawing on cybernetics, he urges the industry to move beyond purely adversarial, agonistic design and toward systems where human expertise and artificial evaluative power work together to produce better outcomes. The blueprint is to build in, not bolt on. Rather than defending old product categories or familiar workflows, it’s time for a more standardized, automated, and sustainable future. Watch the keynote: https://s1.ai/LC25-JAGS

Video Content