For the last several months, we have had early, unbounded access to the latest frontier AI models. What we’ve seen from that vantage point has made it clear that the window for organizations to get ahead of what’s coming is shorter than most leaders realize. We have moved past the era of incremental AI improvements into a threat landscape shift. Our testing has revealed a step-change in capability that demonstrates an intuitive understanding of software vulnerabilities.