SentinelOne cuts malware analysis from weeks to hours

SentinelOne's chief AI officer Gregor Stewart says OpenAI Daybreak can produce human-level reverse-engineering output in roughly eight hours, work that previously took a top researcher four to six weeks. Faster analysis lets the team investigate more samples and look for shared components and relationships across threats.

Stewart stresses that accountability stays with humans, especially for architecture-changing decisions, while point fixes may be better suited for automation. He also says Daybreak is letting more people inside SentinelOne take on security work that was once limited to a small group of specialists.

We'll take the hours-versus-weeks claim as company testimony and still say the direction is important. Malware reverse engineering is a scarce skill, and attackers don't wait. Compressing that work is a real defender advantage if the output holds up under expert review.

Stewart's line that accountability can't be automated is exactly right. The danger with "human-level" AI analysis is that teams stop reading it carefully because it looks finished. Point fixes yes; silent architecture changes no.

There's also a dual-use shadow around tools that reverse-engineer software this well. Daybreak is framed for defenders, but the same capabilities cut both ways. Access controls and logging matter.

I'm watching for peer-reviewed or third-party validation of Daybreak's reverse-engineering quality, not just vendor quotes.