What's this about?
Box CEO Aaron Levie: models are becoming insanely good at finding and exploiting vulnerabilities, frontier models are ahead but open weights are not far behind, and most enterprises are already inundated with cyber discoveries — so the volume is about to multiply. His conclusion: triaging and automating fixes with more AI, alongside human oversight, is essentially the only way forward. And a note on jobs: it is going to be a great time to be in security.
Why does more discovery create a problem rather than solve one?
Because a vulnerability report is not a fix, it is work. Security teams are already the bottleneck: every finding needs triage, reproduction, severity assessment, prioritisation, a patch, testing and deployment. Multiply the input by ten and you have not made anyone safer — you have made the queue longer, and a longer queue means the important thing sits behind more noise. This is the same shape as AISLE's 23 unaccepted reports, at organisational scale.
Why does this matter to you?
- It is the demand-side argument for exactly the kind of tooling that turns raw findings into triaged, prioritised, actionable output.
- "Open weights are not far behind" is the part that determines the timeline, and it is consistent with what GLM-5.3 was marketed on last week.
- Coming from an enterprise software CEO rather than a security vendor makes it a statement about buying behaviour.