I think you’re missing the point, which was that high value targets adjacent to soft targets make escapes a legitimate target, but in low value scenarios vm escapes aren’t worth the R&D

but if you can do it at scale it might still be worth it, like owning thousands of machines