Simple - a few thousand (if it is that many) people out of several million people having a problem is not always evidence of a real problem but rather proof that with a large enough install base some people will always have problems and it might well be for no reason at all beyond shear chance.
If there are 10 million installations of Mavericks (and I've got no idea how many there really are) and 5,000 people reporting a particular problem, that's only .05% of users - which is well within the expected proportion of users you'd expect to have problems with any upgrades. Note, for comparison the recent PS4 release problems and Sony's statement about the 0.4% of users experiencing problems - "This is within our expectations for a new product introduction"
Except you have no idea how many people are affected, you only know how many A) noticed, B) traced the culprit back to kernel_task, and C) reported it. That's likely a very small percentage of the afflicted population.
I'm also not sure how you explain a CPU pegged by a kernel process on a machine which should be doing almost nothing as "shear chance". That's a bug.
If there are 10 million installations of Mavericks (and I've got no idea how many there really are) and 5,000 people reporting a particular problem, that's only .05% of users - which is well within the expected proportion of users you'd expect to have problems with any upgrades. Note, for comparison the recent PS4 release problems and Sony's statement about the 0.4% of users experiencing problems - "This is within our expectations for a new product introduction"