You bring up a good point. I do think interviewing for developers is not a field that has been studied using the scientific method, which is why we seem to have so many rules of thumbs, cargo cult traditions etc. rather than experiment-backed results. I think a few companies do study this as a science (e.g. Google) but most others are either unwilling or don't have enough resources (or some mix of the 2) to do the same.
> Interestingly the top companies seem to be focusing a lot less on experience and a lot more on problem solving/system design/etc.
As long as we're making conjectures: this is likely because a lot of them have tools and processes that are very different from what you see "in the wild". e.g. working at Google means you can leverage all their deployment infrastructure, something you likely didn't have access to outside (this is somewhat changing with k8s and cloud services).
And like science you can base it on the collective experience versus your personal experience. Preferably from rigorous studies but we mostly lack those in CS. So hopefully you're reading books, blog posts, studies, conference presentations, discussing with colleagues, etc.
>this is why in most serious technical interviews today, you will often get asked about your past experiences.
Interestingly the top companies seem to be focusing a lot less on experience and a lot more on problem solving/system design/etc.