I am completely disillusioned with prestige at this point in my academic career. It is all about the strength of the peer review committee for me now, nothing less.
I work on the intersection between engineering and computer science, so I get to read literature from both community and I cannot tell you how many completely **** paper makes it through the "peer review" process at top ML conferences/journals such as ICML, JMRL, ICLR, AAAI and NeurIPS. These "prestigious" journals are publishing crank-level papers with extraordinarily poor citation practices.
Of course, if you look at their peer review process then it is all revealed: two-three reviewers (possibly students) not too familiar with the field or the literature giving superficial and uncritical reviews to not hurt any feelings. There is an over-emphasis on simulation results, which are barely reproducible. Finally these reviews are rarely completely blind as well, because a paper is almost always simultaneously uploaded to Arxiv. Just because they are Google employees doesn't make their research good.
The same probably can be said for a minority of journals in engineering. I have definitely seen a share of bad papers here but I always ensure that they do not make it through.
It is all about the peer review process.