As an researcher I typically only encounter a paywall when I‘m not logged into my institutes VPN. In the rare cases I want to read something the university is not subscribed to, Scihub is indeed a good solution. Open access just moves the cost of publishing to research groups and disincentives proper peer review, because the journal’s revernue is tied to how many paper it publishes. See for example the 15+ Frontiers in ... Journals, which are just barely not predatory and basically publish anything with a pulse. A mix of preprint severs like arxiv and (much later) journal publication seem like a good compromise.
Journals don't incentivize proper peer review, open access or otherwise. Peer review is just something that professors and their assistants are expected to do as part of their research. Which means it's usually paid for by the public, ultimately.
(I have never heard anyone in research who thinks that pay walled journals is a good thing.)