There are alternatives to traditional journals which you may not have factored into your analysis. Search for the Public Library of Science (PLoS) and you will find professionally peer reviewed papers published on a publicly accessible (no paywall) site online.

There remain additional reasons that traditional journals, and even peer review, inhibit the development and publishing of good science and papers, specifically mechanisms employed to prevent demonetization of extant grantees by upstarts, political machination, and bias. Algorithms will combat the latter, but not the former faults with extant publishing methodologies. I'd like to see more concern given integrity in science, as I have experienced some of these impediments in my work, and feel they have prevented far more than my own meager contributions from being available to the world.

Good science gets crushed sometimes by people whose incomes depend on the status quo, and this is bad for society. A side effect is also that 'mad scientists' and frauds are able to undertake work contrary to the public interest in the non-traditional publishing mechanisms that must eventuate due to the failure of extant mechanisms to appropriately and functionally publish. Some people are presently alarmed, and even with some good reason, at the biohacking movement, makers, and other somewhat renegade publishing that has arisen as a result of censorship, gatekeeping, and other failures of scientific publishing to function nominally.

Thanks!