Considering the state of "peer review" in some journals and what is being published, even doing that is a waste of time for some. Yes, smart or careful people willl do that (of course if Crossref data keeps on being accurate...)

Academics love to say "most famously" and then describe a fact that 9 people in the world are aware of

It would be worse for two reasons: 1/ even easier to perform than it is now 2/ it will give the illusion (like many LLM answers) that everythihng is fine, so more practiced. See the first retraction for asking an LLM to make the references section "à la Zotero" and then having "fake references".

I agreed on that point, mine is that it will be mostly the "give me the review paper that is obvious in a minor point of my review literature in an IMRAD format". OK if the diversity remains for limited Matthew effects, though gaming (and farmers) will be paramount.

The GS Matthew effect will be multiplied in AI search because most people won't consider 10 papers but only 5 or even 3 or even 1. "AI search" will be the "Wikipedia" of search : give me THE ONE answer to my question.

Probably, with much bigger problems: quoting non-existing papers (because they will use "AI search" not to read them), putting "hallucination" references in their bibliography and much more massive gaming on the other side to get the first place in "AI SEO".

I thought the cost of WoS and Scopus for CNRS was crazy high (around 1M and 500K/year), but considering the size of users, US universities pay even higher prizes. I hope soon all these institutions will turn to @matthieu.bsky.team for bibliographic tools and open alex for bibliometric tools.

This might be of interest to you tidsskrift.dk/njlis/articl...zenodo.org/records/1090...

Of course, "publishers" not as service providers but as companies that sell their owned journals are still thriving, despite boycotts and petitions.

At least it reminds us that the unified service provider named "publisher" is a historical exception that expanded during the first decade of the electronic age. I always compare that to the 1980-1990's major music industry. Nowadays artists use so many service providers (I agree on the energy pat)