If we forget about politics for a minute and think about the _practical_ side of scientific publishing, we might find that the two are linked.
Imagine how much easier it would be to verify scientific claims if those claims were made in terms of structured statements (which could also be parsed by machine learning algorithms).
@rupertoverall that might actually make it easier to game the system. Although the paper goes some way to provide these structured statements, peer review is more robust as a check, the best would be AI-assisted peer review, for example to propose feedback, counterpoints and protocols
@rupertoverall very true 👌🏻
@rupertoverall hey just using this one after a few days, and thought of you : https://www.explainpaper.com/ hope you like it :-)
@tonic My idea here was not to replace peer review, but to make the content of the reports more accessible. Not only for machine-readability, but also for re-flowing text (Kindle, web/mobile, variable font size, dictation) and to enable mark-up of semantic concepts (molecules, species, reagents, drugs) that would reduce ambiguity immensely.
Human peer reviewers could certainly use this info to assist them (and no energy-hungry AI would necessarily be needed).