The approach taken by the machine learning community succeeded and incurred absolutely none of sky-is-falling shyte you moan about.
I make it very clear that academic publishing has issues. I state it plain as day, I even literally finish my comment with “And to be very clear…:
> And to be very clear, there are very serious problems in the academic publishing world, no question.
However, changing something without even an understanding of what could go wrong is a very serious mistake. As I said in my comment, this doesn’t mean we shouldn’t seek to change and improve things. But if one can’t even imagine what could go wrong, than this person either hasn’t though much about it, or they clearly do not understand the subject very well.
If something needs to be fixed, don’t fall into amature hour. Take the time to actually understand the subject first. Then start fixing.
Now don’t pull bits and pieces of what I said and separate them, please. It’s only a couple paragraphs, It’s not difficult to hold onto the entire context of a couple of small comment paragraphs.
I'm not sure how you can look at political discourse in the United States today and say "yep, let's just blow up academic gatekeepers". Just one anti-vaccine paper from almost 30 years ago (that wasn't even accepted into any publications) and fueled an entire generation of luddites.
The value of academic journals today is trust and the internet has not solved that. Pretend if GP had been searching Google, rather than SciHub. There are millions in the US who have used Google to deduce that placing rocks around your house can cure your migraines all because an enterprising individual had gamed SEO.
"It’s too complicated to fix it" is not what the parent poster was implying. If you take a second to breakdown the actual value that journals provide then one clear way it can become much worse is a model similar to the open web where anyone can publish anything with very loose standards for reproducibility or rigor.
Ok, so you admit that the journals aren’t even able to accomplish what they purport to. So if this is already a problem, yet good research is inaccessible, what again is the point of the paywalled journals?
I'm using that as an example of something that could become commonplace if you blow up the journal system. Then every paper becomes "I feel this is right because blog-A which aligns to my beliefs says so". Institutional legitimacy is not something that is built overnight.