It was a crisis the moment that the Journal system was allowed to be built without any oversight from academia or the government itself.

A whole class of intelligent professionals performing a public good on public funds were tricked into doing most of the work for profitable companies who then charged these same people to read their peers’ work. They were tricked by appealing to their ego and longing for prestige.

And so now we have incessant publication/citation metrics and a tsunami of untrustworthy papers while shitty corporations rob the public funds of money that could much better spent.

Once that system was built, it was a crisis. Not just because it exists (and naturally leads to the current situation and worse). But because it was allowed to come into existence, which speaks to an intrinsic inability to self-govern on the part of the academic community.

“Publish or perish” is an expression that’s been around since forever and it’s well ingrained into every researcher’s mind so…

What did society expect?

(Not so) Fun story: when a friend of mine was doing her PhD she was trying really hard to reproduce an experiment published on Nature by two Harvard postdocs at the time. She was so frustrated because she couldn’t reproduce it, so she approached one of the authors during a conference and he candidly admitted the experiment was utterly wrong, since after publishing it they realized they made a fatal mistake in interpreting the result which invalidated their claims.

They published the original paper honestly, since they were not aware of the mistake at the time, but they willingly decided not to retract it since a paper in Nature is always a paper in Nature and the citations piling up were too important for their career… How about that for the intellectual honesty that scientists project having as an aura?

Anyhow, this nearly killed my friend’s PhD, but luckily she switched to something related she managed to understand and graduated…

Yeah bruh, a lot of the literature is irreproducible horseshit.

This certainly mirrors what I’ve seen on the ground. In the last 10 years, predatory publishers and publication mills went from known issues to the new normal. And yet despite how easy it is to publish, interest in reproducibility seems at an all-time low. It’s jarring, and I’m kind of making a career change out of research as a result, because what I do as a lowly assistant is essentially engineering results and marketing as opposed to anything having to do with discovery or science. I interpret it as capitalism’s going to capitalism.

This has resulted in a 47% growth between 2016 and 2022 in the global number of published papers (Hanson, et.al. 2023). Moreover, we should expect a further spurt of growth following the widespread advent of large language models in late 2022. During the 2016-2022 period there was little net increase in the number of PhD students globally or in the funding of science, both indicators of science activity. Increased paper productivity implies either that scientists became suddenly much more creative over the period, or had spent more time writing, and therefore reviewing papers: an increase in paper productivity but a decrease in scientific productivity.

This is having big real world consequences. In Long Covid research there has been a tonne of near duplication of work and its apparent none of the work is really building on prior work as the sheer volume of papers is impossible to keep up with. Most of its unremarkable in the sense it hasn’t moved further than findings on ME/CFS from decades prior, so much of the work is too shallow to be of use.

Then the other side of this is the psychology side of things which has been publishing some grade A nonsense and none of the findings hold up to any scrutiny once a replication is attempted.

There there is all the widespread fraud where medical images have been fabricated in various ways, the data often shows clear signs of fabrication as well.

Its a real mess and its harming real people who need this research to inform proper treatments.