I keep hearing about the "slopocalipse" coming to scientific publishing, and I can't but think: it was already happening before LLMs became "good enough" at writing academic prose.
-
@albertcardona A significant number of people, myself included, have gone back to the 18th century and just self publish on a blog. Our peers who respect us (and someday I will have one) can read our work and make their own judgements.
This does give me the idea of using "web of trust" to add peer review on top of the blogosphere. After all, publishers have no real value add other than curated peer review...
Ideally, all publishers should run like the Journal of Machine Learning Research: inexpensive, based on latex templates, fully run by academics.
https://blogs.harvard.edu/pamphlet/2012/03/06/an-efficient-journal/ -
I keep hearing about the "slopocalipse" coming to scientific publishing, and I can't but think: it was already happening before LLMs became "good enough" at writing academic prose. What hasn't been addressed is the root cause: the incentives. The publish-or-perish approach to science evaluation, for grants and positions.
Want a solution? Limit publications to one per author per year. You can publish more when in collaboration, i.e., make it fractional. The end goal: make people think what brick they want to supply to the edifice of science. As a bonus, *reward* scientists when they publish *less than* one paper per year.
And the business models based on journal APCs can all die a sudden death now.
@albertcardona I agree about the underlying cause. I think an issue with one paper per year (besides the fact that it would be completely unenforceable, but let us dream for a moment) is limiting publications would be bad for trainees. They do some work and find something, it needs to be reported even if it's not much of a brick. Labs chasing a significant brick tend to burn through trainees to do it.
I suppose they could publish without the PI… -
@albertcardona I agree about the underlying cause. I think an issue with one paper per year (besides the fact that it would be completely unenforceable, but let us dream for a moment) is limiting publications would be bad for trainees. They do some work and find something, it needs to be reported even if it's not much of a brick. Labs chasing a significant brick tend to burn through trainees to do it.
I suppose they could publish without the PI…@steveroyle @albertcardona How about this: Let people publish as much as they want, but evaluate on just one paper per year at max? Lets the ball roll when it’s necessary, but avoid punishing slow science.
-
@steveroyle @albertcardona How about this: Let people publish as much as they want, but evaluate on just one paper per year at max? Lets the ball roll when it’s necessary, but avoid punishing slow science.
@roaldarboel I like that idea! @albertcardona
-
I keep hearing about the "slopocalipse" coming to scientific publishing, and I can't but think: it was already happening before LLMs became "good enough" at writing academic prose. What hasn't been addressed is the root cause: the incentives. The publish-or-perish approach to science evaluation, for grants and positions.
Want a solution? Limit publications to one per author per year. You can publish more when in collaboration, i.e., make it fractional. The end goal: make people think what brick they want to supply to the edifice of science. As a bonus, *reward* scientists when they publish *less than* one paper per year.
And the business models based on journal APCs can all die a sudden death now.
Yes, we need to get back to quality of research not quantity.
-
I keep hearing about the "slopocalipse" coming to scientific publishing, and I can't but think: it was already happening before LLMs became "good enough" at writing academic prose. What hasn't been addressed is the root cause: the incentives. The publish-or-perish approach to science evaluation, for grants and positions.
Want a solution? Limit publications to one per author per year. You can publish more when in collaboration, i.e., make it fractional. The end goal: make people think what brick they want to supply to the edifice of science. As a bonus, *reward* scientists when they publish *less than* one paper per year.
And the business models based on journal APCs can all die a sudden death now.
@albertcardona 'slopocalypse' if you wanted to follow the apocalypse portmanteau formula..!
-
I keep hearing about the "slopocalipse" coming to scientific publishing, and I can't but think: it was already happening before LLMs became "good enough" at writing academic prose. What hasn't been addressed is the root cause: the incentives. The publish-or-perish approach to science evaluation, for grants and positions.
Want a solution? Limit publications to one per author per year. You can publish more when in collaboration, i.e., make it fractional. The end goal: make people think what brick they want to supply to the edifice of science. As a bonus, *reward* scientists when they publish *less than* one paper per year.
And the business models based on journal APCs can all die a sudden death now.
@albertcardona Was it Fred Sanger who won his Nobel on the back of just 14 published papers? I'm probably not remembering that correctly but it's difficult to imagine that being possible today.
-
@roaldarboel I like that idea! @albertcardona
@steveroyle @roaldarboel @albertcardona
Alternatively, castigate institutions/countries that have requirements for publication numbers that exceed that limit.
This is a major driver.
It's not a problem that a well-funded lab with 7 postdocs and numerous collaborators is publishing two Cell papers a year.
It is a problem that a poorly funded lab in India needs to publish two papers a year for the PI to keep their job.
-
I keep hearing about the "slopocalipse" coming to scientific publishing, and I can't but think: it was already happening before LLMs became "good enough" at writing academic prose. What hasn't been addressed is the root cause: the incentives. The publish-or-perish approach to science evaluation, for grants and positions.
Want a solution? Limit publications to one per author per year. You can publish more when in collaboration, i.e., make it fractional. The end goal: make people think what brick they want to supply to the edifice of science. As a bonus, *reward* scientists when they publish *less than* one paper per year.
And the business models based on journal APCs can all die a sudden death now.
I work in way too many fields to publish one paper per year. I have about a dozen ongoing projects, each of which provides regular breakthroughs that need to be communicated to the scientific public.
I would actually like to see more and smaller papers published. We reward big labs publishing monster papers with a dozen experiments none of which are well enough described to replicate. I want to get back to the real science journals about nature (note the LACK of capitals at the start of those two words) in which people publish experiments, theories, or models... separately, and we can judge the literature as a literature not as a paper.
-
@steveroyle @albertcardona How about this: Let people publish as much as they want, but evaluate on just one paper per year at max? Lets the ball roll when it’s necessary, but avoid punishing slow science.
@roaldarboel @steveroyle @albertcardona not sure it helps. It still gives an advantage to publish more because you can select the best one.
-
@roaldarboel @steveroyle @albertcardona not sure it helps. It still gives an advantage to publish more because you can select the best one.
@neuralreckoning @roaldarboel @steveroyle
Plus it's an abuse of the commons – there are finite reviewers and reviewing hours. Better less but good, complete, strong papers.
-
@albertcardona I agree about the underlying cause. I think an issue with one paper per year (besides the fact that it would be completely unenforceable, but let us dream for a moment) is limiting publications would be bad for trainees. They do some work and find something, it needs to be reported even if it's not much of a brick. Labs chasing a significant brick tend to burn through trainees to do it.
I suppose they could publish without the PI…Publishing without the PI should be the norm. Just put the PI in the acknowledgements.
-
Publishing without the PI should be the norm. Just put the PI in the acknowledgements.
This should depend on whether the PI did any of the work. In the fields and labs where I live, the PI is *always* deeply involved in some aspect of the work.
I find the idea that the postdoc somehow does the work without the PI as incredibly weird. It does not describe any of the laboratories I see. I suppose YMMV, but I've seen a lot of labs and none of them look like that.
Papers are collaborations among authors, with (hopefully) all of the authors contributing something. In every laboratory I have seen, the PI is definitely involved at many stages of the experimental design, theoretical and conceptual structuring, and writing.
Science is a collaboration. We all have roles to play and together we make discoveries that can change the world.
-
This should depend on whether the PI did any of the work. In the fields and labs where I live, the PI is *always* deeply involved in some aspect of the work.
I find the idea that the postdoc somehow does the work without the PI as incredibly weird. It does not describe any of the laboratories I see. I suppose YMMV, but I've seen a lot of labs and none of them look like that.
Papers are collaborations among authors, with (hopefully) all of the authors contributing something. In every laboratory I have seen, the PI is definitely involved at many stages of the experimental design, theoretical and conceptual structuring, and writing.
Science is a collaboration. We all have roles to play and together we make discoveries that can change the world.
@adredish @albertcardona @steveroyle
The reason we have open science contribution criteria (e.g., CRediT)… -
R relay@relay.mycrowd.ca shared this topicR relay@relay.an.exchange shared this topic
