(all this said, I made the decision to decline all reviews during my sabbatical and also decline bc of bandwidth reasons sometimes, so no judgement. it's just too much!)

It seems like there's been a "learn to say no to things" movement (punchcards where if you say no 10x you get ice cream, etc.). We also need to teach people that there are limits to that and if you are going to participate in the system you have to help maintain it, even if the incentives suck.

It is also an ethical value that we need to collectively instill & embrace. People (esp younger gen) are getting better at declining excessive service demands, which is a good thing. But people at all career levels who don't review regularly but still submit to peer review journals are freeloaders.

I absolutely think there's training value. And, when I'm on search committees, I'm totally put off when ass't prof candidates only have a couple journals on their ad hoc review list, and conversely impressed when they have a huge number. But, it's hard to list number of REVIEWS, only # of JOURNALS.

Reviews (unfortunately) don’t help one’s cv or visibility much, in proportion to the considerable effort involved. That is what needs to change— add incentives (recognition for hard work/insights, or financial). Publons doesn’t cut it.

When I was a postdoc I was already getting too many reviews to handle. And, when I ask my students to co-review, 95% of the time they say no because of “low bandwidth.” I don’t think junior scholars need or appreciate doing reviews more than senior ones, nor should the burden get shifted to them.

When we coauthored a PB paper, the AE was more thorough and careful than almost any I’ve ever worked with, and the manuscript was longer than my dissertation. If I had done what he did, it would have taken weeks. I quickly resolved to never, ever, ever join PB’s Ed board.

Dear lord, is a requirement for peer reviewing conducting the systematic review yourself?

I know some people love to fantasize that there's a crusade against everything they like so they can wallow in sweet, sweet self-righteousness. But protein???

We keep saying it because it keeps being true. He'll leave such a body count. And it won't always be immediate, clearly discernible effects: these policies will impact people for generations & through both proximal and distal, insidious downstream consequences.

I’m so sorry about your patient loss. That’s a terrible thing to endure but especially as a trainee.

This is a monstrous, deadly plan. It also seems like a political loser for RFK. People of all walks of life take antidepressants and will feel personally attacked.

It’s good to debate better norms for AI research. And lots of us are deeply concerned about Big Tech AI models scraping our scholarly work. Hence viral thread.

But, public claims of “harm” or misconduct are serious and can themselves cause reputational damage, especially for grad students.

The thing is, even if you don't like it, analyzing your paper likely doesn't meet the formal definition of human subjects research so consent is not required. And IRBs define "harm" narrowly and don't include things like being personally offended or mildly stressed out. (1/2)

Still sure how you were "harmed" exactly or even how you were involved if all they just ran your paper through an algorithm.

I guess I have a high bar before I publicly launch a crusade against a grad student's research, but OK.

How are they analyzing you? It seems like they analyzing a paper you wrote. Of course, they need consent for asking you questions about your reactions to it, but they are seeking that.

I can understand not wanting to participate if you don't want to help further the development of these kinds of models on principle. But from an IRB perspective, I don't think your consent is required for them to use your papers.

To play devil's advocate, do you object when researchers use publicly available blogs/tweets to conduct sentient analysis to predict posters' emotional states (much more personal than research summaries & doesn't req consent)? Is it that they contacted you that you found objectionable, or use of AI?

This dude thinks SGM people make up 1/100,000 of the population. Who wants to tell him???