Just as publishers are waking up to the threats from paper mills, they are adopting a new approach to keep one step ahead, as explained in this guest blogpost by Matt Spick. deevybee.blogspot.com/2026/07/how-... #researchintegrity #publishing #papermills

Ahh this is so great I'm gonna have to do my own. Poor old Mondrian. :D

@twpiggott.bsky.social To get you in the mood for the World Cup!

Recommend reading Silverchair's sober analysis of the state of peer-review in 2026. "Before accepting a crisis narrative ... the report asks for evidence. What the report finds is more complicated, and (hopefully) more useful, than a simple alarm." www.silverchair.com/news/the-num...

Three bloody cheers for @pollytoynbee.bsky.social, the best defence of selective screening programs I have read in a long, long time. www.theguardian.com/commentisfre...

I like this take, it helps make sense of the avalanche of I'm not sure exactly what to call it that has in the last year made academic LinkedIn a haystack of bullshit and a needle of information. (Neal Stephenson in Anathem did, in fact, call it bullshit and you should all read that book.)

Here it is: my official audition to play the part of Ozzy in the upcoming Ozzy Osbourne biopic. NOTE: I can also play guitar, sing and drive a car. I am willing to gain/lose weight for the role, and will perform in the nude if it serves the story. #audition #ozzy #blacksabbath #acting #actor

Tangentially, I think this is a structural problem for researchers as well - I feel acutely disadvantaged by not having a university infrastructure around me. However, that infrastructure comes with its own demands that may not ultimately be a net benefit (so I try to fret less about that). 1/2

great, now i have to reconsider atheism

preach

Just to get this on the record, the term "chinfluencers", used to describe researchers who develop high-profile social media accounts by relentlessly posting dubiously helpful science commentary, is mine. Thank you for your attention.

This is what the kids call a deep cut.

I'm sorry, but "Hitler Lewis"? It only gets better when you discover one of his main co-authors is Innocent Benjamin. cen.acs.org/research-int...

the brutality

Very nice essay from your man @iandunt.bsky.social on how to think about AI. It is a tool. (Could we expand on the concept of tools? I am not sure people know how to think about tools either: know your task, and judge the tool against how it performs that task.) iandunt.substack.com/p/how-to-thi...

Ribbit.

old: two spaces after a full stop. bold: one space after a full stop. cold: carriage return after a full stop.

If it makes you uncomfortable, think how I feel? That's my great-grandpa Maurice in the fourth pic.

I don't think we should require anyone to have a licence to be a researcher, but I do think we should require a researcher to have a licence to use a keyboard.

Classic UK approach. Figure out how fast the horse is bolting and where it is bolting to, then shut the stable door. www.theguardian.com/environment/...

writing: i have beautiful cogent thoughts how can they come out so wrong editing: fuck you very much for that writing paul nobody can understand thanks for nothing checking proofs: huh i guess writing paul won the battle of attrition with editing paul reading my published work,ever: <snorts>

I think autocorrect would work a lot better if it assumed you typed a valid spelling correctly, rather than assuming you meant to type a different, correctly-spelled word.

Right that's it I've decided: there is no such thing as construct validity, there is just internal validity and external validity. Construct validity is a type of external validity, where you are measuring a surrogate for the thing you are interested in (e.g. IQ instead of intelligence).

I love methods, but why oh why did they decide to call it Love Methods Week?

Equating an individual with their culture *and* saying that some cultures are worth less than others is racism squared, just get the fuck out of here. As if doing just one of the two wasn't already bad enough.

"Are we still capable of talking to each other," asks the woman hired by billionaires whose economic model is predicated on fomenting polarization to install an actual criminal who dismantles the regulatory state that kept them from fully extracting what remaining wealth there was from the lower 99%

Starting interviews today. 10 points for anyone with the stones to look me in the eye and say "I'm not an applicant, I'm an applican", in a low-key way like they are serious about it.

Crikey, HSBC no longer accept international cheques. They would charge me 30 quid a throw for these, and even at that rate they think it is STILL not worth it.

If you want an object lesson in how generative AI works, look no further than this checklist for (checks notes) "stalitative methods". 😊

SKULL OF THOMAS AQUINAS: TAKE A LEFT NOW PRIEST: No, the GPS says we have to keep going— SKULL: I KNOW A SHORTCUT PRIEST: Do you remember the last ti— SKULL: FOR THOSE WITH FAITH, NO EVIDENCE IS NECESSARY; FOR THOSE WITHOUT IT, NO EVIDENCE WILL SUFFICE

Huh, so in GMail in the UK, the trash can is the "bin" (yay, culturally appropriate) but if you want to search in the bin you have to type in:trash so what the hell

The truth is out.

ChatGPT with an absolute banger.

Metal fans, heed the words of Powl Hughnermunt

Hell yes I am going to get the sysadmin* to enable emoji reactions in gmail. *I am the sysadmin.

The speed with which we went from "risk of bias assessment is bad because it doesn't make enough use of human expertise" to "can't we get an AI to do risk of bias assessment" suggests that human experts just didn't fancy the work.

I KNEW THIS BEFORE THE STORY EVEN BROKE. JFC Guardian.

HMRC are doing webinars to explain Making Tax Digital for self-employeds. My experience of HMRC trying to be helpful has been pretty positive so I would recommend signing up if you are affected. Also, it is cute that someone is still using gotowebinar. 😻 register.gotowebinar.com/rt/467677425...

Given that many of the problems we have in research stem from appeals to authority, Daniel's blog post is a welcome reminder of other ways of solving the problem of proliferation of misused, misunderstood, and insufficiently valid methods.

Today I decided not to frame an article about the future of peer-review in terms of autocannibalism.

WE NEED YOUR THOUGHTS, systematic review practitioners. Fill out this 10-15 minute survey about pain-points and challenges in SR methods in toxicology, environmental health, and related fields. Share or be square! www.surveymonkey.com/r/EHSR25

Don't know if this is your bag, but if it is - register, and listen to two people try to have a sensible discussion about the role of LLMs in research (specifically, systematic review). Lena Schmidt knows her stuff, it should be good! events.teams.microsoft.com/event/e4e8f0...

Seems consistent enough for a society that has had its fair share of eugenicists. 🤷

I did not realise the reviewer comments are available for the tylenol / autism SR. They are a fantastic case study in how editors can screw up, allowing authors to fake out a revision process because they are not sufficiently experienced to see what is going on and/or empowered to call it out.

Just wanted to share a really nice poster by @carrieprice.bsky.social and colleagues. As someone who deliberately invited librarians a lot when I was editing systematic reviews, I can vouch for what they add. Specific info about attitudes and priorities is really interesting. osf.io/8khyp

I have edited 500 systematic reviews and this is one of the worst pieces of work I have ever seen published - certainly in terms of salience to quality. I don't know where to begin. What was the journal thinking? (Probably not much.) Utterly egregious. ehjournal.biomedcentral.com/articles/10....

Want some free career advice? #AcademicChatter

Just gonna say that this is a foundational principle of EBTC (ebtox.org) that I work for, where part of the function of the collaboration is to provide support for people who might be marginalised in their workplaces, and we have a cast-iron don't-be-a-dick policy for involvement.