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.

You know what, this might be a rare occasion on which I have written a half-decent paper. Obviously, my co-authors bear the brunt of the responsibility for this. I hope you like the layered infrastructure of information concept! www.sciencedirect.com/science/arti...

what the heck is "awe-dropping" supposed to mean?

I find it weird how often researchers don't seem to understand what an algorithm is or how one works.

This is an excellent article about peer-review, never mind AI. In fact, never mind AI: it is a reminder that science is a practice where the doing is what generates the results. If you stop the doing, where are the results going to come from? This was a problem before AI, now AI is accelerating it.

🚨 New article 🚨 This paper in EBT cuts through the hype and the fury on AI and LLMs. (Have you heard of PLMs? No? I hadn't either.) Kudos to the authors for a sober and informative presentation on the application of language models in systematic review. www.tandfonline.com/doi/full/10....

This absolute banger

I remember the days when I only had to invite 25 people to comment on a manuscript to get 2 reviewers.

A notice to all predatory publishers: this is how to spam your prospective authors.

THIS BIT OF THE PROOFING PROCESS WAS ACTUALLY GOOD! SOUND THE ALARMS!

christ i hate article proofs a live-tweeted thread 🧵

Very much like this spodcast, especially the new alternating deep-dive / current affairs approach they have recently taken. Sub-editor earned their bowl of chips for this episode title. :)

For you editors, systematic reviewers, and SR commissioners out there: our latest reflections on what works in large-scale SR projects to support international policy (it's basically, comms, practice, protocols, researcher buy-in). www.sciencedirect.com/science/arti...

"Undisputedly the world’s best at running a mile while drinking beer." And it sure as hell beats The Hundred. www.theguardian.com/sport/2025/a...

Some people have an entire baby in the amount of time it takes me to record another one of our webinars. But there we go. It's about how AI (which is boring) is good at boring stuff (which is good!). In this case, spotting issues which the peer-reviewer process will often miss. youtu.be/Cw4O8AN9ZZo

This post on @scholarlykitchen.bsky.social was interesting, about use of citations for providing LLMs with more context to help with verifiable claims (for LLMs, context is everything), and also for fairer (i.e. paid) use of literature by AI companies. scholarlykitchen.sspnet.org/2025/08/07/g...

An example of a useful LLM prompt and response. As an editor, I have a hard time trying to locate reusable recommendations for submitting authors. Software citations is a very common one. Rather than trawl manually through reviewer reports, the LLM does it for me. Boring task. Problem solved.

The mind... boggles www.epa.gov/newsreleases...