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...

It says a lot that I think this is perfectly acceptable attire. Even for work.

I am pleased to announce that I look about average for someone over the age of 18 according to the facial models. Also, you should know that said models pronounced me HOT

Replication Crisis xkcd.com/3117/

Your last chance to register for a webinar from your new favourite journal, about how AI could potentially help improve decision-making in academic publishing. Trying to avoid the hype and focus on how computers can help people do some tasks better. events.teams.microsoft.com/event/65c502...

new mixed-methods RCT preregistration just dropped

Pretty much this, yes.

I'm filing this under "theory of everything" and changing "reviewers" to "people".

Here we are folks: demand gold standard science from others, say "fuck it, anything goes" for their own ideas.

Comprehensive, well-researched article in a major UK newspaper on the issues circulating around publication volume and peer-review in scientific publishing, in case you've not already seen it. www.theguardian.com/science/2025...

hahahaha www.theguardian.com/commentisfre...

Are you interested in AI? Are you interested in improving scientific publishing? Do you want to see how the two are coming together - for better or worse? Yes? Well then, you should sign up for this webinar. events.teams.microsoft.com/event/65c502...

Cool. I also saw this honking great reproducibility study from Brazil, which was more geographic than topic-specific. Not had much chance to digest it but check it out: www.biorxiv.org/content/10.1...

I have to say this often enough that I think it is some hard-wired psychological feature of the human brain at play here, but here we go again: just because you can't do an RCT for the question you are asking, it doesn't follow that your observational study is at any lower risk of bias!

Huh look at that, always had me down as a nerd and jokingly as a goth. www.idrlabs.com/jock-nerd-pr...

Pretty much what my surveillance team says about me too, tbh

Probably going to annoy people by posting about LLMs again but... here is the difference between trying to get something useful out of Google in relation to a half-remembered conversation about domain translation (as it turns out it is called) and using Copilot instead. You tell me which is better.

Anyone need some ice for this burn?

Authors wanting to send a paper to my journal are asked to "submit" and I think that's absolutely correct, yes.

I'm listening to the cricket and they were talking about compound dislocations in reference to Steve Smith's finger. I thought I'd google what that looks like. That was a mistake.

I'm not saying that LLMs are fabulous. However, anyone who has read an abstract that I have written by myself must surely appreciate how much better this machine-assisted effort is. (Preprint here by the way papers.ssrn.com/sol3/papers....)