This is a really interesting read from a UK perspective. Good to be reminded that these issues are international, and push back on suggestions that a privatised model is any kind of solution. It lays out how the financial pressures in the US system drive up long bed waits.

“A city with all the right enemies” - delightful column this:

I hesitate to say it, but Janan is right. London rocks, in part because it pisses off all the right assholes The return of Londonophobia - as.ft.com/r/53682c03-1...

This is such a joyful read. Simultaneously nostalgic and inspiring.

New post: Can AI Replace Social Science Researchers? (No. No it can't. Come on, now.) davekarpf.beehiiv.com/p/can-ai-rep...

“We found that using AI assistance led to a statistically significant decrease in mastery.” Props to Anthropic for studying the effects of their creation and reporting results that are not probably what they wished for www.anthropic.com/research/AI-...

Love this advent calendar thread, now made perfect by including the definitive Scrooge on Christmas Eve. Annual rewatch in our house was last night.

Journalist challenge: Use “Machine Learning” when you mean machine learning and “LLM” when you mean LLM. Ditch “AI” as a catch-all term, it’s not useful for readers and it helps companies trying to confuse the public by obscuring the roles played by different technologies. 🧪

Does global rise in diagnosed #neurodivergence reflect an increased awareness or does it indicate something else? The 62nd Maudsley Debate, chaired by IoPPN’s Francesca Happé, aims to tackle that question. #MaudsleyDebates

Okay, for the folks who asked: here's the majority AI view, writing up the reasonable, thoughtful view on AI that the vast majority of people in tech hold, that gets overshadowed by the bluster and hype of the tycoons trying to shill their nonsense. anildash.com/2025/10/17/t... Please share!

Delighted that @maudsleynhs.bsky.social's Clinical Informatics Service has been shortlisted for @rcpsych.bsky.social Digital Team of the Year. More about our team, supported by @maudsleycharity.bsky.social here www.maudsleybrc.nihr.ac.uk/facilities/c...

The responses to this are just what I needed on a Monday morning. Some things are getting better!

City Hall has triggered a high air pollution alert for Tuesday 12 August as hot weather is expected to increase pollutant levels. We're urging Londoners to look after themselves, stop their engines idling, and refrain from burning wood or garden waste.

Trying to change the population is creepy and dangerous, full stop. But pronatalism is also wrong on the science, particularly the US brand of it. Here's a nice explainer on some of its more egregious logical flaws. From @lesja.bsky.social @karenguzzo.bsky.social and @shelleydclark.bsky.social

It's not enough. But it's something. Please give if you can, and tell your reps to stop this horror. chuffed.org/project/help...

I re-read this every year and am always glad I did.

I learned a great deal from this (very long) thread, definitely worth a read to try and understand what the new pope might mean (inc that this US pope won't/shouldn't please JD Vance and his ilk)

Just a reminder that fascists, authoritarians, dictators of all stripes destroy archives— books, museums, databases, &c— because those things provide windows onto a plurality of other ways we might organize the world. And so preserving, defending, and spreading that knowledge harms those regimes.

Woooooooo ⬇️⬇️⬇️ Someone’s done it - archived the entire datasets of the US CDC before the orange toddler could destroy them. #episky #medsky #academicsky

As someone who’s worked with, on, about NHS data, I can’t emphasise enough that it is far far from a carefully curated, nationally complete and ready to use data set. I keep reading otherwise sensible sounding pieces that comment on the uh-mazing data we can straightforwardly exploit.

AI can make errors when integrated into healthcare systems. ...but so do humans. We don't have good processes to correct these errors for humans or AIs. Some thoughts based on three different clinical encounters as a patient, where my clinician(s) used AI - or not. bit.ly/42bkqwG

This is a very thoughtful piece by @alisonleary1.bsky.social in the @bmj.com. In any organisation, we need the right combination of new enthusiasm and recent training, alongside the experience and wisdom of those who’ve been around a while. #MedSky #PatientSafety www.bmj.com/content/387/...

Seems like it is a day to reissue the @britishacademy.bsky.social report from 2012 on the efficacy of league tables in the public sector and especially to flag this point: often the level of statistical uncertainty involved renders them meaningless www.thebritishacademy.ac.uk/publications...

Still growing thread full of advice worth think about... so much that I'm going to have to read it in stages.

Really useful article on how to think about making clinican interactions with AI useful and avoid turning clinicians into 'liability sinks'. Look forward to seeing more coming out of this project.

A quick thread on schools and phones. I'm quite in favour of older kids having phones and managing their own social relationships/travel/finances and learning about how to be good digital citizen. Seems sensible and useful preparation for more independence and autonomy.

It is my conviction that people who understand, at a fundamental level, what computers can and cannot do, are less susceptible to AI hype.

This is a great piece of writing, and shocking because - if this is our big debate, our big issue, that people are just asking “questions” about - we should be hearing this argument all the time, from the other side, as a mainstream take. That’s who was crowding the streets. Speak for those people.

Late #DigitalFriday entry (hey i was trapped in a train with the heating stuck on 🥵) - reminder that if you want functional, AI free google searches, you can have them at udm14.com

If you know of verified fundraisers for amenities, community centres and businesses destroyed by fascist twats in the last few days, please drop them here for a boost. 👇👇👇

FUPS data came up today. It means imperfect data but little chance of getting anything better. F = flawed U = uncertain P = proximal (proxy) S = sparse We need to learn to act with FUPS data. Paper here on how: bmcmedicine.biomedcentral.com/articles/10....