Happy to share that, after a prolonged peer-review process, our review on the major Neuroinflammatory events in #LongCovid was accepted for publication in Translational Psychiatry by @natureportfolio.nature.com! Coming out soon, but preprint available here: www.preprints.org/manuscript/2...

A take I see often online is that it doesn't matter if code is "clean" - readable, simple, decoupled, easy & fast to test - as long as it works. They're missing the point. Code that's hard to understand, complex, tightly-coupled and hard/slow to test is very likely to break. There's no tradeoff.

"Plutôt que de pondre l’idée d’un coupe-file, il faudrait d’abord se poser la question de la toxicité des réformes accumulées ces dernières années sur la santé mentale des jeunes, mais quel politique veut être mis face à ses responsabilités ? @johanfaerber.bsky.social

Yes and, also, Anthropic's marketing, to both its primary customers (military and surveillance) and to us is that it is ever on the cusp of dangerous breakthroughs. These are never materially defined but form a propaganda mist.

Le harcèlement des malades est officiellement envisagé.

These pause AI nonsense calls have been happening for a long time. It's always about power because they wanna consolidate and stop racing against each other for no gains. Here's a previous time of many they tried the same silliness: futureoflife.org/open-letter/...

anthropic: we have no way of controlling our fully autonomous, "recursively self-improving" model hidden gig workers, RLHF annotators, and seemingly invisible coders & academics that are *actually* improving the model: hold my beer africa.businessinsider.com/news/inside-...

We must be crystal clear about a few points. Long Covid is not a "poorly understood neuroimmune disorder" with no biomarkers. It is a widely-studied biological medical condition. Pathology spanning basically all body systems is well documented, with damage in situ shown across organs

Avec #BienvenueEnFrance, le gouvernement espère récupérer 250 m€ dans 3 ans, soit un mois de dividende de la famille Arnault. Cette année, le gouvernement emmerde le CNRS pour 20m€, soit 2 jours de dividende de la famille Arnault.

Wonderful day for a new #LongCOVID paper! Thanks to an incredible collaboration with @VirusesImmunity and brilliant work done by @keylas3, we studied the effects of injecting antibodies taken from people with LC into mice compared with what happened www.sciencedirect.com/science/arti... 1/

"Camp de concentration" with a growth mindset

Using generative AI for imagery is **directly harmful** to artists, graphic designers, photographers etc. This isn't a vague moral argument: this is just normal, standard theft, except at a completely new scale and social acceptability. We need to reverse that. www.instagram.com/reel/DY2L2-g...

"If data centres were a country, they would be the second-largest destination for gas turbines ordered from Q1 2025 to Q1 2026" New IEA report confirms this isn't a normal industry. It is a fossil-fuelled, high-polluting heavy industry producing socially corrosive products at massive climate cost.

Yeah, existing in and engaging with the OSS community is loads of fun, I have no idea why people don't want to publish new OSS anymore, so weird

🇧🇷 Governo federal lança Tela Brasil, primeiro streaming estatal do país. A plataforma, que quer ganhar espaço contra gigantes como Netflix e Prime Video, estreia com 555 obras, incluindo clássicos como Central do Brasil e Cidade de Deus, e é gratuita, sem publicidade e sem rastreamento de dados.

A new study reveals significant white matter changes in patients experiencing neurological symptoms of #LongCovid, linked to ongoing neuroinflammation. (1/2) www.sciencedirect.com/science/arti...

The incidence and prevalence of #LongCovid worse than anticipated in the US, with 1 of 6 persons affected, and of these 89% had at least 1 chronic condition requiring management. From 58 hospitals, 4 regions, 457,950 patients. Prevalence increased thru mid 2024. jamanetwork.com/journals/jam...

FRAIS D’INSCRIPTION À L’UNIVERSITÉ Pour que la France continue de rayonner, nous devons accueillir les étudiants étrangers, pas augmenter leurs frais d’inscription à l’université ! J’ai déposé une proposition de résolution à l’Assemblée. Explications.

L’Europe continue d’améliorer sa qualité de l’air. La majorité des stations de mesure respectent désormais les normes actuelles de l’Union européenne pour les principaux polluants. Cependant l’ozone au niveau du sol et les particules fines restent problématiques. www.eea.europa.eu/en/newsroom/...

Le secret du New Public Management, c’est qu’il n’est contraignant que pour les institutions publiques, pas pour le privé : « le financement automatique du privé par des fonds publics ne serait pas remis en cause par une évaluation négative ».

As an aside: indoor smoking was a very good visual representation of the spread of exhaled aerosol (airborne particles or droplets). It was therefore a really helpful for intuitively understanding COVID spread and how to mitigate. threadreaderapp.com/thread/13157...

Ebola can become airborne. In the same way Norovirus can become airborne. The insistence on not talking about airborne risks is infuriating.

Yep--that made the list of frequenty unasked questions: medium.com/@emilymenonb...

So, after Bluesky mod shenanigans David Gerard (Pivot to AI) is much less visible here but is still posting over on Mastodon https://circumstances.run/@davidgerard/116617321612970768

“liz you can just use AI as a research tool, don’t be such a Luddite” www.nytimes.com/2026/05/19/b...

in the old world, we had to check refs, but we had the added support of knowing the thing was written by someone *presumably trying to cite correctly*. Now, we have to check references knowing it was written by a machine trying to spew vaguely plausible text, and ACTIVELY FABRICATING REFERENCES

I'm sorry, what? In writing my first monograph, I spent six weeks trying to track down a citation in TWO languages I didn't know. And good thing too, because the citation was wrong. That's scholarship. That's research. You know, the thing we're trained to do?!?

the generational de-skilling is almost the entire point of this technology. its architects are very open about wanting to bring about a new feudal order, with themselves as its barons.

"The old world of tech is dying and the new cannot be born" www.baldurbjarnason.com/2026/the-old... This ones leans a bit more into theory and politics than I usually do when writing about the industry, but this moment in time warrants it, I think.

En fait t’es pas obligé de t’en servir et de les laisser changer comment tu fais ton métier Ou te compromettre à utiliser des outils qui ont un tel coût humain et environnemental Tu as ce choix, beaucoup le font, même s’il représente une bataille de résistance contre la vague de la facilité

Virus in the vents: Study traces COVID spread in high-rise apartment According to a study, a COVID-19 outbreak in a residential building in Spain during the early months of the pandemic likely spread through shared bathroom ventilation ducts. Read more: ow.ly/6S7450Z0f8t

arXiv is requiring people who submit work to the repository to.... read it like, one time before they hit send www.404media.co/new-arxiv-ru...

📺🔥🐘 La co-présidente de l'association était présente ce soir, avec Norah Sahara, sur le plateau de 64' de TV5monde pour parler #COVIDLong en lien avec la crise #hantavirus. #PasPrets #OùEstLaPrevention #OùSontLesTraitements #QAI Pour visionner c'est ici 👇👇👇

Recently, FamilySearch digitized and uploaded tons of microfilmed records, including many from 18th-century Massachsuetts. They're using some sort of AI to transcribe/summarize the handwritten documents. I've noticed that the AI strips out references to race and enslavement in 18thc documents.

Some of the solutions seem counterintuitive, but they include 1. Grades on individual assignments having less value in the system, not more. 2. Helping students build the capacities that allow them to work independently from a place of curiosity. 3. Framing learning as a communal act. (cont.)

> Despite the responses being identical for the UK and US, Copilot produced a rich, detailed summary of how US and UK respondents differed. I know too many people who are using LLMs for exactly this kind of data analysis and refuse to listen to me when I try to warn them

First they came for gender and post-colonial studies and I said nothing because I was in STEM.

OUR NEW BMJ editorial: Andes virus should reset WHO’s default approach to airborne risk. 🧵 1/ For severe pathogens with person-to-person transmission, the starting point should be precautionary airborne protections — not waiting for definitive proof after spread occurs. www.bmj.com/content/393/...

The paper that forced global acceptance that COVID‑19 is airborne: 'Ten scientific reasons in support of airborne transmission of SARS‑CoV‑2', @trishgreenhalgh.bsky.social et al., published in The Lancet (2021). 💪Evidence that conclusively ended airborne denial. www.thelancet.com/journals/lan...

#Hantavirus C'est peut être le moment de rappeler qu'il y a des asso qui font un travail incroyable d'information sur les virus aéroportés, la qualité de l'air et les moyens de prévention : @winslowsp.bsky.social @associationarra.bsky.social tous les mask bloc ex: @maskblocparis.bsky.social

This is lab policy at @epic-biodiversity.org - the abridged version of our policy on generative AI is "no". This is not a guideline or an aspirational document. It's part of our rules of behavior and good academic conduct. epic-biodiversity.org/generative-ai/

It's not the printing press. It's asbestos.

Penser qu'on peut enseigner aux élèves à "bien utiliser les IA", à "conserver un esprit critique" en leur faisant tester ChatGPT ou autre IAg grand public, c'est comme penser qu'on peut leur enseigner à manger de la salade en les emmenant au McDo : c'est au mieux naïf, plus sûrement grotesque

Excited to work with @Invivyd to get this out. One of the most common questions I get from people in the #LongCOVID community are about vaccination alternatives. People still want to protect themselves from SARS-CoV-2 infection but many experience www.medrxiv.org/content/10.6... 1/

"Le cadrage «cerveau en mode défense» requalifie le problème collectif en problème individuel de perception erronée, à traiter par exercices comportementaux… un peu comme on traiterait une souffrance généralisée au travail en préconisant des exercices de yoga en espace-détente."

🚨🗞️ Must read : "Non, le yoga n’est pas la solution au Covid long" sur le Colloque Long COVID EU Project (nous confirmons!), "rapidement interrompu par des personnes vivant avec le Covid long, s’insurgeant des thèses défendues par certains des protagonistes

Fabricated citations in the biomedical literature have increased 12-fold in two years, according to an audit of nearly 2.5 million papers published as a letter to The Lancet today.

On mesure, on mesure : on s'apprête à fabriquer des générations futures qui délégueront la majeure partie de leurs tâches cognitives, créatives et expressives à des automates, et qui par là perdront graduellement d'essentielles capacités humaines et donc toute possibilité d'émancipation.

Imagine being a scientist investigating vaccines in the U.S. After years of research and millions spent, your study is blocked from publication because your data doesn't support RFK's and the FDA's battle against vaccines. This is why people like me were forced to leave.

"we are living through attacks on our science from both the outside & inside wherein non-theory & non-science produced by industry players is presented as computational modelling of the phenomena and capacities we study, and worse as theories thereof (van Rooij & Guest, 2026; van Rooij et al. 2024)"