Our first undergrad year is common among the maths, CS, physics and mechanics degrees. One of my students wanted to choose CS next year, but never did any CS before and was afraid of not being up to par. Well, in the end she got 17.2/20 in Intro to CS (which is really good in the French system). 😊

Generative AI is a machine built to destroy knowledge: «The latest model of ChatGPT has begun to cite Elon Musk’s Grokipedia as a source on a wide range of queries, including on Iranian conglomerates and Holocaust deniers, raising concerns about misinformation on the platform.»

Now that I’ve (more or less) understood how to get the cheese creamy and not lumpy, cacio e pepe is becoming my favourite pasta.

During this 12th COVlD wave, the CDC reports 1-in-3 states have "High" or "Very High" levels. PMC estimates the proportion of residents actively infectious (prevalence): ◾️USA: 1 in 67 ◾️IA: 1 in 27 ◾️MI: 1 in 25 ◾️IN & CT: 1 in 23 ◾️ME: 1 in 21 ◾️OK & SD: 1 in 17 🧵1/4

"Academia shouldn't require us to spend time on this, it can be automated/replaced!" or "We need to change the system!" Fine. Get a working, consistent, good solution in place. Design a better system and get it adopted: then we'll talk. In the meantime, don't defecate in the swimming pool, please.

The number of people arguing, following the #NeurIPS2025 case, that a bit of systematically bad scholarship and borderline academic misconduct is OK because doing things the right way is tedious and time-consuming is concerning. Yes, doing things well might be boring. But doing them badly is... bad

Apparently the NeurIPS chairs decided to reopen the camera-ready submission to ensure "that the proceedings reflect the highest scientific quality.” @neuripsconf.bsky.social, is that true? So the solution to academic misconduct is to give the authors a chance to cover their tracks?

Fuck you people. Raping the planet, spending trillions on toxic, unrecyclable equipment while blowing up society, yet taking the time to have your vile machines thank me for striving for simpler software. Just fuck you. Fuck you all. I can't remember the last time I was this angry.

> So I asked different AIs… [closes the video]

Donald Trump's "Board of Peace" is the result of the world bowing before the global rampage of the US-Israel Axis. Once again, the Palestinian people are being offered as sacrifices, and along with them, the entire global system of international law.

Today I went for the first time to the new site of the I2M maths lab in St-Charles in Marseille, which is in one of the old buildings (i.e., the nice ones that often have pettable cats outside) and it’s super cute. Please notice the tomettes (provençal hexagonal terracotta tiles) on the floor.

Ma boîte mail contient, à la fois : • “AMU fait partie du programme Alt-Impact, qui vise à encourager la sobriété numérique.” • Promotion chez les enseignant·es de la faculté des sciences AMU du séminaire “Faire des résumés vidéo et générer des quiz de vos cours à l’aide d’un outil d’IAg gratuit”

No idea about this, but reposting for solidarity since I’ve already been in the situation of looking for “lost” audio on the web (I got lucky on Reddit, even if in the end the audio was not really worth it).

La proposition de loi LR visant à interdire le foulard/voile via l’interdiction de tout couvre-chef aux personnes de moins de 18 ans de sexe féminin est: -absurde -liberticide -contreproductive -contraire à la laïcité Elle illustre le naufrage du débat politique actuel. Thread 🧵

Tech "I just had COVID again a few months ago." Me "Not good. Are you doing any vascular diagnostics to track possible vascular damage over time caused by the previous COVID infection?" Tech "Don't tell me that. Now I'm worried." Me 😔 Is it not better to know and act to prevent or remain ignorant?

Calling all the PhDs I know to make monumental annoyances of themselves (and honestly, who better)

☀️ Summer School 📚 “Critical AI Literacies for Resisting and Reclaiming" Organisers and teachers: 👉 @marentierra.bsky.social 👉 @olivia.science 👉 myself Deadline for application: 🐦 31 March 2026 (early bird fee) 1/🧵 www.ru.nl/en/education...

kinda feels like they already lost it. (also, what a thing to say out loud)

We have classes for that. They're called "research methods" or similar, and you need to take them before doing a PhD. Maybe you should also have to take them before joining the board of a top ML conference.

What is wild to me is the defense, BY THE NEURIPS BOARD, that fabricated citations do not mean "the content of the papers themselves [is] necessarily invalidated" It does. It very much does. What do you think citing other work is for? What do you think writing a paper is for? What do you *think*?

(sorry if the tone of my first post sounds sarcastic. I am genuinely mind blown by the fact that senior, established researchers in ML do not know better than using LLMs for something they've been repeatedly documented to be bad at, and for which we've had great solutions for more than a decade!)

Just checking, since apparently some are using LLMs to get BibTeX snippets easily. For Computer Science and ML papers, you know about @dblp.org, right? You know you don't have to use LLMs for that... right?

Unpopular fact: We are entering the 7th year of the Covid pandemic. And no, it's still not a winter cold. 👉 N95 masks in high-risk areas (enclosed spaces, crowded or not, social distancing or not / outdoor spaces if close proximity is required) 👉 Indoor air quality 👉 Vaccination

Unpopular fact : nous entamons la 7e année de pandémie de Covid. Et non, ce n'est toujours pas un rhume qui circulerait uniquement l'hiver. 👉 masque FFP2 dans les lieux à risque (lieu clos, bondé ou pas, distanciation ou pas / lieu extérieur si proximité) 👉 qualité de l'air intérieur 👉 vaccination

"Sous couvert d’aide ou de prévention, elles déplacent subtilement les responsabilités collectives vers les individus." theconversation.com/la-fabrique-...

Une partie de moi n'a pas envie de comprendre pourquoi cette position n'est pas celle de l'immense majorité des universitaires. Une autre partie s'en agace. La dernière continue à mener le combat, parce qu'il est déterminant pour les apprentissages, le service public, la démocratie, le vivant...

Whatever my uni declares, I am not going to engage with AI. I am not going to use AI or teach students to use it. I am going to continue to explain to students why it is harmful. Absolutely absurd to allow this anti-intellectual, environmentally devastating technology into the university at all.

No one could have foreseen this.

Fascinant de voir tant d'experts évoquer l'eugénisme "avant", "ailleurs" ou "sous Trump", soit l'eugénisme des "autres", tout en occultant l'une de ses manifestations contemporaines les plus notables, et à laquelle ils participent activement depuis 6 ans : le laisser-faire sanitaire face au Covid.

Oui mais merci de ne pas créer la confusion, en suggérant d'une façon ou d'une autre, que toutes les critiques de ce texte sont réactionnaires, car c'est faux. Il existe une opposition à la proposition de loi sur "l'aide à mourir" de gauche et antifasciste, qui n'est pas fondée sur la religion.

Based on today's CDC & Biobot data, we estimate the following for the week of Jan 19: 🔸1 in 52 people in the U.S. actively infectious 🔸25% chance of exposure in a room of 15 ppl 🔸Nearly 1 million new daily infections 🔸5 cumulative infections per person all-time (avg) 🧵1/5

Still looking for this! #AI #GenerativeAI #AcademicChatter

It’s like when they started calling Covid “endemic”. It’s already here right, so apparently there’s no reason to try to avoid catching it. (In the mean time, people are still dying and getting disabled.)

#Chicago is a raging inferno of COVlD right now. The CDC places Illinois in "Very High" transmission, especially in Chicago. We estimate that 1 in 17 residents statewide are actively infectious. If people are not testing or isolating, there is a 78% chance of exposure in a room of 25 people.

Le déni de la pandémie de Covid en cours, de ses dangers pour tous, et le gaslighting sur la prévention.

Le fait que la pandémie aurait pu entraîner des changements majeurs et pas qu'au plan sanitaire explique les moyens déployés pour normaliser le "vivre avec" et entretenir le déni de pandémie. 1/5

Experts call for N95s over surgical masks as flu and COVID viruses spread. “From an economic perspective alone, the WHN makes a compelling case for respirators as the standard of care for healthcare workers” — @drjudystone.bsky.social via @forbes.com www.forbes.com/sites/judyst...

[mode blasé ON] La pandémie est toujours en cours et la majorité est dans le déni de son existence et de ses dangers 🤷‍♀️ Et je ne vous parle pas du nombre d'artistes que je vois passer sur mon compte d'illustratrice qui sont/ont été très malades/malades à répétition sans comprendre d'où ça vient.🙃 1/5

typical ai use case

are you displeased with today’s AI safety evaluation landscape and curious about what greater conceptual clarity, methodological soundness, and rigour in AI evaluation could look like? if so, consider coming to Dublin to pursue a PhD with me apply here: aial.ie/hiring/phd-a... pls repost

I go further: when someone claims AI helps students do something, they need to actually show what that is and why AI was helpful. Here, it seems to fill in the gaps from an education system that doesn’t teach basic language skills. It “helps” but is really just another form of outsourcing learning.

Has anyone written a nice statement to include in internship proposals along the lines of “I want to actually do research here, so if you’re planning to use generative AI tools please just go look for an internship elsewhere”? #AcademicChatter

One question in our Intro to CS final was: give two words accepted by this finite automata. A student wrote “accepted words: abba (listen to The winner takes it all by ABBA)”. That made me laugh aloud at first, but sorry, it’s actually –1 points since Waterloo is a better song.

Why are ppl using the text extruding machine? It is often wrong + uses a lot of energy. We have search engines, gadver (which OK, are getting worse cause of genAI but you can assess results based on URL & by checking them). Also embarrassing that he tries to offload his responsibility. #ResistAI

Pendant longtemps j’ai cru que le français le plus difficile à comprendre était celui des petits enfants, mais en fait ce sont les étudiant⸱es qui chuchotent des questions pendant les examens : – Monsieur, wshwshwshwshwsh ? – Hein ? – Wshwshwsh, wshwsh ? – Hein ? – UN INTERCALAIRE S’IL VOUS PLAÎT

"I asked ChatGPT..." Lemme stop you right there, buddy.

The increasingly common habit of screenshotting the AI summary of a Google search and presenting it as evidence must be nipped in the bud immediately. That is nothing. That is not anything, it is not a thing.

"Transparency" is utilised to avoid ethical discussions around genAI. The ppl pushing for it have a very particular idea of what science is & how it is done. It is specially concerning that universities & other institutions fail to see the wider context of this fascist technical apparatus. #ResistAI

complete farce of democracy that polling on medical care should affect treatment “sorry, we can’t give you chemo because the data doesn’t support it. oh, i didn’t mean the medical evidence, that’s rock solid, i meant we took a poll of your neighbors and they decided they’d rather you die, actually”

Malaysia and Indonesia become the first countries to block Grok, the artificial intelligence chatbot developed by Elon Musk’s xAI, after authorities said it was being misused to generate sexually explicit and non-consensual images.