more than anything, this should remind you how bad generative has ruined the information ecosystem

CAIL = prerequisite knowledge for a critical perspective, such as to tell apart nonsense hype from true theoretical computer scientific claims. For example, the idea that human-like systems are a sensible or possible goal is the result of circular reasoning and anthropomorphism. olivia.science/ai 🧵

Si nos importa hacer campaña/tener conversaciones efectivas, en vez de hacernos señalamiento virtuoso entre nosotros

A veces veo esas listas tipo "razones para no votar por Kast" y, estoy de acuerdo con todas,pero al mismo tiempo esas son todas las razones por las que mi mamá si vota por Kast (anti inmigracion, reducción del sueldo minimo, anti aborto, etc). Como izquierda tenemos q entender como piensa la derecha

lo bonito si fue ver a la gente votando acá tan lejos de Chile :)

que stress esperar los resultados de la elección wnooo

“While the AI industry claims its models can “think,” “reason,” and “learn,” their supposed achievements rest on marketing hype and stolen intellectual labor. In reality, AI erodes academic freedom, weakens critical reading, and subordinates the pursuit of knowledge to corporate interests.”

extraño trabajar con estudiantes universitarios de chile, lo gringos del maple son full clientes de la universidad. La actividad política es casi nula y se gasta mucho tiempo en pelear las notas en vez del aprendizaje. Obviamente, la culpa la tiene la estructura que te pone 400 cabros en una sala

and then they hit you back with a "oh, so you don't like my wording", and I mean, yeah, because what they say is wrong or imprecise, not just because I have a liking for other words or a different order of them.

It is hard to grade university assignments when it is clear that what a lot of the students need first is some highschool level writing and computer classes. Like, I'm not grading them on that, but it causes some very consistent mistakes of the "you're not saying what you think you're saying" kind

Is inbox zero like patient zero? We need to find it so we can stop the emails from spreading further

Deep truths of Expedition 33

Once again:

Canada's Budget 2025 just dropped. It contains an admission that men and high-income people will benefit the most from the government's new AI funding

AI slop and the destruction of knowledge irisvanrooijcogsci.com/2025/08/12/a...

"a lil bit of austerity is ok we promise the AI will fix it"

One of the best things about finishing the phd is that I don't have to do any more research😂, I'm so done with the whole system, high effort, the professional risk of narrowing your possibilities, not so great rewards

Document Forgery xkcd.com/3160/

heeeeeey I'm a doctor in psychology everyoneeeeeee🥳🥳🎉🎉

plw argentinaaaaaaaa

How are auditory features represented across the #AuditoryPathway? @bathellierlab.bsky.social show that #auditory feature tuning emerges through a mix of gradual & stepwise changes, with early representation of simple features & later refinement of complex ones @plosbiology.org 🧪 plos.io/4oi0XCY

Crazy to think I'll be defending my dissertation in a week. It is also crazy to think about how ANN's have become so relevant since I started. I remember when connectionism used to be this weird dead thing, and while that is still true (lol), concepts from the field have regained importance.

I think this is true in a deep and troubling way. I have been observing this too. I am getting old & maybe I’m wrong, but I really feel basic academic, critical and intellectual, attitudes that were normal and prominent 20 years ago seem to have eroded. I agree, we need to understand how and why.

important on LLMs for academics: 1️⃣ LLMs are usefully seen as lossy content-addressable systems 2️⃣ we can't automatically detect plagiarism 3️⃣ LLMs automate plagiarism & paper mills 4️⃣ we must protect literature from pollution 5️⃣ LLM use is a CoI 6️⃣ prompts do not cause output in authorial sense

I propose we purposefully start poisoning LLM training data

Looks like Canada's approach to "sovereign AI" will involve creating structural dependencies on large US tech companies

Interested in what goes viral online and how social media intersects with rising political polarization? Check out the latest episode of the Stress Puzzle where @jayvanbavel.bsky.social walks us through the implications of this research: www.stressmeasurement.org/podcast/epis...

Love this, gonna send this to everyone who says "AI" is a tool like any other technology

bonus on calculators and AI bsky.app/profile/oliv...

‘There will never be enough computing power to create AGI using machine learning that can do the same [as human cognition], because we’d run out of natural resources long before we'd even get close,’ — @olivia.science www.ru.nl/en/research/...

There is absolutely no good faith reason to use the term “AI” for any technology one is selling. It serves only for dazzling people into thinking the technology has capabilities that it doesn’t. If one wants a technology to be trustworthy, just use a transparent, informative term without hype.

“In this position paper, we expand on why universities must take their role seriously to a) counter the technology industry’s marketing, hype, and harm; and b) safeguard higher education, critical thinking, expertise, academic freedom, and scientific integrity.” — from: zenodo.org/records/1706...

💕🤝🙏

In preparation for dissertation defense, it feels very weird to read things and be like "I did that?"

Today's delight: Tom Morello's *Fuck ICE* playlist - music.youtube.com/playlist?lis...

There's a trend going around TikTok to have ChatGPT create visual alphabets (because the output is always full of ridiculous mistakes). I generated this one to use an example to help explain WHY AI image generators make these kinds of mistakes. It turned out even better/worse than expected.

But they're marching to Bastille Day The guillotine will claim her bloody prize Free the dungeons of the innocent The king will kneel and let his kingdom rise www.youtube.com/watch?v=9L1w...

Only a small % of people engage in toxic activity online, but they’re responsible for a disproportionate share of hostile or misleading content on nearly every platform Because super-users are so active, they dominate our collective impression of the internet www.theguardian.com/books/2025/j...

Since I 1st released this short blog post, many seem to have found it useful. I have now obtained a doi to facilitate citing it. van Rooij, I. (2022) Against automated plagiarism. doi.org/10.5281/zeno... (Thanks to @olivia.science for encouraging me to do so).

heavy como un sector supuestamente progre parece ignorar completamente la relación entre el hippi-naturalismo-new age y el crecimiento de la ultra derecha. Es el anti intelectualismo y la erosión de mínimos comunes en lo que constituye evidencia

which is one of the big reasons why real scientific communication is so important. (And by real I mean "not overly simplified social media content designed for algorithmic success")

I had not thought of this, but of course, AI slop is just another ponzi scheme. It is interesting how so many of the modern ponzi schemes are either pseudo-scientific, or under a 'tech-wizardry' blanket. Basically "We'll use just enough jargon so general public will be confused"

AI's key promise (that you too can grift to success) is also its biggest weakness. Despite ridiculous amounts of investment, Anthropic had to raise prices. Which means its customers (Cursor) had to raise prices. So THEIR customers (grifters) have to raise prices. And their customers (businesses)...

💕🫂🛌

New paper hot off the press www.nature.com/articles/s41... We analysed over 40,000 computer vision papers from CVPR (the longest standing CV conf) & associated patents tracing pathways from research to application. We found that 90% of papers & 86% of downstream patents power surveillance 1/

Hello & HAPPY FRIDAY, as promised an open letter by my lovely colleagues and myself. Everybody, feel free to sign this even if you are not NL-based and ‼️ share ‼️ (anonymous signature is poss): Open Letter: Stop the Uncritical Adoption of AI Technologies in Academia openletter.earth/open-letter-...

This will go down as one of the most important AI Ethics piece.

No more complacency about the data collection practices of social media corporations please. “Why should I care what they collect” this is why. How much longer until app makers sell personal information to third-parties who will weaponize it in this exact way? Likely already happening

What a great time to be living in the country next door to the US /s