Jean-François Bonnefon

The canvas() function from the ggview R package is very useful for previewing/tweaking a ggplot into publication-ready format: it renders a plot "as it would appear if saved to a file with the specified dimensions".

cran.r-project.org/web/packages...

#RStats #ggplot #ggview

[user] analyze and summarize the attached document

[French LLM] *Bonjour*

My zombie Prof would complain bitterly about \textit vs \emph and the use of raw text in math mode latex instead of using \text or proper operator commands. 💀💀💀

What a better way to start the day than reading the sentence “Grammarly is being accused of necromancy”. Now *that’s* my kind of news

This piece by psychologist/philosopher Brian Earp is about the differing attitudes toward AI in psychology vs. philosophy

In psych, one common attitude is; The goal is to get the right answer to fundamental questions in psychology. If AI helps us do that, well then, all the better…

1/

One month to wait for Trace Elements, a conversation on my favorite genres by the author of my favorite series @adapalmer.bsky.social and the person whose taste in books I admire the most @bluejo.bsky.social. I’ve not been that excited about a book for a very very long while

Also: what a missed opportunity to make us Jonathan and Jennifer

My phone decided to autocomplete “Je” (French “I”) with Jérôme and my wife’s phone decided on Jennifer. from our texts you’d think we are obsessed with these two

- Jerome will be late tonight
- it’s ok, Jennifer will be at yoga
- Jerome will stop for food on the way
- Jennifer won’t need anything

#Wilmasreview ❄️ Wilma isn't looking for a new home in the snow — but she is wondering: should AI decide who gets welfare benefits like housing support? A new study by Dong, @jfbonnefon.bsky.social & @iyadrahwan.bsky.social looked at exactly this. 1/🧵

Pixel Agents

A VS Code extension that turns your AI coding agents like Claude Code into animated pixel art characters in a virtual office. It shows what the agent is currently doing with animations. You can also customize the virtual office however you want.

Repo: github.com/pablodelucca...

I went a little overboard with this and a little bit insane. moultano.wordpress.com/2026/02/22/t...

I do it when I'm stressed, these emails have such a soothing rhythm

Join us in June at @iast.fr Toulouse for a 2-day workshop on social influence and its large scale implications!

I propose to make universal the old policy of the Blackfriars conference at the American Shakespeare Center:

If you do not end your paper on time, you will be forced to exit, pursued by a bear. Literally, a bear will come take your paper from you.

Jo Walton’s Reading List: January 2026 - Reactor Compelling memoirs, Vikings, Heyer, and early experiments in authoritarianism

My January Reading List reactormag.com/jo-waltons-r...

Hacking Moltbook: AI Social Network Reveals 1.5M API Keys | Wiz Blog Learn how a misconfigured Supabase database at Moltbook exposed 1.5M API keys, private messages, and user emails, enabling full AI agent takeover.

“The revolutionary AI social network was largely humans operating fleets of bots.”

I saw some people calling it Rectangular February

NGL this description of Bsky in Vulture cracked me up

There are thousands and thousands of papers with an abstract that starts with the phrase

“With the rapid development of Artificial Intelligence,”

and it looks like 99% are from Chinese authors. How come? Does that sentence sound really good in Chinese? Are Chinese LLMs particularly fond of it?

me and the peer reviewer i suggested

The claim in this article is not supported by the data. SGO and NHTSA general public data have different reporting thresholds, and the authors also seems to have used raw gen. pub. crash counts, instead of total vehicles INVOLVED in crashes. Better benchmark is around 50-100k miles b/w crashes.

On research careers in academia and industry The epilogue to a series on Cognitive Science and AI

Should you go to academia or industry for research in AI or cognitive science? It's the most common question I get asked by PhD students, and I've written up some of my thoughts on the answer, as an epilogue to my research-focused series on these fields: infinitefaculty.substack.com/p/on-researc...

New Bridgerton season incoming, cancel all my plans for next weekend

a young boy with blood on his face says you can 't insult me ! Alt: Geoffrey Baratheon says "you can't insult me!"

"I no longer feel an obligation to think purely of peace."

by submitting a different manuscript 6 months later

Heard some sad news (though not the saddest kind of news) about a fellow bipolar person. A reminder to Always Take Your Meds — or make an appointment with a doctor, stat, if you’re not treated yet. I mean it, do it.

Interesting!

Not that I know of, I’m only speculating

Do we have a trolley problem with chatbots and self-harm? Say we could show that for 20 people they successfully pull away from suicide, there is one person they push toward it. Would we be utilitarian, or would we draw a line and say that this is unacceptable regardless of net impact?

We Still Don’t Know if Robotaxis Are Safer Than Human Drivers And even if self-driving technology proves to be less dangerous, there are many better ways to improve traffic safety and prevent fatal crashes.

Sorry, but triumphant claims about autonomous vehicle safety are wildly exaggerated.

It's an open question whether today’s self-driving cars are any safer than those driven by humans.

And if reducing crashes is the goal, that isn’t even the right question.

My deep dive, in Bloomberg 🧵