Antoine Vernet
@antoinevernet.com
Associate Professor in Management at UCL. Networks and Organizational Design. Mostly found reading. Very occasionally on Youtube: www.youtube.com/@antoinevernet
I find arguments that AI can't do judgement or creativity or taste to be especially obviously false in the time of agents. Any long task requires lots of taste, judgement & creativity. I am much more sympathetic to arguments about the quality or diversity of the AI's taste, judgement, or creativity
And people tell me that I am too cautious putting everything on github
Demetri@phdemetri.bsky.social · 3h agoLMFAO
If you made me choose between Claude Code and the rest of the internet, I'd have to think about it
Systems that do not have regulatory barriers or gatekeepers to AI are rapidly getting overwhelmed / transformed: mathematics, paper writing, code, etc. For everything else, it'll be slower but it makes sense to start planning now!
7 months ago I read about gas town and today I read the new thing. what worries me now is either he’s gotten less insane or I’ve gotten more insane.
I agree that coding agents have provided a welcome relief from tiny frustrations from problems that were never big enough to fix before
Ethan Mollick@emollick.bsky.social · 2d agoAn unexpectedly stress-reducing use of Codex/Code is just fixing problems with my various Windows machines: weird driver issues, game incompatibilities, even just tiny stuff that used to annoy me (why does the program that I set to run at startup not run at startup?). Hours saved. Annoying hours.
annoying codex thing last week they just dropped luna prices by 80% but you can't actually make luna subagents from sol or terra fr fr Luna is marked as "V1" and Sol & Terra are marked as supporting "V2" subagent API, and there's a Codex check preventing you from mixing them
I'm so happy to announce version 2.0.0 of my #Rstats package WeightIt is out on CRAN! New features: censoring weights, multilevel propensity scores, improved weights for continuous treatments, bias-reduced ordinal and multinomial models, M-estimation in subgroups Check out the website below!
Weighting for Covariate Balance in Observational Studies
Generates balancing weights for causal effect estimation in observational studies with binary, multi-category, or continuous point or longitudinal treatments by easing and extending the functionality ...
ngreifer.github.io
I'm back with some more #slidecrafting content, specifically how to style your footer, and adding custom footers on the side and top of the slide slidecrafting-book.com/elements#wor... #quarto
7 Elements – Slidecrafting
Tips for styling slide elements in Quarto reveal.js: custom menu icons, logo placement and sizing, footers and custom banners, transparent plot backgrounds, and controlling figure sizing with…
slidecrafting-book.com
"I hate data centers" is such a weird take. It's like "I hate injection molding facilities". I just don't have that strong of feelings about light industrial buildings.
In our information ecosystem accuracy is a liability that constrains your ability to garner attention much more than it rewards you for being trustworthy.
Scott McGrath@smcgrath.phd · 2d ago🧪 Medical professionals are losing the attention war on YouTube. In a 3,958-video sample, independent creators produced the lowest quality content but captured the highest viewership using exaggeration. #MedSky
Alibaba’s Qwen3.8-Max is out, a frontier model they say is competitive with Anthropic’s Fable. It’s a 2.4T parameter model with a 1M context window and will open source the weights next week. Following so shortly after the release of Kimi K3, it’s hard to claim China is months behind the US in AI.
Qwen Studio
Qwen Studio offers comprehensive functionality spanning chatbot, image and video understanding, image generation, document processing, web search integration, tool utilization, and artifacts.
qwen.ai
This. Absolutely this! The bar for what counts as a "publishable paper" has risen considerably, and will keep rising. I was joking to a coauthor this week that, David Card's displeasure notwithstanding, the average WP was going to get longer. We will become a book field--I'm not joking.
Clément Canonne@ccanonne.github.io · 4d ago🧵To juniors: - Move away from LPUs, at all costs. (Nobody likes to think of their own paper as LPU: what I mean here is "a paper just good enough to be published in a fairly good conf even a year ago") These publications are most likely going to die, or not being valued any more, as flood-prone 2/
Using LLMs should allow us to think bigger, harder to grasp, difficult to execute, longer to develop, ideas. That idea that’s been in you brain since forever but you hadn’t had the time because you had to retool and what not. Trust yourself. Go for that shit. Now is the fucking time.
Clément Canonne@ccanonne.github.io · 4d ago"We're all worried," as what it means to do research (in my field, Theoretical CS) seems to be shifting, and shifting fast. What to do? Senior researchers must lead by example, knowing that not everything will pan out. What I'm suggesting below may not work everywhere, but here's my own advice: 1/
Cruelty is the foundation of those 'policies'. Making us feel threatened is the purpose. We have no physical proof of our status, we are at the mercy of a glitch. We are aliens, and we must be made aware of it. Reform's 'plan' to strip us of our Leave to remain is just a darker shade of grey.
Lisa O’Carroll@lisaocarroll.bsky.social · 3d agoExclusive: Some EU citizens in the UK have been contacted by the Home Office to say they were granted settlement status after Brexit "in error". If you know anyone this happened to contact me on lisa.ocarroll@theguardian.com www.theguardian.com/politics/202...
yes. if I say 'this is my code' about AI gen stuff, that's just shorthand for 'this exists because of my decisions and i take responsibility for it' - and sensible people understand that. this ego thrashing about authorship is undignified and irrelevant. get over yourself.
atticus goldfinch@atticusgf.bsky.social · 3d agoEvery time I see a comment like the screenshot, I feel like I know _exactly_ what kind of dev said this, and would bet money they aren't particularly loved by coworkers, bosses, or clients.
One of my pet peeves: People are obsessed with seeing/plotting error bars while also being confused about what they should represent. blog.genesmindsmachines.com/p/the-error-...
The error bar cargo cult
You don't need to understand error bars. You just need to draw them.
blog.genesmindsmachines.com
🧵To seniors: - move away from bean-counting (easy, right?). E.g., when evaluating CVs or grants or applicants, ask for the top 3 pubs, with a paragraph explaining why these are important. Hiring on a CV of 50 pubs was always sort of meaningless, now it's become meaningless-er (!) and gameable. 2/
I have read this Response to my Comment and have written a Rejoinder, showing three things: 1. The multiverse is many correlated tests. 2. The randomization inference is wrong. 3. The trichotomous moderator reaffirms my point. I also enumerate 7 misleading claims, now in print in the APSR.
Cambridge University Press Political Science & IR@cambup-polsci.cambridge.org · last wk.#OpenAccess from @apsrjournal.bsky.social - Still Instrumentally Inclusive - https://cup.org/4bhrhsp "Do individuals in Western democracies shift their views on LGBT+ inclusion when exposed to opposition from Muslim out-groups?" - @turnbulldugarte.com & @bertous.bsky.social #FirstView
There's lots of stuff LLMs can't do, but if you want to disbelieve all the recent math, you have to believe a bunch of Fields Medalists with long track records are just straightforwardly lying about the thing they care most about. Here's a recent blog by Gowers: gowers.wordpress.com/2026/05/08/a...
A recent experience with ChatGPT 5.5 Pro
We are all having to keep revising upwards our assessments of the mathematical capabilities of large language models. I have just made a fairly large revision as a result of ChatGPT 5.5 Pro, to whi…
gowers.wordpress.com
OpenAI announces 10 discoveries from their next model. Observations:: 1) AI is getting very good at math 2) Two years ago LLMs failed at basic math 3) This cost less than $2000 in current API fees 4) OpenAI is focusing on announcing benefits, not just risks, of new models openai.com/index/ten-ad...
