Romain Lesur
LightNews LightNews
rlesur.bsky.social
Romain Lesur
@rlesur.bsky.social
260 followers 1.8K following 7 posts
Head of data science lab @ insee.fr | Interested in data stuff
Posts Replies Media Videos
Reposted by Romain Lesur
anildash.com
Okay, for the folks who asked: here's the majority AI view, writing up the reasonable, thoughtful view on AI that the vast majority of people in tech hold, that gets overshadowed by the bluster and hype of the tycoons trying to shill their nonsense. anildash.com/2025/10/17/t... Please share!
The Majority AI View - Anil Dash
A blog about making culture. Since 1999.
anildash.com
October 17, 2025 at 7:29 PM
Reposted by Romain Lesur
cscheid.net
Quarto friends! I need your help:

We are implementing a new QMD parser in Quarto. It will be super nice. But it's a big change, and we want to minimize the impact.

1. Can you share a link to your Quarto project at github.com/quarto-dev/q...
2. repost this and let your Quarto friends know too?
Epic: are we production yet · Issue #63 · quarto-dev/quarto-markdown
We need to check against many large sites to get a good sense for the impact of this new syntax in practice. autogenerated qmd quartodoc-generated sites (tbd meet with @machow) sites quarto.org Shi...
github.com
October 16, 2025 at 2:13 PM
Reposted by Romain Lesur
olivia.science
Finally! 🤩 Our position piece: Against the Uncritical Adoption of 'AI' Technologies in Academia:
doi.org/10.5281/zeno...

We unpick the tech industry’s marketing, hype, & harm; and we argue for safeguarding higher education, critical
thinking, expertise, academic freedom, & scientific integrity.
1/n
Abstract: Under the banner of progress, products have been uncritically adopted or even imposed on users — in past centuries with tobacco and combustion engines, and in the 21st with social media. For these collective blunders, we now regret our involvement or apathy as scientists, and society struggles to put the genie back in the bottle. Currently, we are similarly entangled with artificial intelligence (AI) technology. For example, software updates are rolled out seamlessly and non-consensually, Microsoft Office is bundled with chatbots, and we, our students, and our employers have had no say, as it is not considered a valid position to reject AI technologies in our teaching and research. This is why in June 2025, we co-authored an Open Letter calling on our employers to reverse and rethink their stance on uncritically adopting AI technologies. In this position piece, we expound on why universities must take their role seriously toa) counter the technology industry’s marketing, hype, and harm; and to b) safeguard higher education, critical thinking, expertise, academic freedom, and scientific integrity. We include pointers to relevant work to further inform our colleagues. Figure 1. A cartoon set theoretic view on various terms (see Table 1) used when discussing the superset AI (black outline, hatched background): LLMs are in orange; ANNs are in magenta; generative models are in blue; and finally, chatbots are in green. Where these intersect, the colours reflect that, e.g. generative adversarial network (GAN) and Boltzmann machine (BM) models are in the purple subset because they are both generative and ANNs. In the case of proprietary closed source models, e.g. OpenAI’s ChatGPT and Apple’s Siri, we cannot verify their implementation and so academics can only make educated guesses (cf. Dingemanse 2025). Undefined terms used above: BERT (Devlin et al. 2019); AlexNet (Krizhevsky et al. 2017); A.L.I.C.E. (Wallace 2009); ELIZA (Weizenbaum 1966); Jabberwacky (Twist 2003); linear discriminant analysis (LDA); quadratic discriminant analysis (QDA). Table 1. Below some of the typical terminological disarray is untangled. Importantly, none of these terms are orthogonal nor do they exclusively pick out the types of products we may wish to critique or proscribe. Protecting the Ecosystem of Human Knowledge: Five Principles
September 6, 2025 at 8:13 AM
Reposted by Romain Lesur
anthonymoser.com
I considered writing a long carefully constructed argument laying out the harms and limitations of AI, but instead I wrote about being a hater. Only humans can be haters.
I Am An AI Hater
I am an AI hater. This is considered rude, but I do not care, because I am a hater.
anthonymoser.github.io
August 27, 2025 at 5:04 PM
Reposted by Romain Lesur
ketanjoshi.co
Good thing no one uses Microsoft Excel for anything related to legal, regulatory or compliance business functions

www.theverge.com/news/761338/...
Microsoft Excel adds Copilot Al to help ... theverge.com The Verget-4.1-mini Al model | 5 successor to the LABS.GENERATIVEAI function Microsoft started experimenting with in 2023. Microsoft notes that you can combine its new Al function with other Excel functions, including IF, SWITCH, LAMBDA, or WRAPROWS. The company adds that information sent through Excel's COPILOT function is "never" used for AI training, as "the input remains confidential and is used solely to generate your requested output." The COPILOT function comes with a couple of limitations, as it can't access information outside your spreadsheet, and you can only use it to calculate 100 functions every 10 minutes. Microsoft also warns against using the AI function for numerical calculations or in “high-stakes scenarios” with legal, regulatory, and compliance implications, as COPILOT "can give incorrect responses." Copy Share Select all Web search Dictionary ...
August 19, 2025 at 5:50 PM
Reposted by Romain Lesur
wcareri.com
Introducing VizDex!

An ever-growing library of personal and independent blogs and newsletters dedicated to data visualization.

vizdexproject.com
VizDex
A library of personal and independent blogs and newsletters dedicated to data visualization.
vizdexproject.com
June 20, 2025 at 3:31 PM
Reposted by Romain Lesur
jondreed.bsky.social
The AI fatigue is real, from @katiecollins.bsky.social.

www.cnet.com/tech/service...
AI Fatigue Is Wearing Me Down. The Hype Obscures What We Really Need to Know
Commentary: It's not just you -- the AI onslaught is endless and exhausting.
www.cnet.com
April 2, 2025 at 6:22 PM
Reposted by Romain Lesur
palaisau.bsky.social
« Comment fait-on comprendre au grand public que lorsque vous commencez à toucher les droits fondamentaux de certains, cela concerne, en réalité, les droits de toute la population ? »
« Il ne faudrait pas découvrir la valeur de l’Etat de droit une fois perdu » : l’alerte de hauts magistrats français
Dans des entretiens au « Monde », des hauts magistrats du Conseil d’Etat et de la Cour de cassation témoignent de leurs inquiétudes face aux attaques contre les principes juridiques mis en place en Eu...
www.lemonde.fr
March 7, 2025 at 5:57 PM
Reposted by Romain Lesur
insee.fr
L’Insee est désormais sur Bluesky ! Dorénavant, vous pouvez également retrouver nos publications en suivant notre compte @insee.fr. Abonnez-vous !
February 26, 2025 at 11:28 AM
Reposted by Romain Lesur
beadmomsw.bsky.social
12 years. RIP my darling boy.
January 11, 2025 at 5:05 PM
rlesur.bsky.social
A fascinating article whose conclusions could certainly be transposed to official statistics
juliaangwin.com Julia Angwin @juliaangwin.com · Dec 9
The key finding of my comparison of how journalists & creators build trust with audiences:

Journalism has placed many markers of trust in institutional processes that are opaque to audiences, while creators try to embed the markers of trust directly in their interactions with audiences.
shorensteinctr.bsky.social Shorenstein Center on Media, Politics and Public Policy @shorensteinctr.bsky.social · Dec 9
With trust in traditional media at an all-time low, a new paper from Shorenstein Fellow @juliaangwin.com examines what content creators — currently engaged in journalist-style work & courting large, loyal audiences — can teach journalists about trustworthiness loom.ly/90Ja7CM
December 25, 2024 at 8:27 AM
Reposted by Romain Lesur
christophmolnar.bsky.social
No one can explain stochastic gradient descent better than this panda.
a panda bear is rolling around in the grass in a zoo enclosure .
Alt: a panda bear is rolling around in the grass in a zoo enclosure .
media.tenor.com
November 24, 2024 at 3:04 PM
Reposted by Romain Lesur
darraghmurray.bsky.social
As a data and analytics person embedded in the business, I’m often involved in robust discussions with IT over access to software & data. It can be frustrating but @deanmarchiori.bsky.social has some great tips on dealing with this conundrum

#dataBS #dataSky

www.deanmarchiori.com/posts/2024-1...
Dean Marchiori - 5 tips for dealing with IT
Getting data science done with your IT team
www.deanmarchiori.com
November 19, 2024 at 10:11 AM
Reposted by Romain Lesur
adambrosiomd.bsky.social
Has anyone built an integration of #llm in #quarto documents such as a function that reads all the previous generated text, some data structure of plot and some instructions to then outputs a text to be rendered? @quarto.org @posit.co @hadleywickham.bsky.social
November 17, 2024 at 8:07 AM
rlesur.bsky.social
Dear data folks, what differences do you see between data cleaning as a data science domain and data editing as an official statistics domain? The objectives are the same. Yet data editing has been an active area of research for decades. The two communities seem to ignore each other.
November 17, 2024 at 7:15 AM