Sign in

Arthur Lapraye

@alapraye.eurosky.social
127 followers 76 following 335 posts

Linguiste. he/him/hän pp by @tailsteak.bsky.social

PostsRepliesMedia
Reposted by Arthur Lapraye
william light @wrl.sh · 13h
ok so I'm hearing both that I don't know how good ai is now because I refuse to use it and that invalidates my critiques but also apparently I'm a fake critic because I inadvertently use ai even when I'd rather not because the providers are shoving it anywhere and everywhere which is it?
1486
Reposted by Arthur Lapraye
Paweł Ausir Dembowski @ausir.bsky.social · 2h
btw interesting sign language fact: Spanish Sign Language has no official status in Spain, but Catalan Sign Language is an official language of Catalonia gotta give it to the Catalans that they're good with minority language rights also within Catalonia.
2164
Arthur Lapraye @alapraye.eurosky.social · 12h
Lu : Quand les Indiens parlaient latin, 2023, Serge Gruzinski, livre qui traite de l'évangélisation des élites Nahua (= "Aztèques" / Mexica), accompagnée de leur alphabétisation en castillan, en latin et dans leurs propres langues.
120
Reposted by Arthur Lapraye
jacobian term @evpok.love · 03/10/2026
j'avais pas idée que ça existait mais ça déchire ! enfin une alternative à Google ngrams!
122
Reposted by Arthur Lapraye
Dr Abeba Birhane @abeba.blacksky.app · 03/10/2026
👇🏾
09916
Reposted by Arthur Lapraye
Peter Eibich @peibich.bsky.social · 02/10/2026
I get that we often think of survey data mainly as data and having more data at lower cost seems great, but at a very fundamental level we run surveys because we want to know what people think. If we don't care enough about people's views to ask actual people, why bother at all?
13814
Reposted by Arthur Lapraye
Disappointed Optimist @disappoptimism.bsky.social · 02/10/2026
Actual LOL
312340
Reposted by Arthur Lapraye
Hisham Zerriffi @hishamzerriffi.bsky.social · 02/10/2026
Pew tested AI synthetic surveys against actual survey data. Conclusion: “This experiment taught us that, at this time, AI models are not an adequate replacement for traditional polling on topics of broad public importance.” This is an all time bad idea. No need to hedge with “at this time.”
pewresearch.org
Can AI Stand In for Human Survey-Takers? Not Really
Synthetic surveys using digital twins do a poor job estimating the public mood and are especially error-prone for some subgroups, including Republicans and Black adults.
1023679
Reposted by Arthur Lapraye
Janne M. Korhonen @jmkorhonen.fi · 02/10/2026
Our tactical encryption remains undefeated, even by most modern Microsoft clankers
A meme picture of Terminator robot searching for a crying anime girl hiding below an office table. The Terminator is labelled "Finnish language," the girl "Machine translation."
1832965
Reposted by Arthur Lapraye
La Alice, lai h'line deusporiante @marwaue.northsky.social · 02/10/2026
C'est fou comment a 14 ans on peut être un insurrectionniste qui menace la sécurité de l'état en demandant du PQ aux toilettes et a 17 ans être un grand bébé chéri d'amour innocent qui se rendait pas compte quand on tiens des propos antisemites
1663306
Reposted by Arthur Lapraye
Lisa Messeri @lmesseri.bsky.social · 01/10/2026
one of the reasons that I warn researchers away from adapting commercial LLMs into their pipeline is that Big Tech has a history of sunsetting popular, usable products and replacing them with worse ones. (i'm trying to ween myself off of google scholar, but it is hard).
1021779
Reposted by Arthur Lapraye
Olivia Guest · Ολίβια Γκεστ @olivia.science · 28/09/2026
I feel this in my bones. We're mourning the death of a career many of us thought we could have, while witnessing our chronic illnesses and disabilities becoming more debilitating. The networks of mutual support that suspended us in meaningful relationships with the rest of society have come undone.
422948
Reposted by Arthur Lapraye
Onboard Navvy @onboard-navvy.bsky.social · 05/07/2026
There's a whole world of linguistics and philosophy, from Austin onwards, from pragmatism really, that so-called LLMs just blow past: Grice and after, Relevance Theory, Politeness Theory, CA: basically the point of language, the important stuff. All smooshed into a planet-burning autocompletion.
3205
Reposted by Arthur Lapraye
Pookleblinky @pookleblinky.bsky.social · 26/09/2026
AI detection software, made out of genAI by the same dipshits making genAI, is being targeted at Black writers so that an Arday can happen but with no human able to be held accountable.
2443186
Reposted by Arthur Lapraye
Arthur Lapraye @alapraye.eurosky.social · 28/01/2026
Je suis ulcéré par tout ce que je vois passer autour des usages de LLM pour rédiger des articles scientifiques ou les évaluer. Je savais qu'il y avait des crétins méprisables dans le monde académique mais j'ai visiblement sous-estimé leur prévalence et leur propension à la démission intellectuelle.
111
Reposted by Arthur Lapraye
Arthur Lapraye @alapraye.eurosky.social · 21/08/2026
J'attends la même indignation envers les professeurs d'université qui utilisent des LLMs que celle qui a frappé Jason Arday suite à des accusation de plagiat. Enfin je l'attendrais si la motivation affichée avait eu quoi que ce soit à voir avec la motivation réelle de cette indignation.
121
Reposted by Arthur Lapraye
Médéric Gasquet-Cyrus @medericgc.bsky.social · 26/09/2026
L’affaire Thélyson Orélien est complexe et pue clairement le racisme, même si ce n’est pas le seul angle avec lequel on peut la traiter. Le texte de Mohamed Mbougar Sarr (Prix Goncourt 2021) est très fin et nous invite à penser au-delà de ce simple texte.
05821
Reposted by Arthur Lapraye
Médéric Gasquet-Cyrus @medericgc.bsky.social · 26/09/2026
J’ai sous les yeux une énième fausse référence de l’IA de Google, une fausse citation facilement vérifiable mais présentée comme authentique. Ce sont les sociétés qui fabriquent les IA que l’on devrait d’abord attaquer et condamner, pour diffuser chaque jour des millions d’informations fausses.
26111
Reposted by Arthur Lapraye
Denis Colombi @uneheuredepeine.bsky.social · 26/09/2026
On en est à un point où les jeunes ne protestent pas pour avoir une meilleure société mais juste pour avoir des toilettes acceptables.
928997
Reposted by Arthur Lapraye
David Ho @davidho.bsky.social · 23/09/2026
Wtaf is this question my university just asked me about my project?
Does the project involve any work that could be interpreted as promoting DEIA, climate change research, or non-compliant to any Presidential Executive Orders?
15145891298
Reposted by Arthur Lapraye
Arthur Lapraye @alapraye.eurosky.social · 17/11/2025
Je souhaite à tous les Pétainistes d'hier, d'aujourd'hui, et de demain, de finir comme Philippe Henriot.
001
Arthur Lapraye @alapraye.eurosky.social · 25/09/2026
I don't think it's very recent and I don't think it has replaced using conjugated forms so much as it coexists with it. One important sociolinguistic variable to take into account might also be that a lot of programmers in f French-speaking context are not L1 French speaker.
130
Reposted by Arthur Lapraye
LittleKuriboh @littlekuriboh.bsky.social · 24/09/2026
The second half of the Two Cake Rule tumblr meme comic, simply depicting the audience staring at two cakes - instead of saying “Holy shit, two cakes!” they are now saying “Holy shit! Cakes!”
171765309
Reposted by Arthur Lapraye
Arthur Lapraye @alapraye.eurosky.social · 19/07/2026
I have a master's degree in linguistics and natural language processing and six years of experience as a professional consultant in NLP and I remain thoroughly unconvinced that "natural language" is a good way to interact with a computer.
132
Reposted by Arthur Lapraye
Hunter Walker @hunterw.bsky.social · 23/09/2026
Wow. Sounds like we should just shut down these companies and stop using it?
341011169
Reposted by Arthur Lapraye
Alex Hanna @alexhanna.bsky.social · 23/09/2026
“AI” IS impacting labor, but not in the ways you think. In some cases, people generating “workslop” are making their colleagues’ jobs more difficult. Snippet from my video Myths of AI and Productivity, in which we at dispel 5 common myths around AI and labor. www.youtube.com/watch?v=3BO0...
There are three stacked screenshots from a video of Alex, a brown trans woman, sitting on a pink chair and speaking into the camera while holding a small microphone. She has short black curly hair and is wearing a blue long-sleeved shirt and light tan pants. Behind her is a cat-shaped pillow with an image of a tuxedo cat on it, and behind her chair are rows of books on a light blue wall. Underneath are captions from the video that say: “But generative AI can also reduce wages in other ways, namely by punishing workers who use the tools and pass on ‘workslop’ to their colleagues.”, “...researchers report that receiving generative AI text or images from colleagues results in lower perceptions of the competence of their colleagues.” and “Getting slop from your colleague makes you think worse of them. And if those workers are women or older, or people of color, then it confirms their co-workers’ biases against them.”
713941
Arthur Lapraye @alapraye.eurosky.social · 23/09/2026
Non seulement Andrew Wakefield et ses épigones style RFK Jr sont des criminels contre l'humanité mais il faudrait inventer une nouvelle catégorie juridique de crime contre l'humanité rien que pour eux. Le militantisme antivaccinal est la cause politique la plus meurtrière et dangereuse du siècle...
100
Reposted by Arthur Lapraye
Dr Abeba Birhane @abeba.blacksky.app · 21/09/2026
i personally got so much shit for signing my name on a letter asking the Irish government to recuse itself from fiscal files during its presidency of the EU but i see the Irish government is now saying the quite part out loud bsky.app/profile/abeb...
218333
Reposted by Arthur Lapraye
Dr Abeba Birhane @abeba.blacksky.app · 21/09/2026
"In a leaked document, the Irish Presidency proposes to EU Member States that the interest of OpenAI, Anthropic, Google, Meta and SpaceX to make profits should take precedence over the fundamental right to data protection" noyb.eu/en/ai-eu-mem...
noyb.eu
AI: EU Member States plan “digital expropriation” of Europeans in the interest of AI companies
In a leaked document, the Irish Presidency proposes to EU Member States that the interest of AI companies to make profits should take precedence over the fundamental rights of Europeans.
19466294
Reposted by Arthur Lapraye
Dr. Lindsay Stallones Marshall (she/her) @lindstorian.bsky.social · 19/09/2026
People designed genAI so it always replicates our existing injustices. To take an already underrepresented group & just replace them w/ chatbots trained on centuries of anti-Indigenous writing, esp when reporting immediately affects their communities, is revolting. There MUST be real consequences.
05525
Reposted by Arthur Lapraye
Kat is ✨phenomena✨ @kathryndeplume.com · 19/09/2026
Yup. I remain wholly convinced that the calls to slow down (and the recent "hacks") are nothing more than a smokescreen to obscure the fact that this is as advanced as the technology gets, and keep convincing people they're capable of more, thus ultimately keep the investment money rolling in.
1296
Arthur Lapraye @alapraye.eurosky.social · 19/09/2026
Si vous voulez tester si un chatbot a vraiment une intelligence humaine, dites lui des trucs délibérément incohérents et vous verrez assez vite l'illusion s'effondrer.
000
Reposted by Arthur Lapraye
Shannon Vallor @shannonvallor.bsky.social · 16/09/2026
oh so he’s a comedian now
BBC headline with photo of Sam Altman: “OpenAI boss says world ‘right to be afraid’ but should trust AI firms”
410319
Reposted by Arthur Lapraye
Arthur Lapraye @alapraye.eurosky.social · 15/09/2026
C'est infernal d'étudier des langues avec un prof quand on est linguiste parce qu'on meurt d'envie de faire des comparaisons et des étymologies dans tous les sens et eux ils insistent pour t'apprendre comment "parler la langue" avec "des vrais gens".
132
Reposted by Arthur Lapraye
Asheesh Kapur Siddique @asheeshks.org · 15/09/2026
Colleges could subsidize books for students rather than paying for premium versions of ChatGPT
6409101
Arthur Lapraye @alapraye.eurosky.social · 13/09/2026
With a teacher : English, German, Russian, Finnish, Hungarian On my own : too many to list but generally not for very long
000
Reposted by Arthur Lapraye
Intenudas @izmarasan.bsky.social · 10/09/2026
A sahelian word > Meaning : Evil eye, curse #Tamajeq : Tagaršak #Zenaga : Tgaršəgt #Songhay : Garsaka #Fulani : Garsa But Tetserret has 'təḍ' which a calque on arabic العين
272
Reposted by Arthur Lapraye
ddɐ˙ʎʞsʞɔɐlq˙uǝʌs@ @sven.blacksky.app · 09/09/2026
Caveat: If we don't all die within the next 10 (ten) years because of rogue AI.
0194
Reposted by Arthur Lapraye
Dawnlight🌞Neni / ART COMMISSIONS OPEN!! @nenilein.com · 09/09/2026
Fun fact: "Reviewing an AI made translation" is more work for translators than just translating it themselves, because AI translations do not pass the sniff test beyond the most superficial. It's a loophole companies use as an excuse to underpay translators, and nothing else.
3439631817
Reposted by Arthur Lapraye
ʤan~ʤɑn~ʤɔn @languagedoodad.bsky.social · 10/09/2026
the security issues with LLMs are not new. Agents are just calls to a server. If there's a rogue agent that just means there's intentionally lax network security protocols. These models might be new, but how they exist in the world—the network computing infrastructure—is not.
281
Reposted by Arthur Lapraye
Anthony Moser @anthonymoser.com · 09/09/2026
"We need to stop burning things for fuel before the carbon buildup makes the planet unliveable for humans" "What if we could just take carbon *out* of the atmosphere?" "Great I didn't really want to change"
1142
Reposted by Arthur Lapraye
Stefan Müller @stefanmuelller.lingo.lol.ap.brid.gy · 08/09/2026
1/ Well, „She sneezed the foam off the cappucino“ may seem unusual/unheard for non-linguists, but it is the standard example in the Construction Grammar literature and of course AI models read about it. So, no surprise that ChatGPT could say something about it.
134
Arthur Lapraye @alapraye.eurosky.social · 08/09/2026
The generativisto-generativist school of linguistics holds that a defining feature of the faculty of language is the possibility of creating infinitely many new sentences about language ("infinite metalinguistic property").
100
Arthur Lapraye @alapraye.eurosky.social · 08/09/2026
L'occasion de reposter l'article de R. Blust sur le changement T ~ K dans les langues polynésiennes
researchgate.net
(PDF) *t to k: An Austronesian Sound Change Revisited
PDF | Although the change of *t to k in Hawaiian has been known and commented on for over 150 years, the widespread driftlike character of this... | Find, read and cite all the research you need on Re...
010
Reposted by Arthur Lapraye
Melanie Mitchell @melaniemitchell.bsky.social · 08/09/2026
Hey @ericlipton.nytimes.com 👋 AI is not “evolving” on its own. *People* are choosing how to build it. *People* are choosing how to train it. *People* are choosing how to run it. Maybe the problem is that they are the *wrong people* to make these choices.
28635174
Reposted by Arthur Lapraye
Eagle Fandom BNF ☕️🦬 @trenchologist.bsky.social · 07/09/2026
Right. Like, MAHA isn’t a push for equitable access to care. It’s eugenics under the guise of “healthy living.” There’s plenty of somewhat normalizing woo and almond mom and ~natural~ in the mix, but that’s just noise. The signal is the strongest survive, and no regard for the lessers who don’t.
151188215
Reposted by Arthur Lapraye
Jeff Doctor @jeff.doctor · 27/08/2026
I've been staying off Bluesky lately because I'm on a deadline for a project examining the history of colonial map making in the Great Lakes region, so image my reaction when one of the first posts I see when I come back is MapQuest trolling the President of the United States
328165
Reposted by Arthur Lapraye
Timnit Gebru @timnitgebru.blacksky.app · 07/09/2026
"Like a number of recent AI breakthroughs, it pasted together ideas from the mathematical literature to build a new theorem. Once again, the LLM’s trick is its superhuman patience for assembling puzzle pieces, not the ability to make some profound intellectual leap."
233758
Reposted by Arthur Lapraye
Marie-Lan Taÿ Pamart @jastrow75.bsky.social · 07/09/2026
Il faut lire les détails, vraiment : la tuerie méthodique (une pelleteuse achetée spécialement pour l'enfouissement !) d'espèces protégées, mais jugées nuisibles au gibier, incluant le Faucon crécerelle ou encore la Grande Aigrette
21394315
Reposted by Arthur Lapraye
Emily M. Bender @emilymbender.bsky.social · 07/09/2026
When Sam Altman, Geoff Hinton, etc or people quoting them say "AI" can do your job, they are making an unfounded claim: they don't have the expertise in what it is to do your job. >>
123749