What on earth did they do to Deepl? The interactive translation features is suddently? shockingly bad.
Sara Horcas-Rufián
@sarahorcas.bsky.social
(Sworn) Translator. Profesora asociada (Adjunct Professor) URV & UAH. PhD student; diss. on translation teacher's competence. ella/she | sarahorcas@gmail.com
🔎 Evaluación y verificación de la información en la práctica profesional de la traducción y la interpretación. Se agradecen respuestas al siguiente cuestionario para profesionales. En defensa de la traducción humana, más que nunca. docs.google.com/forms/d/e/1F... ¡Mil gracias!
Evaluación crítica y verificación de la información en la práctica profesional de la traducción y la interpretación
La evaluación crítica y la verificación de la información son esenciales en el ejercicio profesional de la traducción y la interpretación, mucho más en estos tiempos de control algorítmico y desinform...
docs.google.com
"Sin talento humano, la IA es solo código" Gracias por admitir que es la IA la que necesita traductores y que los traductores no necesitan IA.
«Los editores quieren trabajar con traductores profesionales, no con aficionados esporádicos; sin embargo, la remuneración en pocas ocasiones permite una profesionalización real. Esta paradoja, por sí misma, refleja una situación insostenible».
Tarifas estancadas durante veinte años, pluriempleo y ahora la amenaza de la IA: la precariedad de la traducción literaria en España www.publico.es/sociedad/tar...
Trnslation agencies keep asking translators to lower their rates. The request is nothing put of the ordinary. We get them all the time, true or tale. What never ceases to amaze me is the request that vendors take the burden on themselves so the agency keeps being profitable.
Some life updates: As of October, I’ve officially transferred to URV to complete my PhD (hopefully soon!), and I’ve also started working as an Adjunct Professor at UAH. I’m really excited about both and deeply grateful to everyone who’s supported me along the way.
If experts in their fields are now using AI to read papers for them, then science has truly hit rock bottom.
Today I asked my translation students whether they thought it was fair to be paid per word when, in addition to the translation itself, the agency also profits from the translation memory (TM) that gets updated as they work. None of them thought it is normal not to be paid for updating the TM.
"AI disclosure erodes trust in the AI user." www.sciencedirect.com/science/arti...
The transparency dilemma: How AI disclosure erodes trust
As generative artificial intelligence (AI) has found its way into various work tasks, questions about whether its usage should be disclosed and the co…
sciencedirect.com
Oh yes, postediting, that task that is sold (and paid) as a (light) revision service but is overall as demanding as a translation task. It does not sound like a good deal to me.
OpenAI, one of the most prominent AI companies in the world, just released a model trained BLATANTLY on stolen IP. It's obvious that rights are being violated, and it's also obvious that existing enforcement is not preventing these violations: they press onward. The EU must enforce its rules
"Je commencerai par reconnaître ma situation de privilège : la raison la plus immédiate pour laquelle je n’utilise pas l’IA, c’est que j’exerce mon métier de traducteur dans de bonnes conditions."
Aujourd'hui, c'est la Saint-Jérôme, saint patron des traducteurices. Et ils sont un paquet à être en grève aujourd'hui pour protester contre la dégradation de nos conditions de travail, mais aussi contre l'usage de l'IA. C'est l'occasion de partager ce billet : clement-martin.fr/fr/posts/et-...
«Que los lectores sepan que existimos, que detrás de las traducciones que leen hay personas que aman los libros tanto como ellos, importa. Porque, si no se nos ve, es más fácil eliminarnos del proceso». Feliz #DíaInternacionalDeLaTraducción.
En la revista Calidoscopio, de la Asociación Argentina de Traductores e Intérpretes (AATI), han tenido la amabilidad de hacerme una entrevista. Qué maravilla aparecer en un número tan interesante 😊 aati.org.ar/novedades/bo...
Avui, Dia Internacional de la Traducció, reivindiquem la traducció humana en un sector cada vegada més automatitzat, així com unes condicions justes. Cal ser més conscients que mai que l'associacionisme ens fa més forts i ens permet fer xarxa davant les adversitats. ✊
30 sep, Día Internacional de la Traducción: traductores en huelga en protesta por las condiciones de trabajo cada vez más precarias en las que tenemos que desarrollar nuestra actividad.
Traductrice en grève ! Bonne Saint-Jérôme, les collègues ✊
#agendaURV | Inauguració del curs acadèmic a la Facultat de Lletres, a càrrec de l’escriptora Remedios Zafra, que impartirà la conferència “Los trabajos que amamos. Sobre el entusiasmo y la transformación del trabajo creativo”. 🗓️Demà, 1/10 🕚12h 📍Aula 202 campus Catalunya 👉https://goo.su/Xj0cZUz
Things I had not expected to see today: Deutsche Bank making a numbers based argument that "AI" is a bubble and won't last https://futurism.com/artificial-intelligence/deutsche-bank-grim-warning-ai-industry
Deutsche Bank Issues Grim Warning for AI Industry
Deutsche Bank is warning that AI spending won't "remain parabolic," or continue to increase exponentially, a dire warning for the future.
futurism.com
We should stop using the term "translating" when we are talking about what automatic translation systems can do. What they do is processing texts (to try and replicate it using words from a different language). Translating is a human ability.
Online Event: International Translation Day 2025, Translating Peace and Truth in Times of Global Crisis 📅30th Sept, Tues. 17h-19h BST via Zoom Free & open to all (registration required) Register here: www.tickettailor.com/events/trans...
Register now – International Translation Day 2025: Translating Peace and Truth in Times of Global Crisis – Zoom
International Translation Day 2025: Translating Peace and Truth in Times of Global Crisis – Zoom, Tue 30 Sep 2025 - Register for Online Event Series by Translation VillageInternational Translation Day...
tickettailor.com
In Spanish, the snack or bite you have between meals is tentempié. It literally means 'keeps you standing'. Isn't language diversity beautiful.
In Swedish, a word for what you eat to bridge the gap between meals (or while waiting for the main course to cook) is stödmacka. It means "support sandwich." A similar word in Norwegian is ventepølse, or "waiting sausage."
When they say ai is a time saver what that often means is that it’s a tool that makes it easier for you to waste someone else’s time.
"In a landmark study, OpenAI researchers reveal that large language models will always produce plausible but false outputs, even with perfect data, due to fundamental statistical and computational limits." www.computerworld.com/article/4059...
OpenAI admits AI hallucinations are mathematically inevitable, not just engineering flaws
In a landmark study, OpenAI researchers reveal that large language models will always produce plausible but false outputs, even with perfect data, due to fundamental statistical and computational limi...
computerworld.com
“It would be foolish to think that general-purpose AI could transform the economy with only modest political consequences,” two academics from the University of Toronto argue in a guest essay
Two scholars ask whether democracy can survive if AI does all the jobs
Without taxation there may be no representation, conclude Raymond Douglas and David Duvenaud
econ.st
We have it as a contract point that translations of my work have to be performed by actual humans, I think actual human readers deserve to have actual humans translating and making the works as excellent as they can be regardless of language
Reading so much here about skilled, deeply experienced translators being reduced to much-lower-paying machine translation post-editing (MTPE) — fixing a first-round AI translation. Reminds me of the tremendous labor + frustration of *undoing* a bad copyedit (whether performed by human or machine)
It never ceases to amaze me that every AI implementation on any platform seems to boil down to just two types of user interaction: stealing them, i.e., users are never asked for consent. forcing them to actively resist, i.e, users must actively fight to avoid being exploited. All deeply unethical.
LinkedIn va a entrenar –o ya lo está haciendo– su IA generativa con los datos de sus usuarios. Para "impedir" esto hay que ir a [configuración]>>>[privacidad de datos]>>>[datos para mejorar la IA generativa] y desactivar esta opción que está habilitada por defecto.
'What do I do first?' As we outsource more of our skills and knowledge to AI I think this is a phrase that we will be hearing people say a lot.
LiveAI demo fails on the first prompt at Meta Connect 2025. #Meta #AI #LiveAI
Job opportunity: Two PhD Positions at Stockholm University, Institute for Interpreting and Translation Studies, Sweden. Deadline for application: October 15. Link: su.varbi.com/en/what:job/...
Doktorand i översättningsvetenskap
Institutionen för svenska och flerspråkighet är en av de större institutionerna inom Humanistiska fakulteten med ca 140 anställda och ca 5 000 studenter per läsår. Tolk- och översättarinstitu
su.varbi.com