www.theguardian.com/technology/2...
11.03.2026 06:58
👍 5
🔁 3
💬 0
📌 0
Looking forward to this on Monday!
Our translation slam is coming soon – the French original (source text) is now available! itinortheast.wordpress.com/events/
No knowledge of French is required, as we'll be discussing the translations into English #ITICommunity
First time I've found useful material on TikTok when researching a translation! FYI I searched the web, I don't have the app.
Merci, @manonbrilcuah.bsky.social !
Il n'y a plus qu'à traduire / Now to translate the term 'formariage'...
Automatic transcription of the day: Jean Genet des Ursins (should be Jean Juvénal des Ursins).
AI prediction? Surreal image
First video call on the French government's own platform... which thinks I'm German so gives me all the instructions in German. Did a brief double take when I realised... mais bon 🤷♀️ Tout s'est bien passé
SURVEY:
➡️Help our member Tinka Stoessel (University of Geneva) and join her study about sustainability in the interlingual subtitling of Netflix projects targeting German as the target language🙏
(Survey to be filled out in ENG/DE)
The Department of Languages at @heriotwattuni.bsky.social is facing serious cuts with a proposal to make 11 of its staff redundant, stop offering undergraduate Chinese and German and eliminate intercultural studies courses.
Please sign this petition: heriotwatt.web.ucu.org.uk/2026/02/05/s...
How could I forget to post *this* banger, not in English? @sarahpybus.bsky.social
Peter Fox, Alles Neu youtu.be/qdtLCfEcPL4?...
English conversation Monday, Russian Tuesday, Digital skills Tuesday, German Friday, Spanish Saturday, Games à la carte Sunday
Have you visited the Kittiwake Trust Multilingual Library yet? There are so many reasons to visit! This week's lineup:
I was aware of him through visits to Tanzania in the 90s. I have seen him in concert twice. Amazing energy (and messages)! Gone too soon
Post a banger that isn't in English #kiswahili
youtu.be/ftnccb-AOik?...
All welcome to our translation slam in Newcastle on 19 March!
Dragon of the day: Digital workflows and insufficiency technologies
The idea that aggregating all of the information on the internet leads to intelligence is being put to the test. LLMs are the perfect example of the computer science adage "Garbage In, Garbage Out".
An NYT headline reads “Federal Officers Shoot Person in Minneapolis”
It has gotten so bad that the @nytimes.com has stopped using the passive voice to describe the violence.
I know I'm late to the party, but here's a #ShoutOutForGerman! I'm glad I learned the language and think more people should be speaking it. Take that, Trump! (Thanks @annatwright.bsky.social for showing me this post #ITICommunity)
Amazing sob story: "ChatGPT deleted all the work I hadn't done"
www.nature.com/articles/d41...
Update! For those of you who are coming tonight, the Head of Steam has informed us that the kitchen will be closed - their chef is off sick.
If you were planning to order food, make sure you eat first or grab something on the way!
One of the most insightful political cartoonists of modern times.
Happy New Year! Who's coming along to our get-together at The Head of Steam in Durham on Wednesday 14th? See you from 6.30pm
termine.di.day #DIDit #DUTgemacht
Ohne Unabhängigkeit keine Freiheit!
Heute ist der erste Digital Independence Day (DID)! An jedem ersten Sonntag im Monat wollen wir ab jetzt aus einem Big-Tech-Gefängnis ausbrechen – und natürlich auch anderen dabei helfen. #DIDit #DUTgemacht
Merry Christmas! Coincidentally just visited the edge of Dartmoor and can confirm it was pretty darn chilly
Dragon of the day: 'X when renting' rather than 'experimenting'
Dragon of the day: 'wee chap' instead of 'WeChat' #amtranslating
I'm not saying proofreading is a bad idea!
I am concerned that people will e.g. scan something in a language they don't speak, or in their own language but won't check it, then translate it using AI as well. Thus multiplying the possibilities for misunderstandings...
I mean, the OCR software I have used in the past wasn't perfect, but the errors were usually obvious. The errors here would have been easy to overlook, and if you had given the text to someone without the original they would never have spotted them!
I tried asking Copilot to transcribe text from an image file the other day, and... maybe 85% was correct but some text towards the end seemed plausible but *wasn't the text in the image*.
I probably need to know more about good tools/prompts to use (possibly esp if prompting in a different lang.)
