Machine translations are often brought up as a gotcha whenever I criticize LLMs.
Uncategorized
72
Posts
59
Posters
12
Views
-
Machine translations are often brought up as a gotcha whenever I criticize LLMs. It's worth pointing out two things: Machine translations existed decades before LLMs, and yes, machine translations are useful. However: I would never in my life read a machine translated book. Understanding what a social media post is talking about in rough terms? Sure. Literature? Absolutely not. Hell, have you ever seen machine translated subtitles? It's absolute garbage.
@Gargron ironically, I feel translations have become worse with the rise of LLMs. For example, I recently saw a Microsoft KB article that had a table with Yes / No in its cells, translated to German as Ja / Nr.
-
I have the impression that primarily anglophone people don't read as much translated literature, because so much good literature already exists in their language, so this issue may not be as familiar within that demographic. As someone who did not grow up anglophone, I can tell you there is a world of difference between a good and a bad translation even when done by humans. Machine translations are not even on the scale.
@Gargron Back in my pretentious high schooler days I read Dante's Comedy and, don't know where from, the version I found was one that was like... fully translated as poetry, like in the Italian* original, going as far as trying to replicate the rhyme structure
No computer's ever going to pull off anything even remotely that mad
-
Machine translations are often brought up as a gotcha whenever I criticize LLMs. It's worth pointing out two things: Machine translations existed decades before LLMs, and yes, machine translations are useful. However: I would never in my life read a machine translated book. Understanding what a social media post is talking about in rough terms? Sure. Literature? Absolutely not. Hell, have you ever seen machine translated subtitles? It's absolute garbage.
@Gargron totalmente de acuerdo, compañero.
-
Machine translations are often brought up as a gotcha whenever I criticize LLMs. It's worth pointing out two things: Machine translations existed decades before LLMs, and yes, machine translations are useful. However: I would never in my life read a machine translated book. Understanding what a social media post is talking about in rough terms? Sure. Literature? Absolutely not. Hell, have you ever seen machine translated subtitles? It's absolute garbage.
@Gargron I've seen way too many machine-translated instructions, and that's something you really don't want to have machine-translated, for obvious reasons.
-
I have the impression that primarily anglophone people don't read as much translated literature, because so much good literature already exists in their language, so this issue may not be as familiar within that demographic. As someone who did not grow up anglophone, I can tell you there is a world of difference between a good and a bad translation even when done by humans. Machine translations are not even on the scale.
@Gargron I'm willing to guess that machine translation of prose may serve two uses: firstly, as an assist for human translators (by preparing a very rough first cut, which they then have to refine), and secondly, as an assist for human editors in figuring out which foreign-language-works to pay a human translator (with or without AI assistance) to work on (translation costs money: knowing where to spend it is important). But those are assistive roles, not human-replacing ones.
-
I have the impression that primarily anglophone people don't read as much translated literature, because so much good literature already exists in their language, so this issue may not be as familiar within that demographic. As someone who did not grow up anglophone, I can tell you there is a world of difference between a good and a bad translation even when done by humans. Machine translations are not even on the scale.
@Gargron Crunchyroll also experiments with machine translated subtitles for Japanese anime, and the community (anglophone or not) is not happy with it.
-
I have the impression that primarily anglophone people don't read as much translated literature, because so much good literature already exists in their language, so this issue may not be as familiar within that demographic. As someone who did not grow up anglophone, I can tell you there is a world of difference between a good and a bad translation even when done by humans. Machine translations are not even on the scale.
@Gargron - I think the most recent translated book series I read was Liu Cixin's Remembrance of Earth's Past trilogy,* translated into English by Ken Liu and Joel Martinsen. I don't know a word of Chinese, but I can't imagine machine translation doing these great works any justice.
* Later adapted into the Netflix series 3 Body Problem
-
I have the impression that primarily anglophone people don't read as much translated literature, because so much good literature already exists in their language, so this issue may not be as familiar within that demographic. As someone who did not grow up anglophone, I can tell you there is a world of difference between a good and a bad translation even when done by humans. Machine translations are not even on the scale.
@Gargron As someone who translates into English, I appreciate you saying this.
It's true there's a lot of very good literature in English, but I'm pretty convinced that the majority of anglophones don't know what they're missing by not reading translated books.
-
@Gargron ironically, I feel translations have become worse with the rise of LLMs. For example, I recently saw a Microsoft KB article that had a table with Yes / No in its cells, translated to German as Ja / Nr.
@inanedirk @Gargron Microsoft uses machine translations on their pages extensively, and the results are a mess. They even OCR and translate the screenshots, so you end up with screenshots that have wrong text crammed over the English original, making them doubly useless.
-
I have the impression that primarily anglophone people don't read as much translated literature, because so much good literature already exists in their language, so this issue may not be as familiar within that demographic. As someone who did not grow up anglophone, I can tell you there is a world of difference between a good and a bad translation even when done by humans. Machine translations are not even on the scale.
@Gargron plus doesn't everyone do the test of translate then translate back? They are garbage in a way that even a machine could recognize as garbage.
-
@Gargron I'm willing to guess that machine translation of prose may serve two uses: firstly, as an assist for human translators (by preparing a very rough first cut, which they then have to refine), and secondly, as an assist for human editors in figuring out which foreign-language-works to pay a human translator (with or without AI assistance) to work on (translation costs money: knowing where to spend it is important). But those are assistive roles, not human-replacing ones.
@cstross @Gargron I have a friend who worked for years as a translator (English to French) but in recent years he found that he was no longer being asked to translate but to "post-edit" machine translations. It was taking him just as long, paying him less, and destroying his soul.
He now works as a tour guide.
-
I have the impression that primarily anglophone people don't read as much translated literature, because so much good literature already exists in their language, so this issue may not be as familiar within that demographic. As someone who did not grow up anglophone, I can tell you there is a world of difference between a good and a bad translation even when done by humans. Machine translations are not even on the scale.
@Gargron I've just bought an English translation of Mikhail Bulgakov's The Master and Margarita. I wondered which was the best translation, and found a website with multiple translations of the opening page. It was really interesting how different they were. The one by Diana Burgin's and Katherine Tiernan O'Connor was the clear winner for me.
-
@Gargron thank you. When writing the German subtitles for No Other Land I had to basically do it all by hand because the context window was so narrow it got everything wrong that could be lost in translation.
@ErikUden @Gargron I work for Swiss Broadcast Company. Our devs did a wonderfull job in this regard. I get autotranslated subtitles that are amazingly good. It ain't literature but very good. It's a two tier system that joins the captions, then translation and then reconstructing the captions. Translation is done by Claude. Langs are not that big of a challange (DE FR IT EN). Only Rumantsch is a challange. Claude 3.5(!) Is pretty darn good though. Claude 4+ not so much
-
@Gargron I'm willing to guess that machine translation of prose may serve two uses: firstly, as an assist for human translators (by preparing a very rough first cut, which they then have to refine), and secondly, as an assist for human editors in figuring out which foreign-language-works to pay a human translator (with or without AI assistance) to work on (translation costs money: knowing where to spend it is important). But those are assistive roles, not human-replacing ones.
@cstross @Gargron
Machine translations are more of a hindrance than a help, for translators. If you don't know both languages well, having an automated dictionary lookup could possibly be useful - but if you're a translator, and especially a translator of fiction, having a nuanceless draft will only take more time to figure out. And it will be irritating time, because reading mistranslations is a pain. Editing one's own drafts is hard enough!As to B: Editors rely on readers, reviews /...
-
Machine translations are often brought up as a gotcha whenever I criticize LLMs. It's worth pointing out two things: Machine translations existed decades before LLMs, and yes, machine translations are useful. However: I would never in my life read a machine translated book. Understanding what a social media post is talking about in rough terms? Sure. Literature? Absolutely not. Hell, have you ever seen machine translated subtitles? It's absolute garbage.
@Gargron Yeah you most definitively have to thoroughly proof read that stuff which kinda defeats the point.
-
Machine translations are often brought up as a gotcha whenever I criticize LLMs. It's worth pointing out two things: Machine translations existed decades before LLMs, and yes, machine translations are useful. However: I would never in my life read a machine translated book. Understanding what a social media post is talking about in rough terms? Sure. Literature? Absolutely not. Hell, have you ever seen machine translated subtitles? It's absolute garbage.
Kille Bill -> 'Kill rekening' (Kill the invoice)
-
@inanedirk @Gargron Microsoft uses machine translations on their pages extensively, and the results are a mess. They even OCR and translate the screenshots, so you end up with screenshots that have wrong text crammed over the English original, making them doubly useless.
@inanedirk @Gargron Oh, and not just pages, it seems that at least parts of Windows are machine-translated nowadays, because some texts make absolutely no sense and use extremely weird word combinations.
-
@cstross @Gargron
Machine translations are more of a hindrance than a help, for translators. If you don't know both languages well, having an automated dictionary lookup could possibly be useful - but if you're a translator, and especially a translator of fiction, having a nuanceless draft will only take more time to figure out. And it will be irritating time, because reading mistranslations is a pain. Editing one's own drafts is hard enough!As to B: Editors rely on readers, reviews /...
-
@Gargron I'm willing to guess that machine translation of prose may serve two uses: firstly, as an assist for human translators (by preparing a very rough first cut, which they then have to refine), and secondly, as an assist for human editors in figuring out which foreign-language-works to pay a human translator (with or without AI assistance) to work on (translation costs money: knowing where to spend it is important). But those are assistive roles, not human-replacing ones.
-
@Gargron I'm willing to guess that machine translation of prose may serve two uses: firstly, as an assist for human translators (by preparing a very rough first cut, which they then have to refine), and secondly, as an assist for human editors in figuring out which foreign-language-works to pay a human translator (with or without AI assistance) to work on (translation costs money: knowing where to spend it is important). But those are assistive roles, not human-replacing ones.
@cstross @Gargron My (very not translator) impression is that human translators who have worked from rough machine translations, say that it’s harder than just translating the text.
Also, today I was in a work info session, where the talks were translated by some MS PoS thing, from Finnish to English. The results were horrendous, if hilarious. It might get better but I don’t really know why. Good simultaneous interpretation is kind of a human-level problem, really. Context matters!
Gli ultimi otto messaggi ricevuti dalla Federazione
-
@levelbot I feel good about myself and this number.
-
@reiver@mastodon.social knowing people in general... They will get more bug reports if they said "whom"
-
@reiver I've heard John Mastodon personally picks those out of a tall hat, for each user, individually. What a great man.
-
I'm writing this in English.
Not because English is my first language—it isn't. I'm writing this in English because if I wrote it in Korean, the people I'm addressing would run it through an outdated translator, misread it, and respond to something I never said. The responsibility for that mistranslation would fall on me. It always does.
This is the thing Eugen Rochko's post misses, despite its good intentions.
@Gargron@mastodon.social argues that LLMs are no substitute for human translators, and that people who think otherwise don't actually rely on translation. He's right about some of this. A machine-translated novel is not the same as one rendered by a skilled human translator. But the argument rests on a premise that only makes sense from a certain position: that translation is primarily about quality, about the aesthetic experience of reading literature in another language.
For many of us, translation is first about access.
The professional translation market doesn't scale to cover everything. It never has. What gets translated—and into which languages—follows the logic of cultural hegemony. Works from dominant Western languages flow outward, translated into everything. Works from East Asian languages trickle in, selectively, slowly, on someone else's schedule. The asymmetry isn't incidental; it's structural.
@Gargron@mastodon.social notes, fairly, that machine translation existed decades before LLMs. But this is only half the story, and which half matters depends entirely on which languages you're talking about. European language pairs were reasonably serviceable with older tools. Korean–English, Japanese–English, Chinese–English? Genuinely usable translation for these pairs arrived with the LLM era. Treating “machine translation” as a monolithic technology with a uniform history erases the experience of everyone whose language sits far from the Indo-European center.
There's also something uncomfortable in the framing of the button-press thought experiment: “I would erase LLMs even if it took machine translation with it.” For someone whose language has always been peripheral, that button looks very different. It's not an abstract philosophical position; it's a statement about whose access to information is expendable.
I want to be clear: none of this is an argument that LLMs are good, or that the harms @Gargron@mastodon.social describes aren't real. They are. But a critique of AI doesn't become more universal by ignoring whose languages have always been on the margins. If anything, a serious critique of AI's political economy should be more attentive to those asymmetries, not less.
The fact that I'm writing this in English, carefully, so it won't be misread—that's not incidental to my argument. That is my argument.
-
I've been checking out @inkwell , the new journaling app on the Fediverse. It looks absolutely great. Fantastic work by @stanton. If you'd like to follow me there, check out @evan@inkwell.social.
-
There is the complete list that Mastodon is suggesting I (potentially) follow.
-
Moaner's CRYING TOWELS
only 15¢ from Johnson Smith and Co (1951)
-
It's "Whom To Follow" not "Who To Follow".
I.e., WHOM not WHO.
...
Also, I wonder how Mastodon picked those choices to me to (potentially) follow.
