[tkgally]

Former professional translator here (Japanese-to-English, 1986 to 2005). I have very mixed feelings about MT.

On the one hand, as you and others point out, rates for human translation have been hurt, and the quality of a lot of MT-enabled translation visible to the public is mediocre to poor.

On the other hand, MT is enabling communication among people that was not possible or practical before. A small start-up is able to find customers and suppliers in countries with different languages. People with common interests but diverse languages are able to chat and share ideas on social media. A person receives an e-mail from a long-lost relative in a language the recipient doesn’t know; MT enables them to correspond and eventually meet up. Little of that communication would be taking place if a human translator had to be found and paid for each interaction.

[sinity]

> On the other hand, MT is enabling communication among people that was not possible or practical before. A small start-up is able to find customers and suppliers in countries with different languages. People with common interests but diverse languages are able to chat and share ideas on social media. A person receives an e-mail from a long-lost relative in a language the recipient doesn’t know; MT enables them to correspond and eventually meet up. Little of that communication would be taking place if a human translator had to be found and paid for each interaction.

Yep. I translated tens of KWords between English and Polish by now - thanks to Deepl, which speeds this up by at least an order of magnitude (and the end result is probably better too).

I even planned to translate a book only available in Polish ([this one](https://en.wikipedia.org/wiki/Perfect_Imperfection)) - I translated about 3% already, but unfortunately the author didn't consent to sharing the translation anywhere - which is annoying, because original (Polish) text is freely available on libgen. And English version doesn't exist. And the book is 18 years old at this point. So the potentially lost profits are negligible or nil. Meh.

[OmarAssadi]

Good points!

As an example, as a hobby, I am interested in things like anime video compression and such. Two of the most essential sites for me in terms of resources and information have been almost entirely Chinese. And while they appreciate and respect the international users, at the end of the day, they can’t translate each and every single forum post into English or any other language that someone may need.

And also unfortunately, I don’t have the time yet to invest in learning the language enough to use those sites without assistance. Instead, I rely heavily on things like Google Translate, Yandex, and DeepL. They’ve been totally essential to the point to where a major pain point for me with my preferred browser (FireFox) has been a lack of native MT support (until very recently, which has been nice for Russian and stuff, but I’m still waiting for Chinese support in Bergamot).

Similarly, while I studied in Russia for a while, I’m not a native speaker, and my skills aren’t what they used to be now that I’ve been out of the country for a while. There is a lot of interesting stuff on the Russian internet (Habr, RuTracker, various articles, etc.) that I only really have the level of convenient access to that I do because of MT.

I don’t think it’s all sunshine and rainbows, certainly, as with stuff like DALL-E 2, but it definitely provides a lot of people real value and happiness. Hopefully, humanity will figure out how best to balance the positives and negatives of AI services like these.

I see that both sides (lol, sorry) of this often take pretty hardline stances on “art is art; there is no difference” vs. “these are just copy-pasting people’s hardwork and can’t actually create anything themselves”. I think the value is sort of indisputable, but I worry about how it’ll affect people’s livelihoods — e.g., after playing with DALL-E 2, I think it is totally capable of replacing things like stock photos a large percentage of the time (not always, of course, but it definitely can sometimes).

I’m of the mindset that you should be able to remix, adapt, copy, etc, virtually everything in an ideal world. Maybe I am just entitled, I don’t see why I should have to waste the man-hours recreating or reverse engineering something simply because of a license incompatibility, for example.

Similarly, I don’t see why I shouldn’t be able to transform or straight-up copy and distribute someone else’s art if it provides people with happiness. As an example, talking about anime earlier, I think things like fansubs can provide lots of additional value to people by either providing more natural translations or through extensive typesetting that companies tend not to do because it is not “necessary”. For example, those familiar with the show Bakemonogatari are probably aware of the extreme amount of dialogue displayed exclusively visually, with rapid flashing cards and such. While I think the official subtitles don’t get enough credit 99% of the time, that is a prime example where fansubs which go in and actively mask over the text on the screen and replace it with the viewer’s native language is very helpful. However, despite the value these things provide, it is copyright infringement.

Rationally, I understand that the emotional and protective side of people exists, but at the same time, I don’t “understand” it; I want people to use my stuff — I want people to be happy with their limited time on this planet. In that sense, I think the emotional argument for copyright does not appeal to me.

That being said, pragmatically, people need jobs. We live in a society where you must work. A society where you have bills to pay, groceries to buy, and fees for schools and hospitals. Stripping away copyright protections entirely for things like this would hurt a lot of people right now. And I am really curious how this’ll all turn out in court when the majority of these datasets are trained on things that they have no right to be using.

Will we be in a world where you can essentially bypass all copyright protections by throwing everything into TensorFlow before feeding the computer a little prompt? Or will we be in a world where many of these tools are crippled to the point of uselessness? Or will we be in a world where companies just accept that sometimes someone will take them to court and settlements will just be another business expense?

Aside from the economic issues, if I’m going to be writing some Adderall-infused essay right now, I figure I may as well rant about the other two.

The first of those is biases; while I have always been aware of this problem, it was very evident after playing with DALL-E 2 this past week. The datasets seem to be very European-centric (US, Canada, NZ, AU, etc included in that). There were a number of things that I could not get it to generate correctly or would produce wildly different results based on the prompts I fed it.

As relatively minor, silly example, while I had no trouble getting it to produce white-washed movie posters for by asking it for “Netflix live-action adaptation of Dragon Ball Z”, I couldn’t get it to do the same for India and Bollywood. I am sure someone who has spent more time doing “prompt engineering” than I have could possibly get better results, but the point still stands in the sense that it seems to like to produce what Americans & Europeans are more exposed to.

Another example is, just ignorantly guessing, the way things are weighted — e.g., I wanted to have it produce art in the style of shows like Tatami Galaxy, The Night Is Short, Walk On Girl, and things like the album covers for Asian Kung-fu Generation. All of those listed are done by Yusuke Nakamura. However, perhaps rubbing salt in the wound for artists, I was unable to get the results I wanted by asking for things in Yusuke Nakamura’s style. Instead, I had much better results by using “Yuasa Masaaki”, the director of Tatami Galaxy and The Night Is Short, Walk On Girl. My only guess here is just that Yuasa is a more common occurrence in text, at least in terms of the English internet.

In that sense, I worry that these tools will not only replace people, but they’ll reinforce existing cultural and societal biases even more so than we already do on our own now.

Further compounding issues like that is the censorship and filtering on OpenAI; virtually anything political, violent or sexual is not currently allowed (and may never be? I figure no company wants to be associated with the potential PR disasters that come with it, like Microsoft’s Tay). This is extra problematic given how important all of those things are in art. I worry that we will lose artists and have fewer people pushing the boundaries of art.

And ironically, for a society that criticizes countries like China so fervently for censorship, we rely so heavily on giant capitalist companies which self-enforce much of the same censorship, helping further enforce the status quo and restricting marginalized minorities. I am not a fan to say the least, despite understanding the potential for violent speech, propaganda, etc.

Anyway, that’s the end of my long-winded 6 AM phone rant. Apologies for any weird formatting or incoherent thoughts!

[zyx321]

Speaking of MTL, I have never found a satisfactory replacement for ChiiTrans. An app that shows multiple translations side-by-side, including literal dictionary translation and phonetic translation. It would even show you alternative meanings on mouse-over. Is there any more modern app like that? The western fans of Japanese visual novels were decades ahead of their time.

And I agree on the censorship angle too. Imagine if politicians had to stoop to Youtube Video Essay levels of self-censorship, talking about the parallels between badguy Germany and present day badguyism. Or about the psychological effects of naughty trafficking and childhood naughty abuse. I do not want my political discourse limited to PEGI-6 language.

[egypturnash]

Similarly, I don’t see why I shouldn’t be able to transform or straight-up copy and distribute someone else’s art if it provides people with happiness.

Would you stop doing it if the artists said “this is making me unhappy”?

[samatman]

Weird Al does.

He doesn't have to, parody is inherently transformative under copyright, but he's a decent guy, and he's kept a song or two in the crate because the original artist didn't want a parody released.

I support that, and I also support not caring, as a matter of principle. It's supererogatory; it shows that Weird Al is a good person, but not that someone who releases a parody without asking first is a bad person.

[shlurpy]

You gloss over the economic issues, but those are fundamentally at the center of it. Some copy-able work can be done by semi-amateurs in their free time, and some can be paid for as a 1-off contract without needing exclusivity, but most seems to require trained labor with a guaranteed return on investment.

What it comes down to is that under capitalism, a lot of art like movies, tv, video games, certain more polished professional tools, etc, requires an artifical concept of ownership to ensure that people are able to spend time creating them, because choosing to do too much uncompensatable labor is effectively choosing to starve to death.

But maybe, like me, you see the possibility of a future where we can find a way to compensate art that isn't locked into the grim realities of capitalism. Well, cool, but then political strategy has to be considered. As is, ML seems to be a technology that funnels money away from compensating creators and towards large corporations that can invest in ML and their shareholders. And in our practical reality, more money buys more speech, more speech tends to lead to more power. Thus, self interest rules the day, with the wealthiest uninterested in changing the fundamental economics systems and working to stop it.

That's why I am concerned about ML and copyright. It would be no problem in an ideal world, but in our world, it makes the status quo worse in a way that prevents progress towards that ideal.

[Trung0246]

I wonder what the chinese site in question? It has been long since I dabble into video compression scene beside just use ffmpeg when I needed

[tkgally]

Thanks for the rant! A bit longer than is normal for HN, but I enjoyed it and learned things from it.