Jessie James

@Neineon77

I've been told I'm "pocketable"

she/her
en.pronouns.page/@neineon77

I make my own patches, sometimes draw and I like old/cute video games. I work on computers and retro handhelds/consoles

Jessie James @Neineon773/6/2024, 7:56 AM

leigh @cindervision

lokeloski @lokeloski3/5/2024, 10:19 PM

This page's posts are visible only to users who are logged in.

Amy Dentata @amydentata3/5/2024, 11:24 PM

Machine translation can only really work as a first draft that is then reviewed by at least one human who knows the language. You can tell when translations and captions are done by machine only, because they suck!

lokeloski @lokeloski3/6/2024, 12:26 AM

This page's posts are visible only to users who are logged in.

leigh @cindervision3/6/2024, 1:35 AM

this also happens in technical documentation

I write API documentation for a living. my work gets translated into Japanese and there's an LLM as part of the translation process (I wish it weren't). I have to double check everything that comes back from them because holy shit does it break things, even things that you'd think would be obvious for a machine to parse. as an example, this morning I proofread a doc we got back and found that, for no discernible reason, the LLM had converted only one instance of a word in a code block into katakana, as if it were a loan word.

just one!!! it appeared nine times!!!!!! WHY DID IT DO THIS (the answer is because it doesn't know anything about anything, it is just guessing what word should come next AND ISN'T EVEN ALL THAT GOOD AT IT, FUCK)

anyway so it's falling to me to double-check because, while it may not be my job to translate the docs, at the end of the day it's my documentation coming back with errors and I'm responsible for that.

so yeah this shit fucking sucks to work with!!

You must log in to comment.

in reply to @lokeloski's post:

Luisf @Loosf3/5/2024, 11:22 PM

well

fucking hell, this has basically turned me off from getting it, it looked fairly interesting as far as diablolikes go

but goddamnit

three genders in a wenchcoat @akhra3/6/2024, 12:11 AM

pissed that I'm past the refund window and double mad because yeah, I've genuinely been enjoying it :host-frown:

Nineboat @Nineboat3/6/2024, 7:01 AM

Well that explains why i found the dialog very boring and just skipped through it.

Echo Cian @EchoCian3/6/2024, 7:01 PM

That explains the weird wording that shows up, but it overall seems too good for that, if a bit boring? A friend has been streaming it and a lot of the team seems to be from English speaking countries according to their site map. Maybe the early game got more attention or something.

in reply to @lokeloski's post:

Amy Dentata @amydentata3/6/2024, 1:03 AM

Oh gosh, it would be neat if all companies stopped doing that, but I only brought it up because it's still common practice!

NeckSpike @neckspike3/6/2024, 2:01 AM

backing up MTPE is way worse than starting from scratch, even editing a real janky human translated script is way more consistent and comprehensible.

Even the best mtl I've seen is really dry and sucks the life out of a script.

its miscu @miscu3/6/2024, 3:12 AM

I did a brief test with this on a combo of transcription and translation, and I'm convinced that even if they include every possible colloquialism into this tech, it still won't be able to handle most (if any) turn-of-phrases in media. it's like the popularity of LLM's made everyone forget how severely unreliable machine translation has always been

https://cohost.org/miscu/post/2956211-where-is-transcripti

Gabriel Hawke @GabrielHawke3/6/2024, 8:09 AM

Can confirm, the only way machine translation actually helps is when a company has trained it with its own documents to translate similar documents. Even then, you need a couple of people to go over it, because the machine is stubborn and sometimes refuses to use the terms the company prefers.
To just feed a game to a machine and use whatever it churns out is just greed and disregard for your customers.

Krake @Krake3/6/2024, 12:18 PM

I was thinking the only way a machine translation with a human reviewer would be functional to cover all the constant errors, wrong translations, and inconsistencies machine translations have would be someone: 1. fluent enough in both languages to just translate it, 2. familliar with the material inside and out enough to spot and fix issues, 3. willing to trudge through a book's worth of terrible grammar and rewrite all of it.

At that point just hire someone to translate...

John Colagioia @jcolag3/6/2024, 12:56 PM

This gets at a core problem that I've started talking about: The issue isn't "how good" the AI tools are at all, and arguing about "how close we are" plays right into the idea that this is an ultra-valuable market. The problem is that, even if it's technically perfect, the software is by definition never going to care about the result and its suitability for the purpose.

Vincent Kinian @Video-Game-King3/6/2024, 10:39 PM

If you want to be more precise about it, the problem is fundamental to how large language models work. You can't just feed them an arbitrary metric to judge their output by; they have to work with the metric that was programmed into them at the outset, which is always how closely that output resembles the relevant parts of the corpus. That this is a problem should become apparent the moment you apply this tech to anything that diverges significantly from the corpus, but even if it doesn't, the fact that the text you're working with isn't part of the corpus (why bother translating something the machine's already worked with?) means there's always going to be room for error.

Pinned Tags