Skip to content
  • Hjem
  • Seneste
  • Etiketter
  • Populære
  • Verden
  • Bruger
  • Grupper
Temaer
  • Light
  • Brite
  • Cerulean
  • Cosmo
  • Flatly
  • Journal
  • Litera
  • Lumen
  • Lux
  • Materia
  • Minty
  • Morph
  • Pulse
  • Sandstone
  • Simplex
  • Sketchy
  • Spacelab
  • United
  • Yeti
  • Zephyr
  • Dark
  • Cyborg
  • Darkly
  • Quartz
  • Slate
  • Solar
  • Superhero
  • Vapor

  • Default (No Skin)
  • No Skin
Kollaps
FARVEL BIG TECH
  1. Forside
  2. Ikke-kategoriseret
  3. Machine translations are often brought up as a gotcha whenever I criticize LLMs.

Machine translations are often brought up as a gotcha whenever I criticize LLMs.

Planlagt Fastgjort Låst Flyttet Ikke-kategoriseret
255 Indlæg 170 Posters 0 Visninger
  • Ældste til nyeste
  • Nyeste til ældste
  • Most Votes
Svar
  • Svar som emne
Login for at svare
Denne tråd er blevet slettet. Kun brugere med emne behandlings privilegier kan se den.
  • gargron@mastodon.socialG gargron@mastodon.social

    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.

    vy@mastodon.socialV This user is from outside of this forum
    vy@mastodon.socialV This user is from outside of this forum
    vy@mastodon.social
    wrote sidst redigeret af
    #158

    @Gargron there is a great essay on translation by Simon Leys

    1 Reply Last reply
    0
    • kiloku@burnthis.townK kiloku@burnthis.town

      @Tekchip @Gargron the tiny potential for very rare good outcomes are not worth the constant poisoning of humanity's collective information corpus.

      For every "good" generated content there are dozens of thousands of terrible slop that are difficult to separate from genuine useful information or material when doing research or code reviews, etc.

      Not to mention that these "good" outcomes are much costlier to humanity than creating by hand, with no benefit.

      tekchip@mastodon.socialT This user is from outside of this forum
      tekchip@mastodon.socialT This user is from outside of this forum
      tekchip@mastodon.social
      wrote sidst redigeret af
      #159

      @Kiloku @Gargron the problem is you want to assume they are rare outcomes. I don't believe they are. Unfortunately that's where we're at an impasse. It's literally impossible to measure the good outcomes.

      I agree the environmental outcome is terrible. I don't like that part. What we can look forward to is the technology improving. General computers used to use WAY more power than they do now. The same is going to happen with LLM technology. Hopefully sooner than later. Folks are working on it.

      kiloku@burnthis.townK 1 Reply Last reply
      0
      • kiloku@burnthis.townK kiloku@burnthis.town

        @Tekchip @Gargron the tiny potential for very rare good outcomes are not worth the constant poisoning of humanity's collective information corpus.

        For every "good" generated content there are dozens of thousands of terrible slop that are difficult to separate from genuine useful information or material when doing research or code reviews, etc.

        Not to mention that these "good" outcomes are much costlier to humanity than creating by hand, with no benefit.

        kiloku@burnthis.townK This user is from outside of this forum
        kiloku@burnthis.townK This user is from outside of this forum
        kiloku@burnthis.town
        wrote sidst redigeret af
        #160

        @Tekchip @Gargron (also, most of what "AI" boosters *think* is good generated content is actually laughably bad to anyone who knows the subject matter of the content it generates. I'm certain you've shared something that you thought was indistinguishable from human created content that other people knew and saw a bunch of problems with as soon as they examined it further than a cursory glance)

        1 Reply Last reply
        0
        • cliphead@social.cologneC cliphead@social.cologne

          @Tekchip
          There's no point in explaining, if you don't get "this", tbh.

          @melioristicmarie @Gargron

          melioristicmarie@tech.lgbtM This user is from outside of this forum
          melioristicmarie@tech.lgbtM This user is from outside of this forum
          melioristicmarie@tech.lgbt
          wrote sidst redigeret af
          #161

          @ClipHead sigh. i agree. there are more porous space to beat my head against. 💜

          @Tekchip @Gargron

          1 Reply Last reply
          0
          • melioristicmarie@tech.lgbtM melioristicmarie@tech.lgbt

            @Tekchip
            so... is this a slop account? am i tooting with cheapgpt?

            are you a human playing with toys you do not comprehend?

            dear dogs, may i have the confidence of a mediocre "white" man.

            so... l.l.m.s tokenize english text... and then calculate an average.

            humans making shitty art is qualitatively perfection in comparison to word salad from a calculator. when you enter this into wannabe deep seek... i will be waiting with bated breath for the token response. ; )

            @ClipHead @Gargron

            tekchip@mastodon.socialT This user is from outside of this forum
            tekchip@mastodon.socialT This user is from outside of this forum
            tekchip@mastodon.social
            wrote sidst redigeret af
            #162

            @melioristicmarie @ClipHead @Gargron lol are you an LLM or just don't care to review my profile? Shoot even do a google search. I'm easy to find. Seems like you've lost the plot.

            1 Reply Last reply
            0
            • tekchip@mastodon.socialT tekchip@mastodon.social

              @Kiloku @Gargron the problem is you want to assume they are rare outcomes. I don't believe they are. Unfortunately that's where we're at an impasse. It's literally impossible to measure the good outcomes.

              I agree the environmental outcome is terrible. I don't like that part. What we can look forward to is the technology improving. General computers used to use WAY more power than they do now. The same is going to happen with LLM technology. Hopefully sooner than later. Folks are working on it.

              kiloku@burnthis.townK This user is from outside of this forum
              kiloku@burnthis.townK This user is from outside of this forum
              kiloku@burnthis.town
              wrote sidst redigeret af
              #163

              @Tekchip @Gargron I *know* they are rare.

              tekchip@mastodon.socialT 1 Reply Last reply
              0
              • jawarajabbi@mastodon.onlineJ jawarajabbi@mastodon.online

                @Gargron

                True story: I wanted to read the novel "The Hunchback of Notre Dame" by Victor Hugo some years back, so I went to the bookstore and they had two translations. The first had a serious-looking cover and the other had a trashy-looking one, so naturally I bought the former. Started to read it. It was garbage! So I went back and exchanged for the trashy-looking book. A wonderful translation!

                Moral of the story: you can't judge a book by its cover.

                Also, translation is art.

                kelson@notes.kvibber.comK This user is from outside of this forum
                kelson@notes.kvibber.comK This user is from outside of this forum
                kelson@notes.kvibber.com
                wrote sidst redigeret af
                #164

                @jawarajabbi @Gargron Similarly, I've read two different translations of Les Miserables and fragments of several others, and they're drastically different, despite all being professional human translators working from the same source text and translating it to the same language.

                (The oldest ones are really awkward to read now. They're also old enough to be in the public domain, so every random set of Serious Classic Books is going to print one of the the 1860s or 1880s versions instead of a more modern translation they'd have to pay royalties for.)

                1 Reply Last reply
                0
                • kiloku@burnthis.townK kiloku@burnthis.town

                  @Tekchip @Gargron I *know* they are rare.

                  tekchip@mastodon.socialT This user is from outside of this forum
                  tekchip@mastodon.socialT This user is from outside of this forum
                  tekchip@mastodon.social
                  wrote sidst redigeret af
                  #165

                  @Kiloku @Gargron Please let the rest of us know how to tell when we've seen a good LLM output. Seriously, if we can all tell the good and the bad then we can start gathering some data to have an even more rational conversation.

                  1 Reply Last reply
                  0
                  • alice@mk.nyaa.placeA alice@mk.nyaa.place

                    @gabboman@gabboman.xyz @aeva@mastodon.gamedev.place @Gargron@mastodon.social tbf that's not translation, that's japanese speakers writing english

                    And IMO broken english in an old videogame is so much better than soulless LLM translation. Like yeah, it may be jibberish, but it's a part of the charm

                    aeva@mastodon.gamedev.placeA This user is from outside of this forum
                    aeva@mastodon.gamedev.placeA This user is from outside of this forum
                    aeva@mastodon.gamedev.place
                    wrote sidst redigeret af
                    #166

                    @alice @Gargron @gabboman this is a tangent, but I saw this article float by my feed a few weeks ago and found it to be very entertaining https://legendsoflocalization.com/articles/super-mario-rpg-japan-pop-culture/

                    alice@mk.nyaa.placeA 1 Reply Last reply
                    0
                    • aeva@mastodon.gamedev.placeA aeva@mastodon.gamedev.place

                      @alice @Gargron @gabboman this is a tangent, but I saw this article float by my feed a few weeks ago and found it to be very entertaining https://legendsoflocalization.com/articles/super-mario-rpg-japan-pop-culture/

                      alice@mk.nyaa.placeA This user is from outside of this forum
                      alice@mk.nyaa.placeA This user is from outside of this forum
                      alice@mk.nyaa.place
                      wrote sidst redigeret af
                      #167

                      @aeva@mastodon.gamedev.place @Gargron@mastodon.social @gabboman@gabboman.xyz i haven't read that specific article, but the entirety of that website is really good and i can't recommend it enough

                      aeva@mastodon.gamedev.placeA 1 Reply Last reply
                      0
                      • alice@mk.nyaa.placeA alice@mk.nyaa.place

                        @aeva@mastodon.gamedev.place @Gargron@mastodon.social @gabboman@gabboman.xyz i haven't read that specific article, but the entirety of that website is really good and i can't recommend it enough

                        aeva@mastodon.gamedev.placeA This user is from outside of this forum
                        aeva@mastodon.gamedev.placeA This user is from outside of this forum
                        aeva@mastodon.gamedev.place
                        wrote sidst redigeret af
                        #168

                        @alice @Gargron @gabboman read it there's something really surprising half way through

                        alice@mk.nyaa.placeA 1 Reply Last reply
                        0
                        • gargron@mastodon.socialG gargron@mastodon.social

                          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.

                          S This user is from outside of this forum
                          S This user is from outside of this forum
                          slotos@toot.community
                          wrote sidst redigeret af
                          #169

                          @Gargron Funnily enough, literature is an easier translation target than social media. The latter is ever evolving, the former is frozen in time.

                          1 Reply Last reply
                          0
                          • gargron@mastodon.socialG gargron@mastodon.social

                            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.

                            benaveling@infosec.exchangeB This user is from outside of this forum
                            benaveling@infosec.exchangeB This user is from outside of this forum
                            benaveling@infosec.exchange
                            wrote sidst redigeret af
                            #170

                            Translations are like husbands. There are beautiful translations and there are faithful translations but the beautiful ones are not faithful and the faithful ones are not beautiful.
                            @Gargron

                            1 Reply Last reply
                            0
                            • woe2you@beige.partyW This user is from outside of this forum
                              woe2you@beige.partyW This user is from outside of this forum
                              woe2you@beige.party
                              wrote sidst redigeret af
                              #171

                              @Taco_lad @df @Gargron Have you seen how much money there is in Beanie Babies?

                              1 Reply Last reply
                              0
                              • tekchip@mastodon.socialT tekchip@mastodon.social

                                @Gargron would you know if you've seen a good outcome of an LLM? You'd somehow be able to identify when the LLM got it right?

                                I assure you you've experienced good LLM output and don't even know it. Because that's what good LLM output looks like. Indistinguishable from human output.

                                Your examples are perhaps false equivalencies. Take asbestos. We didn't abolish insulation. We developed better, safer insulation. We didn't stop dying food colors, we just developed safer dyes etc.

                                benaveling@infosec.exchangeB This user is from outside of this forum
                                benaveling@infosec.exchangeB This user is from outside of this forum
                                benaveling@infosec.exchange
                                wrote sidst redigeret af
                                #172

                                Oh LLM output always looks good, as long you don't understand what it's talking, then it looks great.
                                Very beautiful, very plausible.
                                But if you actually understand whatever it is that the LLM is talking about, then it rapidly becomes obvious that it's just spewing all the right words in a random order.
                                @Tekchip @Gargron

                                1 Reply Last reply
                                0
                                • gargron@mastodon.socialG gargron@mastodon.social

                                  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.

                                  pointlessone@status.pointless.oneP This user is from outside of this forum
                                  pointlessone@status.pointless.oneP This user is from outside of this forum
                                  pointlessone@status.pointless.one
                                  wrote sidst redigeret af
                                  #173

                                  @Gargron funnily enough the transformer block used in most LLMs was invented for translation and TBH LLMs are much better at translation than anything before them. Not refuting your point but as machine translation goes LLMs are the best we have.

                                  1 Reply Last reply
                                  0
                                  • gargron@mastodon.socialG gargron@mastodon.social

                                    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.

                                    wally@thepit.socialW This user is from outside of this forum
                                    wally@thepit.socialW This user is from outside of this forum
                                    wally@thepit.social
                                    wrote sidst redigeret af
                                    #174

                                    @Gargron

                                    Also, LLMs are making machine translations worse by adding hallucinated content into the translations:

                                    https://www.404media.co/ai-translations-are-adding-hallucinations-to-wikipedia-articles/

                                    zeborah@mastodon.nzZ 1 Reply Last reply
                                    0
                                    • gargron@mastodon.socialG gargron@mastodon.social

                                      From what I've observed, people who claim that LLMs can replace artists don't understand art, people who claim that they can replace musicians don't understand music, people who claim that they can replace writers don't understand literature, and people who claim they can replace translators don't rely on translations. If I had a button that would erase LLMs from the world but it would take machine translations away (which is a false dichotomy anyway), I would absolutely still press it.

                                      jamesmarshall@sfba.socialJ This user is from outside of this forum
                                      jamesmarshall@sfba.socialJ This user is from outside of this forum
                                      jamesmarshall@sfba.social
                                      wrote sidst redigeret af
                                      #175

                                      @Gargron exactly. To me, all forms of art are still about the human connection-- "Art is a tryst, for in it maker and beholder meet." The artist is communicating something that the beholder receives. With computer-generated art, the sentience on the other end is simply not there, and any "connection" is just an illusion.

                                      1 Reply Last reply
                                      0
                                      • dudinka@mastodon.worldD dudinka@mastodon.world

                                        @Gargron

                                        i do appreciate automatic subtitles extremely for hitting all my humor-chords. may they never evolve.

                                        that being said:
                                        i am lucky and able to read in several languages and read a lot of our bookclub books in original language. i can't count how many times i liked books that many of the others couldn't even finish their translated ones (assumedly) because the language was so poor. (and then we have those who listen to books and it totally depends on the person who was recorded.

                                        golemwire@fosstodon.orgG This user is from outside of this forum
                                        golemwire@fosstodon.orgG This user is from outside of this forum
                                        golemwire@fosstodon.org
                                        wrote sidst redigeret af
                                        #176

                                        @dudinka @Gargron
                                        Screen: *blissful silence*
                                        Machine translation: ..."money"

                                        Definitely makes one laugh 😆

                                        1 Reply Last reply
                                        0
                                        • benedictc@mas.toB benedictc@mas.to

                                          @Gargron we also had Concorde but it wasn’t economically viable. I mention that because I find that economic arguments seem to be heard more readily than moral arguments. (I often find that moral arguments induce temporary deafness in pro-AI people.)

                                          khleedril@cyberplace.socialK This user is from outside of this forum
                                          khleedril@cyberplace.socialK This user is from outside of this forum
                                          khleedril@cyberplace.social
                                          wrote sidst redigeret af
                                          #177

                                          @benedictc @Gargron Concorde wasn't morally viable either, both in terms of passenger safety and damage to the environment.

                                          1 Reply Last reply
                                          0
                                          Svar
                                          • Svar som emne
                                          Login for at svare
                                          • Ældste til nyeste
                                          • Nyeste til ældste
                                          • Most Votes


                                          • Log ind

                                          • Har du ikke en konto? Tilmeld

                                          • Login or register to search.
                                          Powered by NodeBB Contributors
                                          Graciously hosted by data.coop
                                          • First post
                                            Last post
                                          0
                                          • Hjem
                                          • Seneste
                                          • Etiketter
                                          • Populære
                                          • Verden
                                          • Bruger
                                          • Grupper