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. LLMs cannot, by their very nature, lead to AGI.

LLMs cannot, by their very nature, lead to AGI.

Planlagt Fastgjort Låst Flyttet Ikke-kategoriseret
28 Indlæg 19 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.
  • elizafox@social.treehouse.systemsE elizafox@social.treehouse.systems

    LLMs cannot, by their very nature, lead to AGI. They very badly need people to believe it is AGI though, because their growth model does not work without it.

    (This excludes the financials, which are just straight-up fraud on wheels)

    I would say these claims now border on fraud. It will be fascinating in 20 years when we are looking back on this bubble and wondering how so many smart people were so easily swindled.

    AGI / ASI can't happen without an entirely new paradigm. This ain't it chief.

    jaystephens@mastodon.socialJ This user is from outside of this forum
    jaystephens@mastodon.socialJ This user is from outside of this forum
    jaystephens@mastodon.social
    wrote sidst redigeret af
    #3

    @Elizafox
    100% agree.
    I can imagine a hypothetical AGI of which 2026 style LLMs form a small part/interface... But the idea that these models can be the core of an AGI is bullshit on stilts.

    elizafox@social.treehouse.systemsE tk@f.kawa-kun.comT 2 Replies Last reply
    0
    • jaystephens@mastodon.socialJ jaystephens@mastodon.social

      @Elizafox
      100% agree.
      I can imagine a hypothetical AGI of which 2026 style LLMs form a small part/interface... But the idea that these models can be the core of an AGI is bullshit on stilts.

      elizafox@social.treehouse.systemsE This user is from outside of this forum
      elizafox@social.treehouse.systemsE This user is from outside of this forum
      elizafox@social.treehouse.systems
      wrote sidst redigeret af
      #4

      @jaystephens Exactly. We need more forms of machine learning besides this, but AI companies are like furiously going through their cue cards like "LLM... LLM... LLM..."

      jaystephens@mastodon.socialJ 1 Reply Last reply
      0
      • elizafox@social.treehouse.systemsE elizafox@social.treehouse.systems

        @jaystephens Exactly. We need more forms of machine learning besides this, but AI companies are like furiously going through their cue cards like "LLM... LLM... LLM..."

        jaystephens@mastodon.socialJ This user is from outside of this forum
        jaystephens@mastodon.socialJ This user is from outside of this forum
        jaystephens@mastodon.social
        wrote sidst redigeret af
        #5

        @Elizafox
        Fingers X'd the aftermath will be less destructive than previous "AI" backlashes

        aatch@mastodon.nzA 1 Reply Last reply
        0
        • jaystephens@mastodon.socialJ jaystephens@mastodon.social

          @Elizafox
          100% agree.
          I can imagine a hypothetical AGI of which 2026 style LLMs form a small part/interface... But the idea that these models can be the core of an AGI is bullshit on stilts.

          tk@f.kawa-kun.comT This user is from outside of this forum
          tk@f.kawa-kun.comT This user is from outside of this forum
          tk@f.kawa-kun.com
          wrote sidst redigeret af
          #6
          @jaystephens @Elizafox The truth is that we don't have real AI yet. :3
          tk@f.kawa-kun.comT noisecolor@toot.communityN 2 Replies Last reply
          0
          • tk@f.kawa-kun.comT tk@f.kawa-kun.com
            @jaystephens @Elizafox The truth is that we don't have real AI yet. :3
            tk@f.kawa-kun.comT This user is from outside of this forum
            tk@f.kawa-kun.comT This user is from outside of this forum
            tk@f.kawa-kun.com
            wrote sidst redigeret af
            #7
            @jaystephens @Elizafox All we have is misleading marketing pushed by supervillains.
            1 Reply Last reply
            0
            • jaystephens@mastodon.socialJ This user is from outside of this forum
              jaystephens@mastodon.socialJ This user is from outside of this forum
              jaystephens@mastodon.social
              wrote sidst redigeret af
              #8

              @Elizafox
              Yeah the gen AI bros have been annoyingly succesful at squatting on that term so that normal folks pretty much fully conflate AI with their products now. Sigh.

              1 Reply Last reply
              0
              • elizafox@social.treehouse.systemsE elizafox@social.treehouse.systems

                LLMs cannot, by their very nature, lead to AGI. They very badly need people to believe it is AGI though, because their growth model does not work without it.

                (This excludes the financials, which are just straight-up fraud on wheels)

                I would say these claims now border on fraud. It will be fascinating in 20 years when we are looking back on this bubble and wondering how so many smart people were so easily swindled.

                AGI / ASI can't happen without an entirely new paradigm. This ain't it chief.

                enobacon@urbanists.socialE This user is from outside of this forum
                enobacon@urbanists.socialE This user is from outside of this forum
                enobacon@urbanists.social
                wrote sidst redigeret af
                #9

                @Elizafox swindled is the operative word, also in general for capitalism.

                1 Reply Last reply
                0
                • tk@f.kawa-kun.comT tk@f.kawa-kun.com
                  @jaystephens @Elizafox The truth is that we don't have real AI yet. :3
                  noisecolor@toot.communityN This user is from outside of this forum
                  noisecolor@toot.communityN This user is from outside of this forum
                  noisecolor@toot.community
                  wrote sidst redigeret af
                  #10

                  @tk
                  Ai is just a word that many people can define many ways. It's short, recognizable and everyone knows what it means. Today we use it to describe llms, yesterday we used it for video game opponents and Roomba robots, tomorrow who knows.
                  You can say that for you , you define ai differently than other people. But most people use it for llms today and not only linguistically, also technically it's a good word to use for that.
                  @jaystephens @Elizafox

                  jaystephens@mastodon.socialJ 1 Reply Last reply
                  0
                  • elizafox@social.treehouse.systemsE elizafox@social.treehouse.systems

                    LLMs cannot, by their very nature, lead to AGI. They very badly need people to believe it is AGI though, because their growth model does not work without it.

                    (This excludes the financials, which are just straight-up fraud on wheels)

                    I would say these claims now border on fraud. It will be fascinating in 20 years when we are looking back on this bubble and wondering how so many smart people were so easily swindled.

                    AGI / ASI can't happen without an entirely new paradigm. This ain't it chief.

                    noisecolor@toot.communityN This user is from outside of this forum
                    noisecolor@toot.communityN This user is from outside of this forum
                    noisecolor@toot.community
                    wrote sidst redigeret af
                    #11

                    @Elizafox
                    Some of the stuff ai models do today is quite impressive though...
                    The more I read about the hugging face exploit, the more it seems to me there is real progress done, when I thought we are platoing.

                    I doubt llms can get to AGi however that is defined today, but it seems we can use them for a lot of other things.
                    Not to mention the infrastructure they bring. If there is a new paradigm, this computing power will be most likely needed to run it.

                    1 Reply Last reply
                    0
                    • elizafox@social.treehouse.systemsE elizafox@social.treehouse.systems

                      LLMs cannot, by their very nature, lead to AGI. They very badly need people to believe it is AGI though, because their growth model does not work without it.

                      (This excludes the financials, which are just straight-up fraud on wheels)

                      I would say these claims now border on fraud. It will be fascinating in 20 years when we are looking back on this bubble and wondering how so many smart people were so easily swindled.

                      AGI / ASI can't happen without an entirely new paradigm. This ain't it chief.

                      L This user is from outside of this forum
                      L This user is from outside of this forum
                      luc0x61@mastodon.gamedev.place
                      wrote sidst redigeret af
                      #12

                      @Elizafox That's the same that Yann LeCun said leaving Meta, and I followed his researches since the 90s to early 2000s.

                      But if people believes so widely on chem trails and vaccine hoaxes, what would you expect with a well packaged marketing move, arranged as a high technical breakthrough?

                      1 Reply Last reply
                      0
                      • noisecolor@toot.communityN noisecolor@toot.community

                        @tk
                        Ai is just a word that many people can define many ways. It's short, recognizable and everyone knows what it means. Today we use it to describe llms, yesterday we used it for video game opponents and Roomba robots, tomorrow who knows.
                        You can say that for you , you define ai differently than other people. But most people use it for llms today and not only linguistically, also technically it's a good word to use for that.
                        @jaystephens @Elizafox

                        jaystephens@mastodon.socialJ This user is from outside of this forum
                        jaystephens@mastodon.socialJ This user is from outside of this forum
                        jaystephens@mastodon.social
                        wrote sidst redigeret af
                        #13

                        @Noisecolor @tk @Elizafox
                        Semantically and technically, I agree.
                        The issue is about orgs with deep pockets doing a lot of work in the media and in sales pitch slide decks to the C-suite, to conflate hypothetical AGI, LLMs as they currently exist, and claimed but not delivered LLM abilities, under the loose term AI.

                        1 Reply Last reply
                        0
                        • elizafox@social.treehouse.systemsE elizafox@social.treehouse.systems

                          LLMs cannot, by their very nature, lead to AGI. They very badly need people to believe it is AGI though, because their growth model does not work without it.

                          (This excludes the financials, which are just straight-up fraud on wheels)

                          I would say these claims now border on fraud. It will be fascinating in 20 years when we are looking back on this bubble and wondering how so many smart people were so easily swindled.

                          AGI / ASI can't happen without an entirely new paradigm. This ain't it chief.

                          ahltorp@mastodon.nuA This user is from outside of this forum
                          ahltorp@mastodon.nuA This user is from outside of this forum
                          ahltorp@mastodon.nu
                          wrote sidst redigeret af
                          #14

                          @Elizafox The illusion of AGI is necessary to make the chatbot parlour trick work, so they have to push it.

                          1 Reply Last reply
                          0
                          • jaystephens@mastodon.socialJ jaystephens@mastodon.social

                            @Elizafox
                            Fingers X'd the aftermath will be less destructive than previous "AI" backlashes

                            aatch@mastodon.nzA This user is from outside of this forum
                            aatch@mastodon.nzA This user is from outside of this forum
                            aatch@mastodon.nz
                            wrote sidst redigeret af
                            #15

                            @jaystephens @Elizafox unfortunately I suspect that it'll be worse than previous AI Winters. There have been many orders of magnitude more money sunk into it and the exposure is way wider. Previous Winters have been "US Military discovers that it won't be able to replace interpreters and stops funding it", this is so much bigger.

                            jaystephens@mastodon.socialJ 1 Reply Last reply
                            0
                            • elizafox@social.treehouse.systemsE elizafox@social.treehouse.systems

                              LLMs cannot, by their very nature, lead to AGI. They very badly need people to believe it is AGI though, because their growth model does not work without it.

                              (This excludes the financials, which are just straight-up fraud on wheels)

                              I would say these claims now border on fraud. It will be fascinating in 20 years when we are looking back on this bubble and wondering how so many smart people were so easily swindled.

                              AGI / ASI can't happen without an entirely new paradigm. This ain't it chief.

                              moira@mastodon.murkworks.netM This user is from outside of this forum
                              moira@mastodon.murkworks.netM This user is from outside of this forum
                              moira@mastodon.murkworks.net
                              wrote sidst redigeret af
                              #16

                              @Elizafox Took me less than six weeks to figure that out doing early versions of this work. I cannot, for the life of me, figure out how the _fuck_ people think this gets them there. I really can't.

                              1 Reply Last reply
                              0
                              • elizafox@social.treehouse.systemsE elizafox@social.treehouse.systems

                                LLMs cannot, by their very nature, lead to AGI. They very badly need people to believe it is AGI though, because their growth model does not work without it.

                                (This excludes the financials, which are just straight-up fraud on wheels)

                                I would say these claims now border on fraud. It will be fascinating in 20 years when we are looking back on this bubble and wondering how so many smart people were so easily swindled.

                                AGI / ASI can't happen without an entirely new paradigm. This ain't it chief.

                                fandasin@social.linux.pizzaF This user is from outside of this forum
                                fandasin@social.linux.pizzaF This user is from outside of this forum
                                fandasin@social.linux.pizza
                                wrote sidst redigeret af
                                #17

                                @Elizafox

                                Generative "AI" still can't generate "full glass of red wine" and it never will, because there aren't enough pictures to train it on. (because common picture is half of glass of wine)
                                It can never produce anything that didn't get as input.

                                We won't get AGI from that for sure.

                                1 Reply Last reply
                                0
                                • elizafox@social.treehouse.systemsE elizafox@social.treehouse.systems

                                  LLMs cannot, by their very nature, lead to AGI. They very badly need people to believe it is AGI though, because their growth model does not work without it.

                                  (This excludes the financials, which are just straight-up fraud on wheels)

                                  I would say these claims now border on fraud. It will be fascinating in 20 years when we are looking back on this bubble and wondering how so many smart people were so easily swindled.

                                  AGI / ASI can't happen without an entirely new paradigm. This ain't it chief.

                                  tcatinreality@mastodon.socialT This user is from outside of this forum
                                  tcatinreality@mastodon.socialT This user is from outside of this forum
                                  tcatinreality@mastodon.social
                                  wrote sidst redigeret af
                                  #18

                                  @Elizafox

                                  "Border on fraud"?

                                  What techbros are pushing rn is way past the borders.

                                  1 Reply Last reply
                                  0
                                  • elizafox@social.treehouse.systemsE elizafox@social.treehouse.systems

                                    LLMs cannot, by their very nature, lead to AGI. They very badly need people to believe it is AGI though, because their growth model does not work without it.

                                    (This excludes the financials, which are just straight-up fraud on wheels)

                                    I would say these claims now border on fraud. It will be fascinating in 20 years when we are looking back on this bubble and wondering how so many smart people were so easily swindled.

                                    AGI / ASI can't happen without an entirely new paradigm. This ain't it chief.

                                    lethe@digitalcourage.socialL This user is from outside of this forum
                                    lethe@digitalcourage.socialL This user is from outside of this forum
                                    lethe@digitalcourage.social
                                    wrote sidst redigeret af
                                    #19

                                    @Elizafox As @pluralistic said, hoping AGI will arise out of an LLM is like hoping you will get a locomotive by breeding horses.

                                    1 Reply Last reply
                                    0
                                    • notavi10@critter.cafeN notavi10@critter.cafe

                                      @Elizafox
                                      "generative" ai is also not generative, it cant come up with things on its own, it has to take them from other sources (ex. an ai generated movie trailer i remember watching recreated exactly several memes, like the one where leonardo dicaprio is raising a martini glass)

                                      colman@mastodon.ieC This user is from outside of this forum
                                      colman@mastodon.ieC This user is from outside of this forum
                                      colman@mastodon.ie
                                      wrote sidst redigeret af
                                      #20

                                      @notavi10 @Elizafox “interpolative” would be a better word.

                                      1 Reply Last reply
                                      0
                                      • tcatinreality@mastodon.socialT This user is from outside of this forum
                                        tcatinreality@mastodon.socialT This user is from outside of this forum
                                        tcatinreality@mastodon.social
                                        wrote sidst redigeret af
                                        #21

                                        @Elizafox

                                        💯 on the finances

                                        But if the marketing statements only border on fraud, it's because the "truth in advertising" laws are too weak.

                                        1 Reply Last reply
                                        0
                                        • aatch@mastodon.nzA aatch@mastodon.nz

                                          @jaystephens @Elizafox unfortunately I suspect that it'll be worse than previous AI Winters. There have been many orders of magnitude more money sunk into it and the exposure is way wider. Previous Winters have been "US Military discovers that it won't be able to replace interpreters and stops funding it", this is so much bigger.

                                          jaystephens@mastodon.socialJ This user is from outside of this forum
                                          jaystephens@mastodon.socialJ This user is from outside of this forum
                                          jaystephens@mastodon.social
                                          wrote sidst redigeret af
                                          #22

                                          @aatch @Elizafox

                                          #unpopularopinion
                                          I also think it's possible that by sheer force of money, power, oligopoly & will, the current BS will make itself the substrate for a bunch of tech for a generation or two, making the current Microsoft/Apple/Meta era of enshittified mainstream and corporate tech look like a comparatively golden age.

                                          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