Need to let loose a primal scream without collecting footnotes first? Have a sneer percolating in your system but not enough time/energy to make a whole post about it? Go forth and be mid: Welcome to the Stubsack, your first port of call for learning fresh Awful youā€™ll near-instantly regret.

Any awful.systems sub may be subsneered in this subthread, techtakes or no.

If your sneer seems higher quality than you thought, feel free to cutā€™nā€™paste it into its own post ā€” thereā€™s no quota for posting and the bar really isnā€™t that high.

The post Xitter web has spawned soo many ā€œesotericā€ right wing freaks, but thereā€™s no appropriate sneer-space for them. Iā€™m talking redscare-ish, reality challenged ā€œculture criticsā€ who write about everything but understand nothing. Iā€™m talking about reply-guys who make the same 6 tweets about the same 3 subjects. Theyā€™re inescapable at this point, yet I donā€™t see them mocked (as much as they should be)

Like, there was one dude a while back who insisted that women couldnā€™t be surgeons because they didnā€™t believe in the moon or in stars? I think each and every one of these guys is uniquely fucked up and if I canā€™t escape them, I would love to sneer at them.

(Taking over for Gerard this time. Special thanks to him for starting this.)

  • Architeuthis@awful.systems
    link
    fedilink
    English
    arrow-up
    17
    Ā·
    edit-2
    2 days ago

    Saltman has a new blogpost out he calls ā€˜Three Observationsā€™ that I feel too tired to sneer properly but Iā€™m sure will be featured in pivot-to-ai pretty soon.

    Of note that he seems to admit chatbot abilities have plateaued for the current technological paradigm, by way of offering the ā€œobservationā€ that model intelligence is logarithmically dependent on the resources used to train and run it (i = log( r )) so itā€™s officially diminishing returns from now on.

    Second observation is that when a thing gets cheaper itā€™s used more, i.e. theyā€™ll be pushing even harded to shove it into everything.

    Third observation is that

    The socioeconomic value of linearly increasing intelligence is super-exponential in nature. A consequence of this is that we see no reason for exponentially increasing investment to stop in the near future.

    which is hilarious.

    The rest of the blogpost appears to mostly be fanfiction about the efficiency of their agents that I didnā€™t read too closely.

    • mlen@awful.systems
      link
      fedilink
      English
      arrow-up
      8
      Ā·
      19 hours ago

      Second observation is that when a thing gets cheaper itā€™s used more, i.e. theyā€™ll be pushing even harded to shove it into everything.

      Are they trying to imply that when they will make it cheaper by shoving it everywhere? I honestly canā€™t see how that logic is holding together

      • froztbyte@awful.systems
        link
        fedilink
        English
        arrow-up
        8
        Ā·
        18 hours ago

        as I read it, itā€™s an attempt at reference to economy of scale under the thesis ā€œAI silicon will keep getting cheaper because more and more people will produce itā€ as the main underpinning for how to reduce their unit economics. which, yā€™know, great! thatā€™s exactly what people like to hear about manufacturing and such! lovely! itā€™s only expensive because itā€™s the start! oh, the woe of the inventor, the hard and expensive path of the start!

        except that doesnā€™t hold up in any reasonable manner.

        theyā€™re not using J Random GPU, theyā€™re using top-end purpose-focused shit thatā€™s come into existing literally as co-evolution feedback from the fucking industry that is using it. even some hypothetical path where we do just suddenly have a glut of cheap model-training silicon everywhere, imo itā€™s far far far more likely to be an esp32 situation than a ā€œyeah this gtx17900 cost me like 20 bucksā€ situation. even the ā€œconsumer high endā€ of ā€œsure your phone has a gpu in itā€ is still very suboptimal for doing the kind of shit theyā€™re doing (even if you could probably make a great cursed project out of a cluster of phones doing model training or whatever)

        falls into the same vein of shit as ā€œa few thousand daysā€ imo - something thatā€™s a great soundbite, easily digestible market speak, but if you actually look at the substance itā€™s comprehensive nonsense

        • mlen@awful.systems
          link
          fedilink
          English
          arrow-up
          5
          Ā·
          11 hours ago

          Yeah, I agree. I think that this argument that is made there is a false. The logic error imo is claiming this statement is true: things get cheaper as they get used more, therefore if we make it used more, it will get cheaper.

        • Architeuthis@awful.systems
          link
          fedilink
          English
          arrow-up
          8
          Ā·
          18 hours ago

          Could also be donā€™t worry about deepseek type messaging that addresses concerns without naming names, to tell us that a drastic reduction in infrastructure costs was foretold by the writing of St Moore and was thus always inevitable on the way to immanentizing the AGI, į¼€Ī»Ī»Ī·Ī»ĪæĻĻŠĪ±.

      • Architeuthis@awful.systems
        link
        fedilink
        English
        arrow-up
        7
        Ā·
        edit-2
        19 hours ago

        The surface claim seems to be the opposite, he says that because of Mooreā€™s law AI rates will soon be at least 10x cheaper and because of Mercury in retrograde this will cause usage to increase muchly. I read that as meaning we should expect to see chatbots pushed in even more places they shouldnā€™t be even though their capabilities have already stagnated as per observation one.

        1. The cost to use a given level of AI falls about 10x every 12 months, and lower prices lead to much more use. You can see this in the token cost from GPT-4 in early 2023 to GPT-4o in mid-2024, where the price per token dropped about 150x in that time period. Mooreā€™s law changed the world at 2x every 18 months; this is unbelievably stronger.
    • FredFig@awful.systems
      link
      fedilink
      English
      arrow-up
      13
      Ā·
      edit-2
      1 day ago
      1. My big robot is really expensive to build.

      2. If big robot parts become cheaper, I will declare that the big robot must be bigger, lest somebody poorer than me also build a big robot.

      3. My robot must be made or else I wonā€™t be able to show off the biggest, most expensive big robot.

      QED, I deserve more money to build the big robot.

      P.S. And for the naysayers, just remember that that robot will be so big that your critiques wonā€™t apply to it, as it is too big.

    • BigMuffin69@awful.systems
      link
      fedilink
      English
      arrow-up
      21
      Ā·
      edit-2
      1 day ago

      My ability to guess the solution of Boolean SAT problems also scales roughly with the log of number of tries you give me.

    • swlabr@awful.systems
      link
      fedilink
      English
      arrow-up
      12
      Ā·
      1 day ago

      It probably deserves its own post on techtakes, but letā€™s do a little here.

      People are tool-builders with an inherent drive to understand and create

      Diogenesā€™s corpse turns

      which leads to the world getting better for all of us.

      Of course Saltman means ā€œall of my buddiesā€ as he doesnā€™t consider 99% of the human population as human.

      Each new generation builds upon the discoveries of the generations before to create even more capable toolsā€”electricity, the transistor, the computer, the internet, and soon AGI.

      Ugh. Amongst many things wrong here, people didnā€™t jerk each other off to scifi/spec fic fantasies about the other inventions.

      In some sense, AGI is just another tool in this ever-taller scaffolding of human progress we are building together. In another sense, it is the beginning of something for which itā€™s hard not to say ā€œthis time itā€™s differentā€; the economic growth in front of us looks astonishing, and we can now imagine a world where we cure all diseases, have much more time to enjoy with our families, and can fully realize our creative potential.

      AGI IS NOT EVEN FUCKING REAL YOU SHIT. YOU CANā€™T CURE FUCK WITH DREAMS

      We continue to see rapid progress with AI development.

      I must be blind.

      1. The intelligence of an AI model roughly equals the log of the resources used to train and run it. These resources are chiefly training compute, data, and inference compute. It appears that you can spend arbitrary amounts of money and get continuous and predictable gains; the scaling laws that predict this are accurate over many orders of magnitude.

      ā€œIntelligenceā€ in no way has been quantified here, so this is a meaningless observation. ā€œDataā€ is finite, which negates the idea of ā€œcontinuousā€ gains. ā€œPredictableā€ is a meaningless qualifier. This makes no fucking sense!

      1. The cost to use a given level of AI falls about 10x every 12 months, and lower prices lead to much more use. You can see this in the token cost from GPT-4 in early 2023 to GPT-4o in mid-2024, where the price per token dropped about 150x in that time period. Mooreā€™s law changed the world at 2x every 18 months; this is unbelievably stronger.

      ā€œMooreā€™s lawā€ didnā€™t change shit! It was a fucking observation! Anyone who misuses ā€œmooreā€™s lawsā€ outta be mangioneā€™d. Also, if this is true, just show a graph or something? Donā€™t just literally cherrypick one window?

      1. The socioeconomic value of linearly increasing intelligence is super-exponential in nature. A consequence of this is that we see no reason for exponentially increasing investment to stop in the near future.

      ā€œLinearly increasing intelligenceā€ is meaningless as intelligence has not beenā€¦ wait, Iā€™m repeating myself. Also, ā€œsuper-exponentialā€ only to the ā€œsocioā€ that Olā€™ Salty cares about, which I have mentioned earlier.

      If these three observations continue to hold true, the impacts on society will be significant.

      Oh hm but none of them are true. What now???

      Stopping here for now, I can only take so much garbage in at once.

        • swlabr@awful.systems
          link
          fedilink
          English
          arrow-up
          11
          Ā·
          1 day ago

          Youā€™d think that, at this point, LW style AGI wish fulfilment fanfic would have been milked dry for building hype, but apparently Salty doesnā€™t!

          • istewart@awful.systems
            link
            fedilink
            English
            arrow-up
            5
            Ā·
            12 hours ago

            The cult will be the very last constituency to drop away, and likely form the majority of his hiring pool at this point, especially to replace the high-level attrition they suffered last year.

            • skillissuer
              link
              fedilink
              English
              arrow-up
              6
              Ā·
              11 hours ago

              You know, i wondered how come that in modern, tech literate, western allied country presidency was captured by a cult. (South Korea) i donā€™t wonder about that anymore