This is the fediverse, so this is probably going to be extremely unpopular, but yea here goes.

Not saying you have to love it, or trust it blindly. Skepticism is healthy. But there’s a difference between “I use it carefully and check everything” and “I refuse to touch it.”

Historically, refusing new tools has never stopped them. Coopers didn’t stop the steel drum by refsuing to make them. They just ended up out of work while the industry moved on.

The realistic threat isn’t AI replacing you. It’s a colleague who uses AI well doing your job faster. And the skill of using it properly (knowing when it’s wrong, veriyfing output, knowing what not to use it for) takes time to build. Starting late means competing with people who’ve had years of practice.

You can have legit concenrs about ethics, energy, or jobs and still learn the tool. Opting out doesn’t fix those problems. It just removes you from the people who get a say in how it’s used.

PS. Ethically trained models exist and you can run models locally.

  • Uranhjort@lemmy.world
    link
    fedilink
    English
    arrow-up
    2
    ·
    45 minutes ago

    The only ones who are getting left behind are the people irreparably damaging their ability to think by outsourcing it to the mediocrity machines.

    Any new scientific discovery, technology or event after 2022 or so will have the vast majority of its body of work be partially or wholly written by unlabelled AI, and that is pure poison to the training models.

    Every error, every “hallucination”, and every outright lie will be embedded and reinforced by repetition until it is as good as truth, and as a result LLM outputs will become less and less reliable as they gradually poison their own well.

  • kestrel7_7@lemmy.world
    link
    fedilink
    English
    arrow-up
    5
    ·
    2 hours ago

    I’ve traditionally been an early adopter of tech. My attitude is that I’ll start using AI as soon as it does something that I need. So far, I’ve been pretty unimpressed and uninspired by what this software does. I don’t really see any application to my life or work. Whenever I ask people about what they’re actually using this stuff for, I get unsatisfying answers that don’t convince me I need to check it out.

  • uen3@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    8
    arrow-down
    1
    ·
    4 hours ago

    On team “I refuse to touch it.” I’ve been a computer guy pretty much my whole life. My current job is almost certainly my last one doing computer things. I absolutely know I have fallen behind by never learning how to ‘do AI’ in an industry where the only thing that matters anymore is knowing how to do AI effectively. I don’t care, the software business is already dead to me. It means I’ll never make as much money as I did 5 years ago but it looks like neither will anyone else soon. Fuck it. My spite will sustain me. In all seriousness, my next job will probably be a mobile dog grooming partnership, and making ceramic pottery as a side hobby.

  • IronpigsWizard@lemmy.world
    link
    fedilink
    English
    arrow-up
    12
    ·
    edit-2
    4 hours ago

    I work in the field of mental health, should I start inputting individuals trauma and personal histories into a LLM for what reason? So I can risk a data leak that could literally cause someone to take their life if information is exposed in any sort of way?

    Do I need AI for personal notes and insurance documentation dictation that was 100% accurate without AI for some reason? (I personally still type everything myself though)

    My partner and I go camping at music festivals for fun along with some hiking, our primary extracurricular activity, what benefits is AI going to have integrated into that? I use solar for power with an on demand power station that doesn’t use AI, no issues there.

    I have several past jobs where you have to physically monitor someone so they do not harm themselves or others. This isn’t minority report (yet, nor would I want it to be), what benefit is there to AI in that scenario?

    The problem I have with AI is that all of its evangelists are like, “EVERYONE NEEDS THIS FOR EVERYTHING” and I’m a couple years into this crap existing, not needing it, happy and fine.

    I’ve had to hear the same/similar pitch for VR like clockwork every 10 years since 1970. What am I, are we falling behind on, in any of the previously mentioned? Geesh.

  • worlm@lemmy.world
    link
    fedilink
    English
    arrow-up
    29
    arrow-down
    1
    ·
    9 hours ago

    Fall behind in what, exactly? What skill am I not developing by not using it? What do you mean by “learn the tool”? The entire purpose is that there’s absolutely nothing to using it. If I hypothetically decided to stop washing my clothes by hand and use a washing machine for the first time today, would I really be meaningfully worse than someone who’s used one their whole life? I’m aware there are little tricks to getting better results out of it, but honestly I think it’s mostly diminishing returns. AI, that is. Though the same could be said about washing machines.

    I’m going to use my brain and stay sharp while my colleagues continue to outsource thinking to a machine. We’ll see who really falls behind in a few years. I don’t think their “years of practice” prompting will get them much.

    Opting out doesn’t fix those problems. It just removes you from the people who get a say in how it’s used.

    When I choose not to use it, I’m not contributing to its demand.

    you can run models locally

    On what hardware? I’m not buying shit right now.

    Apologies if I come across as angry. This opinion is indeed unpopular on Lemmy but I’m afraid I’m the one with the unpopular opinion in the real world, which I find incredibly frustrating.

  • dan1101@lemmy.world
    link
    fedilink
    English
    arrow-up
    27
    arrow-down
    2
    ·
    10 hours ago

    That is exactly what the billionaires like you thinking. That is why trillions of dollars are being poured into AI data centers. That is why every AI exec says their AI is a threat to the internet or a threat to all humanity, it’s them hyping up their models over the other guy’s.

    These are sloppy tools. If you asked it to prop up a table it might use 10,000 popsicle sticks, or 3 wagons. The table top may or may not be upright, it could be vertical. It still accomplishes the instruction.

    And you want AI to dig into millions of lines of code? AI will find an iteration that fits the instructions, but makes no design or maintenance sense. The databases and code will feel the ramifications of AI code for many years.

    • Taleya@aussie.zone
      link
      fedilink
      English
      arrow-up
      4
      ·
      10 hours ago

      Its like handing out pervitan and going “Look! We have better soldiers!!”

  • Tattorack@lemmy.world
    link
    fedilink
    English
    arrow-up
    12
    ·
    10 hours ago

    Studies have shown that AI use, no matter how “careful”, will result in cognative decline and skill loss, which results in more AI use, which results in more… You get the idea.

    In order to maintain a skill in something, you actually have to use your brain for problem solving. Offloading that to a deep learning algorithm means no longer keeping your brain sharp.

    Oh I have no doubt that these deep learning algorithms, so vapidly marketed as “AI”, will only get better and better. But as a result people will only get stupider. And as the AI continues to recycle old habits, and its own habits, for training data, more and more problems will occur over time. Problems that wouid require humans to fix, humans who no longer have the skills or the intelligence.

    So no, I don’t think I’ll fall behind.

    • Seralth@piefed.seralth.com
      link
      fedilink
      English
      arrow-up
      10
      arrow-down
      2
      ·
      9 hours ago

      Literally every single study I have seen on that topic has also had the same exact problem.

      It assumes a person will not replace the saved time with any other skill. That they will proceed to not research, learn or engage in any other activity at all.

      Every single time that gets accounted for those studies are proven flat out wrong.

      Fundamentally those studies are right in the sense that yes, if you don’t use your mind and you don’t use your brain, you don’t exercise it. You go stupid.

      People overwhelmingly will do something else and develop a different skill if a tool or a utility makes a previously hard task suddenly easy.

      Yes, people will lose certain computer skills. That just means that they now have the available time and resources to work on other skills.

      Some people are going to be lazy pieces of s*** and not. But that’s true of every time. Anything is made easier.

      It’s the same argument you could have against the very concept of a user interface.

      The invention of a GUI made nearly everyone lose the ability to use a terminal. We should get rid of GUIs they’re rotting the mind of people and making them lazy and stupid.

      • Undeclared_Mustard@lemmy.ca
        link
        fedilink
        English
        arrow-up
        2
        ·
        3 hours ago

        The invention of a GUI made nearly everyone lose the ability to use a terminal. We should get rid of GUIs they’re rotting the mind of people and making them lazy and stupid.

        You’re forgetting one thing: the proliferation of the GUI made computing an accessible function for people who never learned to use a CLI in the first place. It didn’t make people forget how to use a terminal, it just made people who can use a terminal more valuable. AI doesn’t create any such pathway because it’s being used to replace the people who knew how to use a terminal, and doing so poorly.

      • Annoyed_🦀 @lemmy.zip
        link
        fedilink
        English
        arrow-up
        4
        ·
        5 hours ago

        The invention of a GUI made nearly everyone lose the ability to use a terminal. We should get rid of GUIs they’re rotting the mind of people and making them lazy and stupid.

        Guess what? They’re doing just that but instead of syntax-based terminal it’s a chatbot. Are we smart yet

      • Tattorack@lemmy.world
        link
        fedilink
        English
        arrow-up
        9
        arrow-down
        1
        ·
        9 hours ago

        People overwhelmingly will do something else and develop a different skill if a tool or a utility makes a previously hard task suddenly easy.

        The studies haven’t shown this at all. Not even close. What they have shown is that AI will gradually get used more and more and more till it replaces even the most basic requirement for thought in their daily lives.

        This has absolutely no similarity to the development of the GUI. It’s not one skill that gets replaced with another. It is the complete removal of critical Human thought and problem solving. EVERY skill a person does is built upon these two things, and it’s being eliminated.

      • perturbedstarstuff@lemmy.world
        link
        fedilink
        English
        arrow-up
        5
        ·
        9 hours ago

        The saved time:

        Expected speedup: 24%
        Measured slowdown: 19%
        Post-task belief: 20% faster

        After the experiment ended and the timing data was up, the developers still walked away convinced AI had sped them up.

          • NotMyOldRedditName@lemmy.world
            link
            fedilink
            English
            arrow-up
            1
            ·
            27 minutes ago

            If its the one I think it is, it was having people use AI in their workflows that weren’t familiar with it yet. The people who were, did perform better, but overall because of all the people that werent it was worse.

            They re did it with people being more familiar and the results showed a modest improvement overall.

            Essentially, like all new tools and technology, it takes time to become proficient with it.

            Also, the models and tooling would have gotten better between tests unless they controlled for that too.

  • socsa@piefed.social
    link
    fedilink
    English
    arrow-up
    27
    ·
    11 hours ago

    It’s a colleague who uses AI well doing your job faster

    My professional experience so far has been that “faster” doesn’t actually work unless you abandon the whole concept of responsibility entirely and just blindly trust the AI. Writing code has never been the long tail of software engineering. It has always been code review, software management and debugging, and so far AI has actually made that significantly slower in my experience.

    Just this week my colleague and I pointed out a very laughably obvious flaw in a “true believer’s” code, and he just… argued about it, not because the AI can’t be wrong, but because of the implication of him submitting code he clearly did not understand. So far, that’s the norm, not the exception.

    Obviously it’s not going anywhere, but the pendulum is so far past the equilibrium it’s going to settle into eventually, that the loudest evangelists might as well just be wrong.

    • Ek-Hou-Van-Braai@piefed.socialOP
      link
      fedilink
      English
      arrow-up
      6
      arrow-down
      10
      ·
      10 hours ago

      This all or nothing extreems aren’t reality.

      You can have AI help you add unit-tests you wouldn’t have thought of.

      Once you are “Done” writing tests, and can’t think of more scenarios, AI more often than not can. And can help you add better test coverage.

      I’ve heard many people say this, what is the downside here?

        • Seralth@piefed.seralth.com
          link
          fedilink
          English
          arrow-up
          3
          arrow-down
          5
          ·
          10 hours ago

          You understand you can… Just read up and learn on the thing the slop machine suggests right?

          You don’t have to trust it. And you shouldn’t.

          But dude is right, the slop machine will 110% come up with legit suggestions because it’s the stolen collective knowledge of the internet.

          Your job is to exercise that thinking thing between your ears and validate information before implementing it. Regardless of the source.

        • Ek-Hou-Van-Braai@piefed.socialOP
          link
          fedilink
          English
          arrow-up
          7
          arrow-down
          7
          ·
          10 hours ago

          Then read the extra tests it came up with and make sure you understand it.

          It catches things like a null check you might have forgotten, trivially easy to understand.

          You don’t even have to have it write any tests if you want, you can ask it “what Unit-Test scenarios did I miss?” And write them yourself.

          • socsa@piefed.social
            link
            fedilink
            English
            arrow-up
            3
            ·
            9 hours ago

            That last scenario is the correct usage IMO, and largely how I’ve been using it. However, in my experience the actual result in actual organizations is that every software engineering taboo just falls to some combination of ego, laziness, and whip cracking, one by one, until something gets properly fucked, and only then are strict rules established.

            My major beef is that it is too tempting to abuse it, and the abuse inevitably becomes too voluminous to properly inspect, and something inevitably breaks. The same thing happens with humans, but humans moving at human speed make it exponentially easier for humans to comprehend. For bullshit slopware, code nobody actually has time to understand is honestly fine, but for serious systems it is absolutely not ok.

  • yesman@lemmy.world
    link
    fedilink
    English
    arrow-up
    167
    arrow-down
    4
    ·
    16 hours ago

    YSK that the Luddites were not anti-progress or anti-technology. They were an early labor movement opposed to the sweeping social changes brought about by industrialization and capitalism. Just like most anti-AI, they were the very people who worked with, and best understood the machines.

    The Luddite’s reputation as dumb people who hate the future is 100% capital propaganda that’s credulously slipped into history. I think it’s ironic that you frame AI as “the future” when your argument is the same song they’ve been singing for two centuries.

    People will decide how things go. How AI will develop isn’t a force of nature, it’s a choice that people will make. Believing AI is inevitable is a marketing strategy by the people who don’t want you or me to have any say in the matter.

    • tabarnaski@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      27
      arrow-down
      2
      ·
      14 hours ago

      Completely agree on how Luddites have been misrepresented, and how there are tons of similarities between the industrial revolution and what we are going through now with ai.

      That’s why I wouldn’t compare ai “haters” to Luddites.

    • nialv7@lemmy.world
      link
      fedilink
      English
      arrow-up
      4
      arrow-down
      7
      ·
      12 hours ago

      i wouldn’t say most anti-ai people best understand ai. just from personal experiences.

  • Retro_unlimited@lemmy.world
    link
    fedilink
    English
    arrow-up
    57
    arrow-down
    3
    ·
    14 hours ago

    People that use AI are the ones falling behind.

    Losing the ability to think for themselves, losing the ability to create. Its sad

    • Mikina@programming.dev
      link
      fedilink
      English
      arrow-up
      17
      arrow-down
      1
      ·
      edit-2
      13 hours ago

      Yeah, it’s like saying “Using Pervitin for sports will cause you being left behind. People who use it will be faster than you.”

      This argument falls apart when you consider the extreme athropy of skills, cognitive thinking and general frustration threshold. It is making you unable to work without the tool, and dependant on it.

      And the tool does not make any financial sense, and is literally billions of revenue in debt. And once the honeymoon phase ends, and they will have to figure out how to actually start making money, you and your company, will be priced out of all your “new and faster skills”, while anyone who avoided it will be waaay better than you for the jobs that are left.

      I’m pretty certain that only the megacorporates like Meta or Google will be able to keep using AI native workflows. Your startup the semi-sucesfull web dev company you work at will not be able to afford it.

      The end of this fiscal year is going to be great, once every company sees their AI bill and yearly result. Looking forward to it, I expect a lot of posts about “our AI spending is limited, what do I do”.

      • Seralth@piefed.seralth.com
        link
        fedilink
        English
        arrow-up
        4
        arrow-down
        2
        ·
        10 hours ago

        It’s unlikely people will be priced out of it. Not in any meaningful way. That logic shows a misunderstanding or ignorance of the current landscape.

        One of the biggest unspoken problems is the good enough effect.

        Local and mod scale models that can run efficiently and meet or exceed the needed utility level for most small to medium applications is basically here right now.

        Which undercuts the big companies entire business model to some degree.

        At current pace, a skilled developer can use the existing models to do real work. While your avg person with no skill can only really do very small scope things or generate slop.

        But if those models keep at pace then larger and larger workloads keep falling past the good enough line. Each time that happens the entire business model of the big ai companies becomes more and more shaky.

        Open weight models by their design undercut the very value and service the big companies are selling.

        The same thing happened in game hosting on a smaller scale. You had dedicated small companies early for huge premiums.

        But then as the software got better, more efficient faster, you stopped needing specialty technology a lot of compute. And you could start hosting it on your own computer or a small server. At that point it became more efficient to sell cheaper and in bulk than it was to sell a specialty good.

        All those early movers in the market died out basically overnight and then all that was left were the small bulk sellers and self-hosters.

        If openweight models keep at their pace, there is a legitimate chance that it’s going to be cheaper to buy a bunch of old server racks and cheap older gpus in a generation or two to run models at a fraction of the cost instead of paying these big companies.

        The only thing these big companies could then sell is management in other services around it. Which is still a fraction of the value and profit of what they’re doing now which still is going to collapse their entire business model.

        Lot of people seem to be acting like gpus are just never going to get any faster as well. And a faster GPU into more efficient GPU means that models go farther and farther.

        A lot of the value right now comes from the simple fact that really only one generation maybe too. Arguably worth of GPUs are even viable for this sort of work.

        And two or three generations of gpus. We’re going to see speeds and efficiency. Make what we’re doing now seem kind of quaint.

        Yeah the models will be bigger than and more bulky and need that more compute. But again the only thing that matters is the good enough line.

        If you could run anthropics mythos class frontier b******* on $1,000 server that you’re leaving in your little small business is closet. Then even if they never got any better at all from what they are right now.

        We have passed the good enough line.

        The entire industry is self-cannibalizing and and basically living in a pretend world as if technology is never going to advance. In everyone seems to be in the same delusion. The haters most of all seem to be forgetting technology advances and improves with time.

        And that means that these companies can’t keep doing what they’re doing because what they’re doing requires technology to stop advancing and it’s not going to stop.

  • mrmaplebar@fedia.io
    link
    fedilink
    arrow-up
    62
    arrow-down
    4
    ·
    14 hours ago

    Ridiculous and false. “Falling behind” is bullshit FOMO bait that doesn’t really reflect how things work.

    First off, as someone approaching 40, I learned how to use a computer multiple decades after the computer was invented. And as someone who learned to program in the 2010s, I was able to learn C++ 25 years after it came around. Despite all that, I work as a programmer today. It’s never too late to learn a genuinely useful skill.

    Second, “using AI” is not some kind of trade or skill that needs to be honed, like painting, playing a musical instrument, or speaking a foreign language. Rather it is a substitute for skill, and a replacement for thoughtful execution. It doesn’t take years of dedicated practice or specialized knowledge to learn how to skip steps and cut corners. Anyone competent who wants to “learn to prompt” can do so in a day.

    If anything, people who use the computer to think and work for them are arguably paying for it in ways that we aren’t fully aware of yet…

    What happens when you skip leg day at the gym? Your legs get weaker! Likewise, there are already studies that show that offloading thoughtful tasks to AI is bad for your brain and decreases you mental skills.

    The realistic threat isn’t AI replacing you. It’s a colleague who uses AI well doing your job faster.

    If your work can be delegated to AI, then why would someone hire you to do it at all? If I were your boss I would cut out the middle man and use AI to do the work myself.

    And maybe you’re thinking “but the AI isn’t good enough to do my job well!”

    Maybe so. But that hasn’t stopped other people’s bosses from trying to replace them, has it?

    The goal of a company is to make the most profit for its owners, right? Capitalism 101. It does so by maximizing its revenue and minimizing its costs.

    You are a cost. And so if it is cheaper to pay for tokens than it is to pay for man hours than we are ALL equally replaceable, especially by lazy and stupid management who have no eye for quality, artistry, thoughtfulness, etc.

    All of our higher ups are just one stupid decision away from deciding that they can do everything with AI instead of having a team of people.

    And that is a big problem for society…

    You can have legit concenrs about ethics, energy, or jobs and still learn the tool. Opting out doesn’t fix those problems. It just removes you from the people who get a say in how it’s used.

    If you “have legit concerns about” the ethics of something, but continue to do it, what kind of person does that make you? A hypocrite at best.

    In my view, today’s generative AI industry is built on a brazen and unprecedented exploitation of other people’s stolen labor at a scale not seen since slavery. And you don’t need to take me at my word on that.

    When I see people make stupid Ghibli slop, I don’t see it as harmless or cute in the slightest. I can only think about artists like Hayao Miyazaki, Kazuo Oga, Hideaki Anno, and so many countless other people whose work and vision I respect who have been ripped off and have had their life’s work thrown into some silicon valley corporate tech dweeb meat grinder without so much as basic consent.

    As such, how could I possibly use it in good faith without betraying my values?

    I’ll have no part in that. It’s unworthy of dignity or respect, and I am if nothing else a proud person. If that’s the future of society, I’ll opt out.

    PS. Ethically trained models exist

    Please name some. I have searched quite a bit and have only found 1 model that seems to be ethically trained.

  • Zensate@lemmy.world
    link
    fedilink
    English
    arrow-up
    73
    arrow-down
    2
    ·
    15 hours ago

    Fall behind what? The race to slopify every product or the one to destroy the earth the fastest?

    • ContactClosure@lemmus.org
      link
      fedilink
      English
      arrow-up
      17
      ·
      12 hours ago

      Exactly. I don’t want to participate in the matrix, I’ll keep growing potatoes. If they think I’m the one that got duped that’s perfectly OK with me.

    • Ek-Hou-Van-Braai@piefed.socialOP
      link
      fedilink
      English
      arrow-up
      4
      arrow-down
      18
      ·
      12 hours ago

      It will depend on what work you do. Like if you’re a plumber by all means, don’t use it.

      If you build basic websites, you’re just going to be far far slower than your colleagues or competition.

    • StabbingSky@lemmy.zip
      link
      fedilink
      English
      arrow-up
      20
      arrow-down
      3
      ·
      15 hours ago

      Adobe’s genAI offering is trained on licensed works. AWS’s offerings are also trained on licensed works.

        • Aatube@kbin.melroy.org
          link
          fedilink
          arrow-up
          11
          arrow-down
          5
          ·
          15 hours ago

          Do you need a frontier model? There is an extreme amount of community around Llama and it works very well.

        • Ek-Hou-Van-Braai@piefed.socialOP
          link
          fedilink
          English
          arrow-up
          7
          arrow-down
          6
          ·
          16 hours ago

          Depends on the model, but that’s not my point.

          You can even train your own, or the “Ethical” models might get a lot better in the near future.

          • breadsmasher@lemmy.world
            link
            fedilink
            English
            arrow-up
            12
            arrow-down
            6
            ·
            16 hours ago

            Yeah. you can “train ethical models”. but are they useful. im not defending AI here. Just saying - “full open source” llama is wildly behind “closed source frontier models”.

            No point training something that turns out useless.

            • Ek-Hou-Van-Braai@piefed.socialOP
              link
              fedilink
              English
              arrow-up
              8
              arrow-down
              5
              ·
              16 hours ago

              Then don’t use them, use a open weight model. The are very capable and can be run locally or on a rented server.

                • Seralth@piefed.seralth.com
                  link
                  fedilink
                  English
                  arrow-up
                  2
                  arrow-down
                  1
                  ·
                  10 hours ago

                  You… Uhh… Kind of don’t have a point is the problem. Your dancing around a kind of vagary which means any response to you is going to appear off topic.

                  Convey a defineable point, topic and logic and people will be able to debate or converse with you about it.

                  Doing what ever it is here your doing is only going to result in people fumbling in the dark trying to meet you half way to nowhere.

  • jpreston1986@lemmy.world
    link
    fedilink
    English
    arrow-up
    6
    ·
    11 hours ago

    My boss just gave me a lesson in how to use AI to do some of my job, and… well it didn’t really do it that much faster than me, and I can’t really trust the output from it, so I’d have to verify the info, and the only way to do that is to just, do that actual work, so it really just makes my job harder…

    That being said, I think it will definitely be a useful tool in the future, however, I still view it as I would any other piece of technology. Right now it’s in it’s beta stage. It’s almost working right. It’s almost user friendly, there are ways to optimize the data centers so that it doesn’t destroy our environment.

    These are all things that will improve, and I expect it to, over time. But right now, it’s still the hallucinating, lying, environment destroying shiny new toy that requires a lot more work.

    I’ll use it when it’s fixed. When it’s stable.

    I’m the kind of gamer that’s waiting till GTA 6 gets a PC port, so I’m used to waiting for a bit to get a better, more polished, finished product.

    • Seralth@piefed.seralth.com
      link
      fedilink
      English
      arrow-up
      1
      ·
      10 hours ago

      It’s an overpriced, expensive novelty toy at the moment.

      Some of the smaller models are useful for a little things. But most of those run on CPU and aren’t really all that helpful for any real.

      And two or three generations of gpus and improvements to hardware models that are currently considered frontier probably will run on reasonable home equipment. Not your Grandad’s computer, but like a high-end gaming PC or workstation more than likely.

      At that point we’ll probably hit the good enough threshold where it’ll be useful and reliable enough to do mundane tasks with little to no error.

      And that’s probably where all this is going to end up after the bubble popping. It’s all just going to kind of be mundane and boring. It’s going to change the world but not in any overly large way at least not quickly.

      It’s probably just going to be like the.com bubble and everything is just going to slowly shift over like a 20-year span.

      The whole business model of a giant company selling AI subscription honestly is cannibalizing itself when you step back and look at it. They’re making more efficient models that are more compact and faster generation after generation at a ridiculous speed.

      The moment a new generation of hardware comes out to run these things even faster, you’ve suddenly just massively hurt your ability to sell the lower end of your stock.

      Given enough generations of hardware and models. You’ll reach a point where the small models are doing what the giant models could do generations ago. At speeds faster than those models could do generations ago.

      And what the hell does your average small business even need with a super fast super big model when something they can run off computer that they have right there that they own in their own office can just do.

      You just end up at a point where you can’t sell your product. People aren’t going to need frontier class models forever. The better the mid-tier wants to get and the better the low tier ones get the less of a value proposition the frontier ones become.