• SmokeyDope@lemmy.world
    link
    fedilink
    English
    arrow-up
    13
    arrow-down
    1
    ·
    edit-2
    3 days ago

    Some people need something to rage and virtue signal against. Those who work in private STEM sectors or took machine learning classes years before the LLM craze already understand the tool is here and are willing to learn to work with it if applicable in their job or daily life.

    Those who don’t understand anything about the science of machine learning and are angry at the how megacorporations got away with unconsentually scraping their copyright infringed data off the internet for the first iterations of training data still get to let off some steam by calling it ‘hyped autocomplete just as bad as NFTs that will never do what a person can’.

    If I were an artsy type whos first exposure to ML was having my work stolen followed by the thief bragging to my face about how copy protection laws dont matter to the powerful and now they can basically copy my honed style 1 to 1 with a computer to sell as an product, I would be unreasonably pissed too and not interested this whole 'AI’thing. Megacorps made chatGPT and stable diffusion using my work therefore AI bad. I get it.

    That said, I’m not an artsy type or an idealist. I’m a practical engineer who builds systems to process the flows of information and energy with the tools available at my dispersal. Theres more to machine learning than proprietary models made with stolen information to be sold to th masses. Instead models are just the next new way to process large datasets full of complicated information. Its just that now were taking cues from natures biological information processing systems. Whether such processes prove more certain and effective to the old analog and digital ways have yet to be seen. Perhaps using these new tools will open up entirely different ways of treating information for all of society. Perhaps it will be just another niche thing for researchers to write papers about. Time will tell.