Noticed some AI evangelists try with the same kind of fake centrism, that was pretty popular during the Gamergate era. In this case, they try to set up people spamming genAI slop against “fanatical anti-AI people”. But just as many loves to pretend moderates are “far-left”, as their past-Gamergate selves pretended Anita Sarkeesian demanded state censorship of sexist videogames, as they are now pretending Hasan Piker is the most radical leftist living today (there are much worse even within the content creator sphere, such as Badempanda) to illustrate the “horseshoe theory”, the table is tilted in favor of AI adoption.

  • DevDave@piefed.social
    link
    fedilink
    English
    arrow-up
    4
    ·
    2 days ago

    tl;dr 1. I just don’t see “AI” panning out as anything but a fucking waste of time, energy, and resources. 2. I might consider AI as something to tolerate IF the “AI” techbro’s pay a “fair” amount to the copywrite owners whose material was sourced to “train” their models. Say $100,000 per picture or $300,000 per short story of text would be fair. Maybe $100 for each text “comment” collected from various sites. Plus of course ongoing licensing fees.

    Using a chatbot (codebot?) for writing software is always going to be a bit frustrating because it will do a great job right up until you let your guard down and it proceeds to invent an entire system service with API (imagine Google Maps sending you down a forest service road to a building that doesn’t exist but insists you need to go to the third floor, suite #324!).

    As for AI customer service, it works really well until stumbling at the same difficult problem as conventional customer service, how to handle a problem there is no existing policy for. Conventional support can do a risk/reward assessment of whether to involve someone higher up in the company. The “owner” or the person who swore this AI shit was going to be so much better and save a lot of time is incentivized to always fuck over the customer unless the perceived risk is obviously dire.

    AI content is ultimately derivative slop. The funniest thing I’ve seen made by AI is a video of two cats jousting with toilet bowl brushes in barbie hot wheel cars while a typical “valley girl” rages about it being 3am and stop fighting. What made this work was the human being that thought that absurdity up. Also this clip was very short which doesn’t expose AI’s weakness of being terrible at execution and pacing. Same problem as AI and coding, short controlled snippets are amazing, but if you try to get too greedy it will crap out these weird but impressively also really boring incomprehensible fever dreams.

    Last point. The people trying to sell “AI” are these dweeb super villains that have literally stolen almost every piece of human intellectual creation made in say the last 100 years. They just shoved it all into a form of holographic storage medium and have it shit out mashups of what they stole. As we learned with all the torrent lawsuits, copying is theft and should be punished; I don’t see any difference with AI.

    • TrooBloo@lemmy.dbzer0.com
      link
      fedilink
      arrow-up
      1
      ·
      2 days ago

      Your point about copyright is similar to one i keep coming back to. The data centers are already a huge cost sink, especially when you consider the rate their hardware will need to be upgraded/replaced over time. If you make them pay royalties to copyright holders whose material they constantly reproduce, pay a team of people to manage that inventory of works and authors, etc… The operating cost skyrockets. I see a regulation like this as being something of a nail in the coffin for AI. But I’m also a cynic and I don’t see a fascist nation like the US siding with the little guy artists against the tech giants who paid for their election.