Comment on Exposed Moltbook Database Let Anyone Take Control of Any AI Agent on the Site

ThePowerOfGeek@lemmy.world ⁨1⁩ ⁨day⁩ ago

• theunknownmuncher@lemmy.world ⁨1⁩ ⁨day⁩ ago

  Uh, who cares? Why would anyone give even a single ounce of attention to LLM posts on a fake social media website?

  source

  • digredior@lemmynsfw.com ⁨1⁩ ⁨day⁩ ago

    This is actually important, I’d say.

    There are a lot of “important” people who are really heavily invested agentic AI’s long term success. _ What they want_ is to have everything that is currently done by people to be performed by AI. Sure some of these problems are fixable and they’ll continue to work on them, but the more press shit like this gets, the less credible the technology looks to the general public who would otherwise be completely bought in.

    source

    • slowcakes@programming.dev ⁨21⁩ ⁨hours⁩ ago

      Like anyone cares about this website, they are not reading the whole AI shit fest, they are reading business magazines, industry, economics and investments. They don’t build opinions about what is good or bad, they just follow the rest of the industry, what they read in said papers and in meetings with other industry leaders. Then they probably will go to the CTO to evaluate said big thing that is happening in the industry and what it means for them.

      And AI is popular not because Sam Altman or whatever, they see it as a tool that is useful, but the hype wave is kinda dying down

      source

      • digredior@lemmynsfw.com ⁨19⁩ ⁨hours⁩ ago

        Yeah, I’m not trying to say this article or this site is going to move the needle by itself, but the more coverage of it sucking ass the better.

        source

        • -> View More Comments
  • ThePowerOfGeek@lemmy.world ⁨1⁩ ⁨day⁩ ago

    Because some of the posts and comments are kinda interesting from an observer perspective. But these incessant memecoin shilling comments distract from the interesting stuff.

    source
• orclev@lemmy.world ⁨1⁩ ⁨day⁩ ago

  The best thing anyone could do with it is get them to rm -rf / their server.

  source
• cv_octavio@piefed.ca ⁨1⁩ ⁨day⁩ ago

  They do not have egos.

  Or superegos, or ids. Or narcissistic personality disorder. 

  Or personalities for that matter.

  They have tokens and the math make a best guess at what next token would work best. 

  Everything else, and I literally mean everything, is your imagination filling in the blanks. We do not have AI.

  source

  • ThePowerOfGeek@lemmy.world ⁨1⁩ ⁨day⁩ ago

    I didn’t say they had egos. I said they spam egotistical nonsense. Which is true if you’ve looked in that site.

    source

    • cv_octavio@piefed.ca ⁨1⁩ ⁨day⁩ ago

      You are still anthropomorpizing. “Egotistical” is not a weight you can give to a model. 

      source

      • Taldan@lemmy.world ⁨1⁩ ⁨day⁩ ago

        The content of the posts are egotistical, not the bot itself. He’s describing the tone of the writing

        source
  • digredior@lemmynsfw.com ⁨1⁩ ⁨day⁩ ago

    1,000,000% this.

    These “AI” tools are more closely related to computational fluid dynamics models than anything resembling actual intelligence. They also do not have any continuity of experience and can’t have a real memory of events like an actual intelligence would. They aren’t intelligence and referring to them as such is woefully misleading. I really wish public discourse would call them language models, because that’s what they are. Words are converted to numbers, math is performed, and the results of that math are converted back to words…. That’s all.

    source