The actual research paper https://arxiv.org/abs/2603.18030

Also it pointed to where it is on github which I think means its possible to just tell chatgpt to tell me how to run this on my PC and then play with it, looks like you can just toss in anthropic/openai/etc api keys and go on your way with a POSIX controlled ai managed by any os https://github.com/kehao95/quine

Oooh I just realized you could probably schedule it to fetch content for yourself, if a web browser lets you use it via api. This would be like the chatgpt feature that is basically scheduled prompting but since this lets you mess with real hardware and software you can have a shell command that lets you fetch a “web surf internet rabbit hole”

This would be the way to avoid algorithmic nonsense because you can specify or have it infer your values based on content on your PC and it would do the job of an algorithm but with asuch or as little control as you want and not be beholden to any platforms

#!/usr/bin/env bash set -euo pipefail

mkdir -p “$HOME/digests”

cd “$HOME/agent-runs/web-rabbit-hole”

export QUINE_MAX_DEPTH=2 export QUINE_MAX_AGENTS=4 export QUINE_MAX_CONCURRENT=1 export QUINE_MAX_TURNS=20

quine " Fetch recent content from my chosen sources. Look for threads related to:

  • interspecies communication
  • lucid dreaming / dream yoga
  • NixOS
  • AI agents as OS processes
  • federated media

Write:

  1. a short digest
  2. the best 5 links
  3. one rabbit-hole path worth following

Save it to $HOME/digests/$(date +%F)-rabbit-hole.md. "

  • confuser@lemmy.zipOP
    1·
    6 hours ago

    Man I wish I was an LLM in this case, it is pretty lame to get spat on for sharing something I found interesting and useful to other people who I know would find this interesting and useful. The incoherence of the post is just rambling without wording things nicely

    • Arcanoloth@lemmy.ml
      6·
      6 hours ago

      I know I come across like an asshole here, and I don’t want to hurt feelings, but, in case you are human, you should really consider your impulse control, writing coherence, and, maybe, how a given topic is usally regarded by the intended audience. In short, next time, please stop and think before posting something as emotionally charged as that, ok?

      • confuser@lemmy.zipOP
        2·
        6 hours ago

        Oh I consider it quite well usually, just I didn’t care to do it this time because I figured it is reasonable for people to click the 2 links to figure out what this is if for some reason the title wasn’t enough and then they could read the rest if they wanted to but I guess this is too much to ask of people nowadays since we are too used to only reading things that are nice.

        This was not an issue with other communities I have shared this with so I have to believe that the Lemmy community is just built different I guess.

        There was nothing emotionally charged about the original post it was just thinking out loud.

        Thankyou for actually being somewhat reasonable

        • Arcanoloth@lemmy.ml
          3·
          5 hours ago

          Also: I apologize for my wording. I had an extremely strong emotional reaction, and I shouod have stopped and thought.

          • confuser@lemmy.zipOP
            1·
            5 hours ago

            Thats OK we all get buzzed with buzz words haha.

            There is one thing that I am curious about though, do you think you’d use the thing in the original post?

            Wanting to know if the utility has landed for at least one person.

            • Arcanoloth@lemmy.ml
              2·
              5 hours ago

              Thx. No, I have yet to see anything from an LLM that I consider remotely useful. And I wouldn’t be caught dead leting one anywhere near any system integration (Just look at all of the AWS outages in the past months). I continue to remain utterly baffled whenever anyone uses them for anything (other than maybe image generation, but I’d rather pay a small artist for those); I mean, their “summaries” are just as much random garbage as their citations… They look reasonable, but that’s all they do. I think LLMs are a neat parlor trick that is going to cost us, as a species, a lot, without getting us any closer to AGI.

              • confuser@lemmy.zipOP
                1·
                5 hours ago

                Hmm I think aws outages would not be an issue if a local LLM or a self hosted one was used, this seems like the kind of thing that lends itself well to this actually for once.

                I don’t think image generation is remotely useful in this format because I think it would be best used to just fetch things in a way that is like a universal fuzzy search.

                I think the quality of the summaries and citations found are more to do with prompt skill than the ai itself, like for example we can look at how my original post caused so many people to have a visceral negative reaction where it was unwarranted just because it was not to their liking, but the content itself is actually somewhat reasonable.

                I think nobody will settle on what agi actually is so I consider this not worth thinking about if the topic is restricted just to workflow things.

                Something I just realized is that this can be used on android with little to no adjustment and it could even use apps to varying degrees

        • Arcanoloth@lemmy.ml
          3·
          6 hours ago

          I’m trying. For the record: The global-warming-increasing tech-bro-billionaire-owned LSD-tripping babble-automatons are not exactly universally applauded. Saying anything about them requires care.