• NotMyOldRedditName
        link
        fedilink
        English
        arrow-up
        5
        arrow-down
        2
        ·
        8 hours ago

        Hence the you’re just delusional because you really don’t understand what’s going on.

        • artyom@piefed.social
          link
          fedilink
          English
          arrow-up
          3
          arrow-down
          6
          ·
          8 hours ago

          You’re the delusional one if you think everyone and everything was coded with AI. That’s just incredibly ignorant.

          • NotMyOldRedditName
            link
            fedilink
            English
            arrow-up
            6
            arrow-down
            2
            ·
            7 hours ago

            I’m not saying it was entirely coded with AI, but AI has had some involvement.

            You clearly stated you think any use should be disclosed, as per the old steam policy, and that you wouldn’t use anything if it was disclosed as used.

            In some shape or form, it has been involved with PieFed, and it would require a disclosure if they were being honest about it.

            There are over 40 contributors on codeberg repository. That’s 40 people that you have to belive have never asked ChatGPT a question about the work they were doing on it, that didn’t base a descision off the Google search result AI response at the top, that didn’t use any of the IDE code completion tools built right into the software they use to write the software.

            It’s just not happening.

            If you think you can use software like this that isn’t tainted by it, you’re just flat out wrong at this point and moving forward in the world.

            • artyom@piefed.social
              link
              fedilink
              English
              arrow-up
              3
              arrow-down
              5
              ·
              7 hours ago

              AI has had some involvement

              it has been involved with PieFed

              You keep suggesting this but you can’t provide any evidence whatsoever. And until you do I’ll continue to reject your nonsense.

              • NotMyOldRedditName
                link
                fedilink
                English
                arrow-up
                8
                arrow-down
                1
                ·
                edit-2
                7 hours ago

                Here there’s talk on their repository about using an AI detection service (which is AI) to try and flag AI.

                https://codeberg.org/rimu/pyfedi/issues/605

                I tried a 3B version of the same model and it couldn’t detect that the sample text that came with the repo was AI. So we’d need a GPU capable of loading an 7B model which means at least 8 GB of RAM. afaik the demo script does two detections as examples of what’s possible but either one will do the job.

                They were talking about a service they thought maybe they could even have run on your machine.

                Doesn’t matter if they used it or not in the end, AI was used in the development process.

                Tell yourself whatever you need to so you can sleep at night, but of the 40+ people contributing to the project and they’re openly talking about AI on their repository, well I can’t help you…

                I’m expecting no reply to this message as it would constitute you using a service that involved AI.

                Edit: and pay attention to the names in that thread and who owns the repository.

                • artyom@piefed.social
                  link
                  fedilink
                  English
                  arrow-up
                  1
                  ·
                  4 minutes ago

                  Doesn’t matter if they used it or not in the end

                  Uh…that’s the only thing that matters.

                  AI was used in the development process.

                  No it wasn’t.