I have a feeling this is AI-generated, to be honest. I first found it on a German site IIRC, but couldn’t find any significant matches apart from that.

I’ll try to avoid posting AI art in future, but I have to admit… some of it is pretty flippin’ impressive!

EDIT: No, I guess I’ll post more AI art in future, based on the positive response and discussion below. I’m frankly a little torn, but I’ll abide.

  • Piecemakers@lemmy.world
    link
    fedilink
    English
    arrow-up
    1
    ·
    edit-2
    1 year ago

    Again, I appreciate your candor and genuine interest in discussion (as an OG from that other place, this is a breath of fresh air, frankly), and I hear what you’re saying. My background is also in art from back when Photoshop was first offered as collegiate courses and that alone was a divisive aspect for a hot minute — though that too quickly normalized, and the detractors either knuckled down to learn the new skills or simply found other things to protest, I guess. I suppose what I’m getting at in all of this is less of a judgement of AI’s current capacity and more a forecast of its likeliest future forms, both technically and culturally.

    Since very near the beginning of this boom, I’ve been fiddling around with various AI tools in my free time (not just graphics), and work in 3D sculpting/printing, so the whole landscape is supremely interesting to me on a number of levels.

    In truth, I feel that it’s more a question of process automation than spontaneous generation in that a finished, viable piece is not attainable by a simple word string or “barest of keywords”, all due respect — though I completely understand that it can seem easy and unchallenging to go from text to marvelously vibrant dynamism in ultra detail. The fact is that the vast majority of skillfully created “AI generated” artwork (graphical or otherwise) requires not only thousands upon thousands of hours of high level coding, database training, and testing, but even then the end-user is required to compose and refine each prompt in the pursuit of their vision, making innumerable creative choices along the way.

    Ultimately, I’m looking forward to seeing how this all plays out, and I’m excited to see what else you’ll share in that regard here, for sure!

    • JohnnyEnzymeOPM
      link
      fedilink
      English
      arrow-up
      2
      ·
      1 year ago

      Again, I appreciate your candor and genuine interest in discussion (as an OG from that other place, this is a breath of fresh air, frankly)

      Thanks, certainly, and same. Indeed, I guess one nice thing that’s come out of our little exodus to the FV is that there’s more positive, invested user-engagement, and less casuals & ‘herd activity.’ Haha, at least for now.

      Also, and just personally, it makes me feel honored that I was evidently the first to create a BD-themed community in the FV, and as a result, I have much residual goodwill towards the sub, my instance, and Lemmy in general.

      Anyway, it doesn’t sound like we’re far apart on the issues, moreso that we’re coming from different activity & experience POV’s. It also occurs to me, based on your explanations, that it might just be that the traditional concept of “artist” is less useful than formerly. Sorta makes me curious if William Gibson coined any good anticipatory terminology in his “Neuromancer” series (that I haven’t read for ages).

      1) The fact is that the vast majority of skillfully created “AI generated” artwork (graphical or otherwise) requires not only thousands upon thousands of hours of high level coding, database training, and testing, 2) but even then the end-user is required to compose and refine each prompt in the pursuit of their vision, making innumerable creative choices along the way.

      1) Right, for sure. Those folks are like arch-mages to me, or whatnot. Indistinguishable from magic. :P

      2) “Innumerable,” really? I’ve never really got that impression from the MidJourney posters, but what do I know? I’ve only played with the lower-tier free AI’s so far, and despite the quality being relatively crap (plus random-ish), it was pretty clear who was doing the heavy lifting there. Do you know of any good examples (like, articles) of people talking about the process of refining their prompts and settings?

      • Piecemakers@lemmy.world
        link
        fedilink
        English
        arrow-up
        2
        ·
        1 year ago

        Maaaan, you’re gonna drop Gibson and Clarke in the same reply? 🤣🤘🏼 I like your style, choom.

        As for pt. 2, I’ll have to look in the morning, but I’m pretty sure I’ve got a couple bookmarks I can share (though related Discord communities [prompt honing, training model techniques, etc] have proven surprisingly supportive and more than a little elucidating, really)

        • JohnnyEnzymeOPM
          link
          fedilink
          English
          arrow-up
          2
          ·
          1 year ago

          Haha, thanks. I’d completely forgotten where that quote came from, but me likey.

          Funny, I just randomly discovered this little project, which was evidently built with Stable Diffusion + ControlNet.

          So this is getting to be stoopid, my gawking on the sidelines like this, but I need to revive paypal or something and try this stuff out…