• 2 Posts
  • 17 Comments
Joined 2 years ago
cake
Cake day: June 11th, 2023

help-circle


  • I like this license. I think some people in this comment section do not understand the purpose. An (open source) license is a tool that is used most effectively in determining what kind of entity uses your software. Use an mit or bsd license and everyone will use your software. Use a gpl license and large corporations might use it sometimes, but wont if there are alternatives and most likely wont incorporate it into any of their products.

    This license might not be fully enforcable, but it is 100% poison to companies. Using this license(especially with the copyleft module) will almost certainly guarantee that only small projects will ever use it. Sure, it will kneecap your growth and your code will never see any real usage, but for some projects that might just be the best option.






  • I disagree with your points fundamentally, and I believe the difference is in how we interpret both art and the creation of art. I do not believe that a prompter is able to convey enough intent for it to count.
    This could be compared to someone commisioning a drawing for, for example, a story. The story and direction they give, that would be the prompt or what lead to it, in this case, would display their intent. The drawing itself, however, would not display their artistic vision, but that of the artist they commissioned to draw it. Now, they might coördinate with said artist to get their visions to align as closely as possible, but as I said, models have no vision, and so none can be aligned with. You could ‘find’ an image generated by such a model that aligns with what you wish for, but there is no intent behind it.

    The environmental damage is inherent to the technology, as matrix multiplications are inherently not very efficient, and any given model runs a lot of them. Running a model at home seems more efficient because you only generate for yourself, but if every user of diffusion where to do this, the cost would not be better.

    I do not understand what you see in the video you sent me. It does not, to me, seem to carry a message. Sure, some of it’s imagery can be aesthetically pleasing, but I cannot interpret it as carrying any meaning.

    Oh, and dw, I did not mean to call you stupid, I think the ideas about art you have specifically are stupid. That does not necessarily carry over to any other part of you.


  • I judge art on the basis of three things:
    The intent of the artist,
    The context surrounding the art,
    My own interpretation of the art

    A stable diffusion model is not much more than a set of statistical functions executed over a large array of numbers. Therefore, the model cannot have intent.
    The use of the model to generate images damages the environment, makes use of work made by artists who, by design, cannot be credited for said work, and no or very little artistic effort went into the generation. Therefore, the context is pretty loathesome.
    The third point depends on the image, although I find that most images do not have much in the way of creativity or artistic direction, and come off as “bland”, “samey”, “wrong”. The fact that there is no intent makes it hard for me to read intent. Therefore, my interpretation is usually not very favourable.

    These are my thoughts. I believe your ideas about art and how we should judge it (which is what you are prescribing) to be quite stupid, but you live your life however you want, I guess.