4 ms·
I've been dabbling with text-to-image generation quite a bit lately. I also do graphic design, drafting, and occasionally doodle with a pen or pencil. I really
by crumpled 4y ago
I've been dabbling with text-to-image generation quite a bit lately. I also do graphic design, drafting, and occasionally doodle with a pen or pencil.
I really enjoy generating images, and I really enjoy drawing even though I'm not really good at it. I like to draw with my five-year-old, but I've also exposed him to Dalle-Mini. I think showing him dalle-mini was a mistake, and I'm going to avoid repeating it. (We are a low-media household.)
My concern about AI art has to do with children learning art skills. Will it be hard to encourage developing art skills? If a person can just ask for an image rather than learning how to make one, will they be motivated to do the work to build the skill? Or will it suddenly become a waste of time to the young mind?
- elefanten 4y agoIs that much different from being able to find ready-made images to look at in various media? I could see a difference in the case of wanting to see a custom image and being able to summon it via description vs. needing to learn to produce it. I’m not particularly artistic and never developed any such skills… so I don’t know: how crucial is that particular motivation to those who acquire art (production) skills?
- crumpled 4y agoI think you described the difference pretty well. If you just want to look at an image of a thing, finding one is good. You might even find a real picture of something. ( In the future, this isn't a given. ) Plus you'll have to filter out images that are irrelevant, or even harmful. If you need to use the picture for something, there's a lot of advantage to crafting your own picture. (licensing, privacy, attribution, ownership, et al) When I wrote my comment I was trying to think of a parallel from the past, and I kept thinking of calculators. An even better example would be the decline of hand writing as a skill.
- syntheweave 4y agoThe difference between seeing art and drawing art is essentially one of a knowledge capture skill. When you learn to write, you have to copy the shapes of the letters. Often this is first done by tracing the lines or filling in an outline with color. Then you learn to reproduce it on your own, and eventually have it memorized. The same is basically true for drawing, but you're doing more of the early parts of the process more often. To draw a figure from imagination you have to learn the shapes. To do that you have to find a real figure in pose, and carefully copy it to get the proportions of the shapes as accurate as you can(which is assisted by various measuring techniques to correct your work). Then, after a few thousand of those, you try to draw an imagined figure and voila, your hand makes recognizable shapes - maybe a bit simplified, out of proportion, or with poor anatomy, but something others can parse correctly. And so doing a lot of drawing is like building up a vocabulary, becoming more observant and aware of subtle visual differences and therefore more able to communicate them back. I can't see AI replacing this: it can be a good servant and give a lot of options, but it can't mind-read that you want a hand posed slightly differently.
- astrange 4y agoYou can ask for an image, but that's not a physical object, it's an image. The thing that creates physical objects is called an inkjet printer and already exists. Neither of them create paintings though. So your kids are safe to learn that.