No biggie! Tbh, the fact that I’m working in advertising means riding lots of waves of the ‘hot new thing’ all the time. The Metaverse being the current bandwagon…
AI is an interesting proposition but the type of polish and flexibility my day to day work requires doesn’t make me fear for my job just yet. Not when you can spot something midjourney spat out from a mile away and everything looks like a fever dream oil painting, especially in concept art circles where everyone thinks they’re Craig Mullins now. IMO It’ll probably end up just being another tool that might help out part of the process, but certainly not be a ‘be all end all’ solution. Not just yet
Also curious how all this AI output eventually feeding back into itself will just end up being a singularity of sameness.
As long as any AI concept is programmed on pixel based interpretation I think this will not change in the near future. I cannot even imagine how complex it would be, but having an AI based on vector or even vertex approaches could definitely be more usable in these applications
So no 3D involved in art at all? And also no movies with CGI, Compositing or VFX? Also no digital art at all? You would be shocked if you would know how much AI is actually involved in enhancing, denoising, optimizing, simulating, etc.
I think I got something like 15 images. You get 4 variations with each prompt and you can upscale or evolve any of them as much as you want - either contributing to that image cost
I’ve started to see the use of AI generated images used in media already - it can be an effective replacement for stock imagery in a lot of places. The ability to create a scene very useful if you’re writing a creative story etc.
There are definitely characteristics I see in the AI’s as you mention - I actually quite like some of this character, especially in the cruder AI’s, almost a personality or style! However with the rate they’re improving no doubt that’s not far off being nullified.
Have you seen the results from Dalle 2? They go way beyond oil painting fever dream.
Yes, Dalle 2 is pretty good. And you can iterate with it better as you can mask bits of the result and have them changed. It has some weird limitations though, like not being able to use people that are still alive in your prompts? Probably afraid of weird deep fakes
Speaking of stock imagery, we had some good fun trying to make AI come up with stock photos from scratch
A piece of advice: you can use 4 credits for a 600 iteration picture and 1 credit for upscaling it. This way you’ll only make 1 picture a day, but it will be much better.
As an artist, I find this AI-generated art genre interesting. Looking at the invention of the camera, and how that affected painters, which then led to some of the most interesting and unique painting styles and movements, maybe this could create a similar catalyst? Who knows? I am sure some painters felt threatened that this new machine (called the “camera”) would take their job of painting accurate portraits and landscapes, but others embraced it and felt freed up to venture into new styles. And the other side was that it also created the whole new art form of photography. More for everybody.
I have seen artists on forums dismissing AI-generated art, or worried that they will be out of the job, and I’m just not sure I see that happening. My suggestion would be to find a way to flip the narrative, and use it as inspiration, even if you disagree and want to rebel.
I have not yet used one of these AI generators, but there might be some uses for me. Maybe as an idea generator, or as parts/pieces for collage work. I am still somewhere in the middle and am trying to see it from both sides. I just don’t think AI-generated art is the end for artists, or really that AI anything is the end of…everything. Well not yet anyway.