Oh dear. What strikes me is not just the lack of any consistency of style, which you noted, by the lack of any sense of objects. It's not like it is playing with object, distorting them deliberately. It just doesn't know what an object is, what shape it is, or how the pieces fit together.
Yes, and that seems to be a limitation endemic to Nightcafé specifically. but since I'm locked out of Dalle2 and I don't want to pay for Midjourney I have to work with what I have. Perhaps there are specific styles associated with specific AI engines?
I actually think the muddled mess aesthetic works really well with the song's disjointed, fragmentary lyrics. Illustrating a more traditionally-written song with these kind of images might be a disaster.
Indeed. I'm amused by how the AI tries so hard at some points, yet fails spectacularly at creating anything coherent. The image for "Calm" worked out pretty well. But what is with those people with three or four rows of teeth??
I think at best AI art can induce lols, at worst (and mostly) it’s extremely disturbing and indescribably lacking.
I’m interested to see controlled studies where individuals are asked to discern AI art from human made art. I’d hazard we’d be able to most the time but this may change. Most of those images and other DALLE images I’ve seen have these weird curves everywhere.
Oh dear. What strikes me is not just the lack of any consistency of style, which you noted, by the lack of any sense of objects. It's not like it is playing with object, distorting them deliberately. It just doesn't know what an object is, what shape it is, or how the pieces fit together.
Yes, and that seems to be a limitation endemic to Nightcafé specifically. but since I'm locked out of Dalle2 and I don't want to pay for Midjourney I have to work with what I have. Perhaps there are specific styles associated with specific AI engines?
I actually think the muddled mess aesthetic works really well with the song's disjointed, fragmentary lyrics. Illustrating a more traditionally-written song with these kind of images might be a disaster.
I like it. Looks like something that the transporter in Star Trek coughed up. Or Reanimator or From Beyond.
Indeed. I'm amused by how the AI tries so hard at some points, yet fails spectacularly at creating anything coherent. The image for "Calm" worked out pretty well. But what is with those people with three or four rows of teeth??
Wow. Spooky.
I think at best AI art can induce lols, at worst (and mostly) it’s extremely disturbing and indescribably lacking.
I’m interested to see controlled studies where individuals are asked to discern AI art from human made art. I’d hazard we’d be able to most the time but this may change. Most of those images and other DALLE images I’ve seen have these weird curves everywhere.
The hype is sure strong with AI, but yes, it still has a long way to go before it can reliably produce reasonable-looking imagery on command.