Spaces:
Running
on
Zero
Everything after 2.1 kills intrigue rather than opening it up
As usual with later SD, this is a boring model. Contrasted from SD2.1, it's a devolution. This model "wants" to emphasize and snap back to banality. If you consider SD generations to be an exhortation for the viewer to think about something or experience an aspect of the world, the sensation I get from 3.5, and XL, is "shutdown". Or that I'm inside a flotation tank filled with perfectly pH-balanced saline. Nothing is out of place. It's like a just-so story. Why bother caring?
The SD 2.1 results are weird. There's crap on the ground. Why? Where are they? What's happening? I'm curious.
This is the prompt/NP I used on both:
prompt: people with things in places, something major happens, high detail, in the foreground is a person
negative prompt: undetail, line art
Note how open-ended the prompt is! 2.1 is capable of putting trash on the ground and making lopsided, multifaceted landscapes when you don't specify one way or the other.
These are some 2.1 results. I did a little selecting but I don't think that matters.
2.1 has the capacity to expose me to something I don't understand, just for "people with things in places"!
These are some SD 3.5 Large Turbo results. I didn't do very much selecting because you instantly are aware of "the clock ticking" when you're on Zero hosting! If I had somehow selected for the worse, it really doesn't make any difference. Additional substantiation for the obvious faultline that I'm asserting can be found with a few minutes' generating.
a boring market
a boring parliament or lecture hall. Who cares?
commuters or something
an old lady in a plaza with pots doing exactly what old ladies in plazas with pots do
Everything and everybody is role-appropriate. Everything in its right place. There is no possibility of surprise or the out of the ordinary. There is nothing in particular to think about. It’s a great model if you like lobotomies.
And yet, it’s on Zero hosting which means that the thing with WORSE results is MORE restricted. Bullshit. Who’s minding the store? Why is there a presupposition that as new things come out, they get better? Why is there an Emperor’s New Clothes disconnect that 3.5 sucks, yet is being portrayed as something BETTER, which therefore comes with restrictions based on higher resource requirements?
The spotlight moves along, to 3.5, 3.5 Turbo, then probably to 4, and whatever.
2.1 does more with less! Yet 2.1 is the one that is subject to neglect now, and it took the recent mutiny to bring it back.