That's because it's just copying them directly from the images in the training set. It's not generating anything, it's just smoothing.It got really good at making images that look like they're just screenshots from a film/tv show, and the more imagery of the show there is on the internet, the better it does. It adds enough minor imperfections to make it all look very real without making too many mistakes that reveal it to be an AI image. I doubt any of these would stand out as AI generated if you just stuck them in an article.
View attachment 505620View attachment 505621View attachment 505622View attachment 505623
Sure, that's exactly what a hopeless pessimist with no curiosity or investment in the future would say.Using an absolute shitton of compute to generate a bunch of disjointed, lifeless filler scenes that subtly shift perspective for 3 seconds is only impressive to drooling retards.
Just a quick attempt. You'd need to clean them up and manually animate them I suppose. Could be a good way to do concepts at least.
View attachment 505743
I literally work on platforms for Conversational and Generative AI for use in call center applications, using the very latest models and APIs. I understand how these technologies work.Sure, that's exactly what a hopeless pessimist with no curiosity or investment in the future would say.
You saw how fast the image generation improved and yet you're still underwhelmed by the current capabilities of the video version and don't expect rapid improvements to it over the next few months? You're the drooling retard, dude. Go tell your cats why you're right. You'll be proven wrong by the time they lose interest.
I've seen things you people wouldn't believe. Attack ships on fire off the shoulder of Orion. I watched C-beams glitter in the dark near the Tannhäuser Gate. All those moments will be lost in time, like tears in rain.