Artificial Intelligence
- Started
- Last post
- 1,637 Responses
- milfhunter1
Make-A-Video is a state-of-the-art AI system that generates videos from text.
- uan3
- drgs2
- utopian0
- Here's how the best AI art generators compare: https://www.creative…utopian
- Yeah this was posted beforegrafician
- and all of them can do 1000+ variations for the same prompt. (would be interesting to see the actual number calculated, at some point they must output the same)uan
- grafician-4
- The thing with these is you can only make like one as an artist - then on to the next thinggrafician
- But now expect to see TONS of these from other artists too
twigs was always a trail blazer...grafician - still talking to yourself i see..._me_
- _me_ still downvoting for no reason LOLgrafician
- this is one of her caprisongs, the rest of the vids were released months ago so this one must have taken a whilekingsteven
- yup, uses features that were introduced recently in SD I guessgrafician
- Beeswax4
Adobe will buy one of these platforms and make it part of the creative cloud where you will be able to generate layered images based on your prompts and then tweak it using regular tools.
- I would welcome that. They should scoop up Stable Diffusion and integrate it in PhotoshopNBQ00
- Ai AI captainkingsteven
- Already plugins doing this lolgrafician
- https://www.youtube.…PhanLo
- what plugins, graf? (except stable diffusion)?NBQ00
- @NBQ there's a Ps plugin and a Figma plugin to generate AI stuff
Look around the AI threads here.........grafician - Yeah, already been discussed ad nauseam here. They've been prepping their Ai for a few years now. It's called Sensei.garbage
- you can use beta.dreamstudio.ai for stable diff. and the plugin https://christiancan…
to access it from psuan - So once again, if you are in your mid 30s and have spent decades learning the things I'm sure most of us here know, you have wasted your time.garbage
- AD positions will be replaced by automation in our lifetimes.garbage
- Instead buying Figma for $20 billion, Adobe should've purchased an AI company.utopian
- I've been trying things over on Openart..
there's some invites if anyone's interested...
https://openart.ai/i…exador1 - @utopes It's in the thread. They've been building their own AI engine for years now. Myabe skimmed over because it looks like we're talking about Illustrator.garbage
- They will buy up and assimilate the competition, but if you haven't seen it, they've been working on Sensei for a long, long time.garbage
- You can trace back the history to their acquisition of Omniture in 2009. Anybody that thinks "oh if this scares you, you aren't talented enough"..garbage
- ..is just wrong. Adobe has spent over a decade building a job-killing Frankenstein, fueled by creative input that isn't just free, but paid for by the cloud.garbage
- drgs2
- https://media.tenor.…palimpsest
- this gives me hope. for my career.shapesalad
- PioneerDJ3031
- the shadowing is amazingutopian
- You know how some dogs are breed to be cute, but actually it handicaps them - I feel like this with her.shapesalad
- lol shape. Yes she's stuck in a 2d pic forever.NBQ00
- https://twitter.com/…PhanLo
- grafician-3
Almost certain all those beautiful characters/faces in MJ and SD are from artstation characters and shopify fashion shops posing and face models included in the LAION-5B dataset
You can pretty much check for yourself anyway here: https://rom1504.github.io/clip-r…
but overall from all the tests done, unless you include "model" "character" or "artstation" in your prompts you're not getting anatomically corect anything worth generation/usage...
- Also all images were included at 1:1 pixel ratios so 90% of generated images are cut or out-of-framegrafician
- Not to mention you generate dozens of images for ONE useful or sometime for some prompts hundreds and nothing useful
A LOT of time wastedgrafician - Also using seeds and variations is useless 90% of times (depending on the model) you get 90% of the same seed image (unless u retrain or img2img)grafician
- So yeah, the tech for now is just for fun, with clearly BIG limitations for commercial usage in any image generation capacity for studios/art directorsgrafician
- i'm pretty sure some are the best of thousands of images. you get 10 images per min running SD on a 3090. much easier to tweak prompts as you gokingsteven
- The idea is not to tweak prompts too much, but to hit the "generate" button hundreds of times for wildly various results and IF a few are actually usable...grafician
- But agreed SD is more accesible, but also DALL-E has the same issues so I can generalize easilygrafician
- yeah. i love it but i'm not sure who these people are hyping it for commercial use. i think about how two generations ago was trippy animal faces on fruitkingsteven
- now we've got something that's a weird creative collage machine. as i mentioned before the possibilities for inpainting with this 'generation' of AIkingsteven
- that sort of sideways development of A.I. assisted tools is really only possible since SD so it will be interesting to see what is possible in a yearkingsteven
- i know a few guys using forks of SD that don't even take text prompts, others making body horror with filters disabled. it's going to get real weird.kingsteven
- drgs0
- PhanLo3
- https://www.youtube.…imbecile
- ^Pretty much. :-)PhanLo
- Actually many of these prompts seem to be banned on Midjourney. I once tried "sensual" and it said it was banned. lolNBQ00
- grafician-1
Wow
https://www.reddit.com/r/StableD…
img2img gone wrong to video, but with fascinating amazing results
- Goes right with my conclusions that AI models so far are useless for commercial applications - unless taking inspiration as an art director some whatgrafician
- Just because it looks "cool" doesn't mean it's actually useful so A LOT of time waste - but sure, it's fun if you have the time to waste on itgrafician
- this sort of thing you just leave running overnight. i'm still not getting your 'commerciality' thing. i've not read anything but scare stories suggesting thatkingsteven
- it's for that purpose. it's a medium in it's own right.kingsteven
- This "commercial" usage:
https://www.creative…grafician - Jessica Walsh on using DALLE for a rebrandinggrafician
- thats marketing harnessing the scare of A.I. that validates itself by utilising the current gen of tech. might as well be a 1950s robot eating mashed potatokingsteven