|
Sally Sprodgkin posted:lol. lmao didn't stop me TIP posted:"musclebound Abraham Lincoln making bedroom eyes tintype photo"
|
# ? Aug 3, 2022 05:01 |
|
|
# ? May 29, 2024 21:06 |
|
TIP posted:I couldn't make "sexy abraham lincoln" or "seductive abraham lincoln" and describing either of those as porn is pretty ridiculous Yeah. I mean that was the joke with "man getting hosed down with ranch" and my mayonaise man. The joke was cum. It was funny seeing which one made it through. Also "bodyslam" and "piledriver" were not acceptable terms. Looks like Wombo can spell (print your prompt in the art randomly). "synthwave van" Weirdly just threw back odd looking tape decks and stereo equipment in neon hues. "hip hop rap music"
|
# ? Aug 3, 2022 05:12 |
|
Bloody Hedgehog posted:
I really like those. More upscaling and filters would have given it a sharper, more defined look I guess? edit: The other day I was thinking about how 40k-esque these new image-AIs are. Like, finessing the algorithm to get it to produce the image you want feels like an odd combination of technical know-how (which filters/commands cause what to happen) and associative imagination, completely independent of understanding how the code is working on a fundamental level. Sometimes with my choice of words and attempts at creating a certain lighting or mood, it feels a bit like a tech priest waving incense censers at a piece of tech, while other times I seem to stumble on a "magic word" whose associative significance to the AI is lost on me. That's my impression as a code-illiterate person who's really fascinated by this stuff, anyway. I prompted this one with something like "complex dorter trade city, antiquity fantasy style, ivalice, Akihiko Yoshida style, dark fantasy style, uphill street." For me the most interesting thing about Midjourney is that a lot of it is just experimentally throwing in terms via your associative imagination and seeing what the magic well throws back at you. I'm getting better at manipulating it to get the type of images I want, but the learning process itself, and occasionally just going wild with strings of adjectives, is really magic. I was thinking writing a forums CYOA-style game/story could be really fun using Midjourney. Also, I read the bit about the "creative commons" in the online manual. It seems like as long as you're not a company making $1 million+, you could for example use these images combined with something like the makeplayingcards.com website and publish a board game using this art with no issue? (Assuming you're not using images from a specific franchise, I mean.) Ceramic Shot fucked around with this message at 05:32 on Aug 3, 2022 |
# ? Aug 3, 2022 05:21 |
|
Tried more food today also doesn't work work, but still adorable
|
# ? Aug 3, 2022 05:24 |
|
I was trying to get midjourney to meld two images for so long last night. I ended up trying all sorts of combinations of random poo poo, mixing the style levels and image weights of the source image prompt, trying different styles and detail/realism prompts etc, additional text prompts to try add weight to stuff I wanted in there. Everything came out looking fucky. Eventually I just punched in 'x dressed up as y' and it worked perfectly. My 1990s computer toucher brain simply cannot comprehend natural language parsing in computer systems. yes I know the language parsing still sucks a lot of the time
|
# ? Aug 3, 2022 06:21 |
|
Sally Sprodgkin posted:I was trying to get midjourney to meld two images for so long last night. I ended up trying all sorts of combinations of random poo poo, mixing the style levels and image weights of the source image prompt, trying different styles and detail/realism prompts etc, additional text prompts to try add weight to stuff I wanted in there. Everything came out looking fucky. Funny you mention that, it's how I finally got a dog with human hands after god knows how many tries: "dressed up as a baker" I'd done a lot of these but hadn't noticed any hands before, but once I saw them on a pig lawyer I knew it could be done
|
# ? Aug 3, 2022 06:28 |
|
AARD VARKMAN posted:Tried more food today.
|
# ? Aug 3, 2022 07:37 |
|
No more steroids grandma! The Cringe Man Ogdred Weary fucked around with this message at 09:32 on Aug 3, 2022 |
# ? Aug 3, 2022 08:52 |
|
|
# ? Aug 3, 2022 10:28 |
|
What will be interesting to see is if things that are generated by this AI tool get used in media (games,comics,movies,tv,album covers)
|
# ? Aug 3, 2022 12:50 |
|
Mr Ice Cream Glove posted:What will be interesting to see is if things that are generated by this AI tool get used in media (games,comics,movies,tv,album covers) I feel like the novelty value alone will kickstart its relevance in broader media. Definitely don't envy graphic designers now, but I'm kind of giddily anticipating what kinds of applications people way smarter than me will figure out. I've read that other types of AI are already involved in things like calculating actors' "star power" vs. cost and optimal release dates. It'll be no surprise when they make stuff like focus groups look ridiculously primitive. Is it just me or does this kind of feel like a pandora's box moment like "the internet" was? edit: here's the Creative Commons related stuff in Midjourney's "rules" tab on their Discord channel: If you are not a Paid Member, Midjourney grants you a license to the Assets under the Creative Commons Noncommercial 4.0 Attribution International License (the “Asset License”)... If you are an employee or owner of a company with more than $1,000,000 USD a year in gross revenue, and you are using the Services to benefit your Employer or company you must purchase a corporate membership plan to use the Services or copy the Assets for your company. Seems ok for small independent types to be using in their products? I'd love to hear a real lawyer interpret this though. Ceramic Shot fucked around with this message at 13:39 on Aug 3, 2022 |
# ? Aug 3, 2022 13:06 |
|
Mr Ice Cream Glove posted:What will be interesting to see is if things that are generated by this AI tool get used in media (games,comics,movies,tv,album covers) I have a feeling it will be everywhere in 10 years and overused in uninteresting ways. The same way low poly 3D models were slammed into everything in the early 2000s. We figured out where when and how to use them in TV movies and games, but that took time. People just discovering AI art for the first time constantly miss flaws that people that have been playing with them for 4+ years already see instantly. I can't wait for a bunch of TV shows that we look back at as basically N64 games. A few amazing gems so many that were fun rides buts completely unplayable now and stuff that was just experimental and didn't work in the first place. There's a place for AI art, either assisted or just raw expression of the machine. Something is going to take off made by an AI and everyone will dogpile in trying to do AI movie in the cheapest most repeatable way.
|
# ? Aug 3, 2022 13:18 |
|
In 10 years this is going to be unrecognizable. 3D, realtime, personalized. "Make me a new season of The X-Files, 1997" It's going to get crazy.
|
# ? Aug 3, 2022 13:32 |
|
Mola Yam posted:In 10 years this is going to be unrecognizable. 3D, realtime, personalized. "Make me a new season of The X-Files, 1997" It's going to get crazy. That's exactly my point it will be barely working in that sense and people will then air that as the new season thinking it's good. But when you look at it in even 15 years it will look like the janky garbage it is that vaguely resembles the X-files but when you look closely at anything is obviously not. Molder just slowly morphs into Skully while Skully morphs into Molder for no reason mid conversation which model is producing which voice changes as it ticks to 51% the other character, maybe there's a part where there's 2 51% Molder's and they both talk and Skully is mute because she doesn't exist in this context. Really weird and strange things that you have to actually watch and not just listen / background watch to notice. And no this wont be part of the plot it'll be a weird thing that the AI just does. If you watch the image AI and let them run way too long you get some really really weird stuff going on.
|
# ? Aug 3, 2022 13:36 |
|
Everything is so based on existing art (obviously), I'm curious if you combined the text ones with the image ones if you could get a pitch for a new IP with concept art and plot ideas, from a well worded prompt. And then theoretically even scripts, though a lot of the text stuff I've seen seems to forget what's happened or who is in a given scene, so there's probably more sophisticated models to come.
|
# ? Aug 3, 2022 13:42 |
|
I appreciate your skepticism, but it's hard not to see at least a significant economic disruption in terms of creative-service industry jobs. I'm really mediocre at visual arts, but with the right prompts I can now get effectively unlimited professional quality art for $30/month. Lately I've been designing analog board games as a hobby, and now, with the greatest of sympathy towards artists, I nonetheless see no reason to take a chance on a "human element" when I can produce multiple high-quality pieces of media in about 30 seconds as long as I spend a modicum of time researching the right terms/filters to finesse the AI into giving me what I want.
|
# ? Aug 3, 2022 13:46 |
|
gently caress You And Diebold posted:I like in Shadowrun: hong kong where poetry chat bots keep gaining sentience and going rogue so there is a ban on making poetry bots for your chatroom I put that through MJ without any additional settings. I've been curious about seeing what DALL-E does when you just feed it an image without a description for a prompt. I wanted to see how much creative license it would take with the composition. Using those two images, it gave me these: A nice little bonus is that google translate returns a handful of phrases when you import those last three. It's mostly gibberish, but sometimes amusing ("After birth, cobalt is born into a super shortbread. Can it breed?"). Zombie Killer 5000 fucked around with this message at 14:02 on Aug 3, 2022 |
# ? Aug 3, 2022 13:57 |
|
|
# ? Aug 3, 2022 14:24 |
|
AARD VARKMAN posted:Tried more food today As someone from Philadelphia, those actually look accurate; the meats look dry as shoe leather. <3
|
# ? Aug 3, 2022 15:12 |
|
"two cute emo girls playing Monopoly in their dorm room. the room is decorated with posters of 80s bands and christmas lights. realistic photograph."
|
# ? Aug 3, 2022 15:56 |
|
precision posted:"two cute emo girls playing Monopoly in their dorm room. the room is decorated with posters of 80s bands and christmas lights. realistic photograph." I'm the mongay poster in the background
|
# ? Aug 3, 2022 16:01 |
|
One of more of those girls may also be members of the Borg, not sure what it was trying to do there lol
|
# ? Aug 3, 2022 16:05 |
|
I like the homebrew diy version of Monopoly
|
# ? Aug 3, 2022 16:12 |
|
gentlemen, i bring you... THE FUTURE! "a screenshot of the final build of Star Citizen"
|
# ? Aug 3, 2022 16:15 |
|
"a movie poster from 1983. the title of the movie is "Revenge of the Jedi". it depicts an evil Luke Skywalker"
|
# ? Aug 3, 2022 16:19 |
|
Not really on par with other brilliant stuff posted, but this captured exactly what I was picturing perfectly.
|
# ? Aug 3, 2022 16:55 |
|
Durzel posted:Not really on par with other brilliant stuff posted, but this captured exactly what I was picturing perfectly. Lower right is just having a small nap. He is using the laptop to warm his always-cold diabetes feet.
|
# ? Aug 3, 2022 16:57 |
|
|
# ? Aug 3, 2022 17:07 |
|
Durzel posted:Not really on par with other brilliant stuff posted, but this captured exactly what I was picturing perfectly. love it. I need to spend more time researching different mediums to mess with, a lot of my favorite stuff has come out of trying a simpler concept with a particular artist
|
# ? Aug 3, 2022 17:08 |
|
|
# ? Aug 3, 2022 17:17 |
|
precision posted:gentlemen, i bring you... THE FUTURE! Got inspired
|
# ? Aug 3, 2022 17:33 |
|
prompt was 'a screaming man in a garfield suit'
|
# ? Aug 3, 2022 18:25 |
|
DALL-E apparently has a very firm idea of what it thinks I'm eating for lunch
|
# ? Aug 3, 2022 19:01 |
|
"A robot gentleman throwing a sixpence to a poor disheveled human artist in the streets of the multiverse of sadness"
|
# ? Aug 3, 2022 19:44 |
|
Sally Sprodgkin posted:They're all like that to some degree, I think they are hyper sensitive to the picture-AI version of the 'hyper advanced AI chatbot immediately becomes a racist nazi' headline that put a bullet in that chat AI on twitter a while ago. Craiyon uses open source and open models, they don't have any restrictions. Stable diffusion supposedly is going to release their code and models later and it has already shown that it's able to create portraits that look like actual people.
|
# ? Aug 3, 2022 19:45 |
|
|
# ? Aug 3, 2022 20:56 |
|
well, i respect the effort anyway: one hundred different Dragonball Z characters standing on a flying turtle in the style of leonardo da vinci
|
# ? Aug 3, 2022 21:35 |
|
Ceramic Shot posted:Is it just me or does this kind of feel like a pandora's box moment like "the internet" was? we are seeing nothing less than the automation of intellectual labor. artists and soon people who right code, just aren’t going to be as necessary to businesses. people are rapidly not as required by those who control the economy.
|
# ? Aug 3, 2022 23:49 |
|
naem posted:we are seeing nothing less than the automation of intellectual labor. who's to say that being able to come up with the exactly correct and detailed prompts to create art isn't just as much a rare talent as creation of art in the first place? because like, as an artist myself, when you think about say music. the issue with making good music has almost nothing to do with being talented at playing any instrument. many people learn how to play an instrument very well but can only "write songs" in the sense that they can string together chords and notes. so, hopefully, you know, it shifts to a situation where good artists are the ones who can most intelligently utilize the technology - by, for example, having the artistic capacity to type "orson welles dressed as sailor moon, realistic photograph"
|
# ? Aug 4, 2022 00:14 |
|
|
# ? May 29, 2024 21:06 |
|
Abhay Khosla of Abhay Khosla's Bram Stoker's Dracula has been making comics using MidJourney and the results are pretty good. fez_machine fucked around with this message at 00:26 on Aug 4, 2022 |
# ? Aug 4, 2022 00:18 |