[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/ic/ - Artwork/Critique


View post   

File: 1.83 MB, 1281x1040, 1694152072609771.webm [View same] [iqdb] [saucenao] [google]
6999671 No.6999671 [Reply] [Original]

New year, new thread. The krita experiments continue. Feel free to join or ask questions.

previous threads
>>6984652
>>6965971
>>6958183

>> No.6999675
File: 279 KB, 1686x578, 1695021037531951.jpg [View same] [iqdb] [saucenao] [google]
6999675

ignore the seething autists.

itt i want to focus more on making basic value or color layers to guide the output.

>> No.6999682

>>6999671
Her expression totally changed.
I can get why this would be impressive to somebody who just wants to make a pretty picture, but what about the artists who are trying to make things with actual intent? If I'm trying to create a picture with a girl's shocked expression cut-off by the image border, I would be quite frustrated if the AI just kept giving me an end result where the girl is smiling instead. Fine control is important, and you lose it when you rely on directing an AI to give you the final result.

>> No.6999692

>>6999671
Why do you keep making these threads when you know you are unwanted here?

>> No.6999696

>>6999671
>>6999675

looked much better before

>> No.6999714
File: 102 KB, 768x512, 1703154148030305.jpg [View same] [iqdb] [saucenao] [google]
6999714

>>6999682
it is completely trivial to change it back if you have the ability. you know that.
the point is to figure out how to integrate AI, not just to rely on it blindly. you're still in the mindset of doing things in one shot.

>> No.6999719
File: 529 KB, 520x779, 1677643365717407.webm [View same] [iqdb] [saucenao] [google]
6999719

>> No.6999735

>>6999671
I tried krita but its not working for me, I just don't know how to use it like that

I'll stick to Stable diffusion for now and just edit the output in clip studio

I'll wait for a better app that can be user friendly someday

>> No.6999739 [DELETED] 

there's some fairly new thing that you could try. i haven't tried it myself yet.
https://github.com/Stackyard-AI/Amuse
as for krita, i'm using it with an existing comfy instance.

>> No.6999743

>>6999735
there's some fairly new thing that you could try. i haven't tried it myself yet.
https://github.com/Stackyard-AI/Amuse
as for krita, i'm using it with an existing comfy instance.

>> No.6999748

>>6999739
is this local install or an always online app?

>> No.6999750

>>6999748
none of these are online.
they work locally after you set it all up.

>> No.6999753 [DELETED] 
File: 3.85 MB, 498x198, 1687426807503754.webm [View same] [iqdb] [saucenao] [google]
6999753

/home/tat/Videos/room.webm

>> No.6999756
File: 3.27 MB, 492x198, 1697300766609277.webm [View same] [iqdb] [saucenao] [google]
6999756

>> No.6999757
File: 3.85 MB, 498x198, 1680951989360159.webm [View same] [iqdb] [saucenao] [google]
6999757

>> No.6999759

>>6999671
>krita experiments
I am interested in using this for background, feels like a good skill to pick, how do I go about this?

>> No.6999761

>>6999759
https://github.com/Acly/krita-ai-diffusion
try to get this set up.

>> No.6999815

>>6999671
Anon, you should use the image generated as a base to go further with your own art or draw/trace on. Taking it as *the* finished artwork is what faggots who dont enjoy drawing but still pushes through their art endeavor like Shad Brooks did.

Like, take parts of it that looks right and use as references. That's what AI is for, to assist you. Not to autocomplete yourself so that all traces of the original artist is gone and replaced with mixture of machine trained slop.

>>6999682
He doesn't enjoy drawing. He just wants pretty picture out quick instead of experimenting. I found usage for AI like backgrounds (usually colors) and some details on clothing that I rarely draw (jeans, sneakers, for example). You proompt, then you see the 100 different proompt results, 90 of which are trash, and keep the ones that you could use, and then you use it back as reference and draw yourself so that the artwork remains cohesive and has your identity.

>> No.6999837 [DELETED] 
File: 353 KB, 720x730, IMG_20240101_065246.jpg [View same] [iqdb] [saucenao] [google]
6999837

>>6999671
Bitch ass spammer

>> No.6999839

>>6999815
>Anon, you should use the image generated as a base to go further with your own art or draw/trace on.
that is the idea anon.

>> No.6999844

>>6999671
This is the the only useful AI thread on the board so congratz.

Could you show one where the sketch is more defined? So far your uses have been an unfinished rougb sketch but id like to see how much of a more finished work it can translate

>> No.6999849 [DELETED] 

>>6999844
It's tripfag he prob made the other AI thread or his discord did since they seem to follow his ass

>> No.6999851

>>6999839
>that is the idea
time to execute.

>> No.6999853 [DELETED] 
File: 339 KB, 575x1088, IMG_20240101_070803.jpg [View same] [iqdb] [saucenao] [google]
6999853

>> No.6999860
File: 2.75 MB, 520x779, 1687500700913317.webm [View same] [iqdb] [saucenao] [google]
6999860

some basic experiments with pure color areas.
also i'll say that LCM/krita (the fast stuff) is way worse at recognizing difficult stuff.
like the mask >>6999714, it wouldn't render it right no matter what, but without LCM it worked. with bigger resolutions it would likely work as well.

>>6999844
i'm on it.
but the harder i try, the more my style shows through, which isn't necessarily a good thing for these threads, so i've been trying to balance that. but i might just put more masks on things from now on lol.

if you have lineart i can try to give it a try for you tho. it doesn't even have to be particularly clean

>> No.6999864

>>6999761
Does it have to be krita? Can it be CSP or PS?

>> No.6999868

>>6999860
in case people don't understand what is happening in that webm:
the first 4 images are the templates i used to make the rest. tried various settings/models to see how well it could follow the palette while still making up things on its own.
next up: more deliberate and specific templates. and maybe values first.

>>6999864
this is specifically a krita plugin. you also need VRAM to run this. dunno what the minimum reqs are though, i'm using it on a upper midrange gaming rig.

>> No.6999879 [DELETED] 

>>6999860
>/g/
also did your skills atrophied to the point where you have to trace every little thing
>your sketch falls below even beg at times

>> No.6999887

>>6999864
oh and i think there are PS plugins. but i'm not sure they are for this kind of "live painting" stuff. i dunno.
i wish to have it on CSP as well. it's what i'd normally use. krita has been awkward to get used to.

>>6999879
you're not wrong, i'm pretty far from my peak.
but you're confused if you think there's any tracing anywhere in any of these examples.

>> No.6999890 [DELETED] 

half your use cases when it goes about AI tripfag basically starts going into photobashing.
especially since you love it to do the rest for you with your shitty traces

>> No.6999892 [DELETED] 

>>6999887
Your peak is ending up as Chris for how much you can do
actual begs have made it and are better received then your shilling ass
though I could guess the how crippling thinking is for you at times

>> No.6999893 [DELETED] 

>>6999757
>beg ass retard wants show you his skill

>> No.6999897 [DELETED] 

>>6999675
>wants to edumacate (YOU)
>still needs to still learn color value/color theory
>scribbles look like they traced
Hhhhmmmm...

>> No.6999899
File: 1.66 MB, 490x248, 1689496197312585.webm [View same] [iqdb] [saucenao] [google]
6999899

>>6999897
>>6999893
>>6999892
i drew this while you were sperging out :)

moving on. let's see what we can do with this...

>> No.6999904 [DELETED] 

>>6999899
So tripfag is all you do is basic bitch anime girls that's a dime a dozen to see.
That's hardly even worth talking about honestly.
Try again when you have anything interesting to actually show

>> No.6999907

I’m glad you made your own containment thread. That’s progress. Now move it to /g/
No one wants your shit here.

>> No.6999913

>1girlwhitevoid just like all the data it's ripped off from
Top kek. The existence of Game maker means 999 shitty games from prebegs instead of just one.

>> No.6999915

>>6999899
These no construction outline meme anatomy scribbles you're feeding it are this "gotcha". kek.

>> No.6999954

>>6999899
this is interesting but whats with the obsession with putting that horrid face mask thing on top, looks uncanny and drops any appeal your image could have had

>> No.6999958

>>6999675
You're the seething autist here bro
You couldn't be bothered actually learning how to finish drawings so when ai came around you went apetismo mode and based your whole personality around it.
Duc will forever be a lazy cuck

>> No.6999960 [DELETED] 

pababababab

>> No.6999961

>>6999860
>the harder i try, the more my style shows through which isn't necessarily a good thing
Peak personality disorder

>> No.6999964 [DELETED] 

niggers tongue my anus

>> No.6999967

>>6999899
Wow riveting

>> No.6999968 [DELETED] 

>ai le bad even doe it makes artfags obsolete

>> No.6999970

ywnbaw

>> No.6999972 [DELETED] 

sude sude side

>> No.6999974

>>6999968
>black guy that fucks your wife le bad even doe it makes you obsolete

>> No.6999975 [DELETED] 

how is this board this slow were fonna get banned before reaching the get

>> No.6999976

TEST

>> No.6999977 [DELETED] 

>>6999968
>forgot your trip
see your vendetta brainless bitch ass to /g/

>> No.6999979 [DELETED] 

i am the black guy crakka now bend over

>> No.6999980 [DELETED] 
File: 53 KB, 894x894, 1693734710105.jpg [View same] [iqdb] [saucenao] [google]
6999980

>>>6999968
>>forgot your trip
>see your vendetta brainless bitch ass to /g/

>> No.6999981

>>6999974
>BWC le bad even doe it makes BBC obsolete

>> No.6999987

If you use AI in any way possible for your "work", kill yourself, genuinely.

>> No.6999988 [DELETED] 

>>6999977
posted about random tripfag that no one cares about again award

>> No.6999992 [DELETED] 
File: 63 KB, 622x893, IMG_9985.jpg [View same] [iqdb] [saucenao] [google]
6999992

>If you use AI in any way possible for your "work", kill yourself, genuinely.

>> No.6999996 [DELETED] 
File: 382 KB, 640x480, ecobbc.png [View same] [iqdb] [saucenao] [google]
6999996

suck my dick or something yeah

>> No.6999998 [DELETED] 
File: 13 KB, 214x235, 1701892113755.jpg [View same] [iqdb] [saucenao] [google]
6999998

win

>> No.7000003

>>6999981
>bWC
>5.7 incher

>> No.7000004
File: 81 KB, 187x255, tropical forest dweller.png [View same] [iqdb] [saucenao] [google]
7000004

>>6999671
>>6999671

>> No.7000008 [DELETED] 

>>6999998
swedish fail

>> No.7000011 [DELETED] 

>>6999998
No get

>> No.7000015

Try again next year

>> No.7000018

>>7000000

>> No.7000019 [DELETED] 
File: 130 KB, 800x999, 396.jpg [View same] [iqdb] [saucenao] [google]
7000019

>>7000015
Get was deleted and as always janny was to useless too stop it from being wasted
>shame

>> No.7000020

>>6999996
>>6999998
>>7000004
>>7000008
>>7000011
>>7000018
>>7000019
Reddit tier behavior

>> No.7000022 [DELETED] 

>>7000020
your mother

>> No.7000024 [DELETED] 

>>7000020
It's honestly deserved given that this is an AI thread on /ic/
A good portion of AI threads are just either Reddit posts or Twitter ones

>> No.7000027 [DELETED] 

>>7000020
>like I care for the integrity of this thread

>> No.7000029

>>6999954
Because he can't draw faces even while referencing an image. Instead he spends his time playing with a new toy instead of learning how

>> No.7000032

>>7000027
No I appreciate you lot shitting up a worthy thread but it's still le redditor funny internet number posting

>> No.7000038

ai lost

>> No.7000041 [DELETED] 
File: 95 KB, 720x452, IMG_20240101_090551.jpg [View same] [iqdb] [saucenao] [google]
7000041

>Get to hot for janny to handle

>> No.7000046

>>6999671
Get cancer and die a horrible death faggot
LMAO@ ur pathetic existence

>> No.7000066
File: 78 KB, 512x768, 1699958836607372.jpg [View same] [iqdb] [saucenao] [google]
7000066

lol

>> No.7000094

>>7000066
The fuck is with the hands your so inconsistent as hell that it's hard to tell what you just let the AI do for you
>then again you live to be walking advertisement

>> No.7000097 [DELETED] 

>>7000066
>hair melting and hands fucked
>"looks good to me" t.tripnigga

>> No.7000099

>>6999671
i wish i had a decent gpu but not, so i just have to cope and use the ai in the modest ways i can find. Aside from that, i always think on several ways to use the shit if i ever get the change to buy a modern gpu to work on Krita.

I think that one ideal way to work should be, of course, already know how to draw, then, work just on the style generation with the model/loras, so you can get the closest style in the prompt to your actual style, or at least the ideal style you want. Then work only in the "non iconic" aspects of the drawing, i mean, all that is not the face and some core keys of your drawing workflow.
For example if you make several pinup fanarts, then use the prompt to get the closest bodies, leave the head faceless and draw with your style, or use the AI to make the face but use it as reference.

More simple ways to work that i used before:
>prompt poses for reference to just copy, add the character by yourself.
>make your drawing nude and use img2img to prompt clothes, is a easy way to generate lingerie, bikinis, and some clothes, since is generated over your drawing, you just need to copy the shit in your style.
>add some shit in the background, you can make the basic idea in your drawing, finish the character(s) and your sketch the background, and use img2img to complete that, you can later delete your character and complete that background, or actually make the background from zero as a sketch, in boths ways later join the backgorund in a layer or copy the background in your style.

>> No.7000102
File: 3.86 MB, 532x298, 1674157605414306.webm [View same] [iqdb] [saucenao] [google]
7000102

>> No.7000109 [DELETED] 

>>7000102
>all that work yet still has it on a basic bitch anime setting

>> No.7000112 [DELETED] 
File: 3 KB, 96x90, 1704127519587156.gif [View same] [iqdb] [saucenao] [google]
7000112

>>7000041

>> No.7000122 [DELETED] 

>>7000102
why do you blurry everything like a fagget

>> No.7000124 [DELETED] 

>6999860

>> No.7000128

>tracing and using AI
what's the point?

>> No.7000129 [DELETED] 
File: 96 KB, 720x223, this nigga.jpg [View same] [iqdb] [saucenao] [google]
7000129

>>7000099
aren't you saying similar shit like this every thread
anon too just so you know the way you type sounds like this one anon who basically wanted to use it for Loli porn

>> No.7000132 [DELETED] 

>>6999671
Nigga really links dead threads

>> No.7000133

>>7000129
>like this one anon who basically wanted to use it for Loli porn
That would be mikufag who has supposedly been on /ic/ forever just like the tripfag

>> No.7000135
File: 644 KB, 684x1029, 1690936699016912.png [View same] [iqdb] [saucenao] [google]
7000135

i think i'm done with the values. if anyone sees any more improvements, send help.

that being said, i realize now that it is a good idea at all to separate the values from the color for AI work. because you can't reintroduce color without also massively changing everything else. it's probably better to work with color right away.
value mockups are still informative tho.

>>7000122
that's just the quality of the webm.
4chan has a 2min and 4mb limit on webms. and i'm pretty new to video shit.

>> No.7000145 [DELETED] 

>>7000135
>hair still melting

>> No.7000149 [DELETED] 

>>7000133
So guess it's true than the longer you've been here the more NGMI you actually are than

>> No.7000152

>>7000145
This the kinda nigga call a picasso wrong

>> No.7000159 [DELETED] 

>>7000152
so tripfag you mean to tell me your anime girl is suppose to be abstract.
>inb4 your prompt has realistic in it

>> No.7000163
File: 3.55 MB, 532x298, 1692494476543703.webm [View same] [iqdb] [saucenao] [google]
7000163

>>7000145
anything can be fixed easily.
look, a generic fix.

>> No.7000167 [DELETED] 

>>7000163
>fixed
can still not see shit and practically copy pasted in the end

>> No.7000170 [DELETED] 

>>7000018
shame the get was deleted

>> No.7000171

>>7000167
nvm this nigga just retarded

>> No.7000174 [DELETED] 

>>7000041
wasted Get

>> No.7000177 [DELETED] 

>>6999998
failed again anon
you were close though

>> No.7000178

>>7000163
Are you even drawing or just erasing to reveal the AI layer underneath?

>> No.7000184 [DELETED] 

>>7000178
honestly very hard to tell for tripfags stuff >he can't make better quality WEMBs but can use all the AI pluggins

>> No.7000195 [DELETED] 

>>6999998
Jak thread when

>> No.7000199
File: 610 KB, 588x867, 1703921217574571.png [View same] [iqdb] [saucenao] [google]
7000199

eh, maybe there is some use in getting the values done. since i can always use the next AI gen as an overlay color layer or something.
(although afaik that isn't the ideal way of transferring color to a greyscale layer)
not really ideal but worth considering.

>>7000178
obviously i didn't draw the hair there, save for a bunch of manual fixes afterward to bring it together.
at this point this is a mix between AI and my input. the sketch is obviously mine. see >>6999899
some of my other inputs
>manually redrew the arms and the hat at some point. and enlarged the breast a bit as well.
>the mask in general is manual only, since the AI is dogshit at recognizing it (at least using LCM).
>a lot of the value decisions like the bright background and the blending into the hat, breasts and ball are also manually done. i also added darker tones to the very bottom of the image (water)


>>7000184
those are different things. i only installed a video editing program today. and only to speed up the video. i'm just not that familiar with it.

>> No.7000202

>>7000102
>>7000199
what the actual fuck is this slop supposed to even be?

>> No.7000204 [DELETED] 

>>7000199
>just installed video program
>made several WEMBs throughout the year
>marketing spiel
>"I can always use the lastest AI..."
when tripfags about never relax

>> No.7000208

>>6999671
if you're going to draw anyways why not just learn to draw? why handicap your growth?

>> No.7000209 [DELETED] 

>>7000019
The get was he only true reason for this threads existance

>> No.7000213 [DELETED] 

>>7000208
he wants to sale you AI since he already given up on that
>and needs to constantly make AI threads to spam for his marketing firm

>> No.7000223

>>7000204
nah, those other webms were from twitter. like all the krea-ai ones, they were never mine in the first place. i made them fit using online converters.
i did record some of my webms (ubuntu has a built in function for that), but i only adjusted them using online converters before today.
i needed something else for this thread since i wanted to speed up the recordings to make them fit under 2 mins, which i couldn't do with those online converters.

i don't know why i bother explaining this to blatant retards who are trying their hardest to find fault with me tho.

>>7000208
lol.

>> No.7000237 [DELETED] 

>>7000223
I say fuck off to Twitter because you seem to only want to justify yourself

>> No.7000243 [DELETED] 

>6999750

>> No.7000251

>>7000223
>trying their hardest to find fault with me
You're leveraging the unethical data scraping of real artists so that you can avoid putting in effort and pretend to be a skilled painter. That's pretty repulsive, my guy. What you're doing isn't far removed from photobashing bits of pro work directly into your canvas and pretending you're doing something profound and clever. You hurt real artists and you hurt your own development. Go to /g/ you aren't welcome here.

>> No.7000264

>>7000251
i disagree.
you can take any of my current examples or of the last three threads and tell me what exactly it is that i supposedly have taken from other artists.

>> No.7000274 [DELETED] 

>>7000264
last 3 threads
>2 links are dead
what did he mean by this

>> No.7000278

>>7000274
>what is the archive

>> No.7000285
File: 941 KB, 1114x798, 1694014821186719.png [View same] [iqdb] [saucenao] [google]
7000285

somehow pure white backgrounds are always easier to work with. it was the same with black backgrounds i think.
makes me wonder if the AI's attention isn't spread as thin in these cases or something.

>> No.7000286 [DELETED] 

>>7000278
actually link them then retard

>> No.7000287 [DELETED] 

>>7000285
blog post on Twitter nigger

>> No.7000289 [DELETED] 
File: 258 KB, 718x742, IMG_20240101_125044.jpg [View same] [iqdb] [saucenao] [google]
7000289

/pol/ happenings. With AI I wonder how well deep fakes propaganda will be happening for any of the current conflicts

>> No.7000296
File: 718 KB, 1082x772, 1674050557137722.png [View same] [iqdb] [saucenao] [google]
7000296

>> No.7000305 [DELETED] 

>>7000289
Neat

>> No.7000306 [DELETED] 

>7000296
>hair and hands are still fucked
>fixed

>> No.7000317
File: 1.19 MB, 1111x808, 1676445005074924.png [View same] [iqdb] [saucenao] [google]
7000317

taking a break now.
don't burn the house down, retards.

>> No.7000332 [DELETED] 
File: 676 KB, 832x832, makes me wonder if the AI's attention isn't spread as thin in these cases or something.png [View same] [iqdb] [saucenao] [google]
7000332

>>6999899
>>7000066
>>7000102
>>7000135
>>7000163
>>7000199
>>7000285
>>7000296
>>7000317
*BRRAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAPPPPPPPPPP*

>> No.7000333 [DELETED] 

>>7000317
>colors and linework absolute shit if you zoom in
>wants to edumacate you still

>> No.7000339

>>7000332
Imagine its smell

>> No.7000342

>>7000264
That argument only works if you trained it on your own work. You didn't. You downloaded shit that was either already trained since the faggot behind it gave two fucks about the works he scraped, or you yourself gave two fucks and trained one yourself.
>It didn't store those images
Nigger it wouldn't be a thing without them in the first place. You use it. And on top of that you use it as a crutch. That makes you a double nigger. You disagree because you don't give a shit. You just want results.

>> No.7000343 [DELETED] 

>7000317
how long is taking a break is for you tripfag you seem to be coming here near everday

>> No.7000344 [DELETED] 

>>7000199
you seem to love (YOU)ing yourself like it's a credible source worth listening too

>> No.7000347

>>7000342
wrong.

you people are just so confused. you can't tell the difference between lora and base model training.
you cannot train an AI on your art from the ground up. you can only finetune a lora on it. and that WOULD make it better as far as making it adhere to your own style.
but it would still be trained on other people's art and photos. do you understand that?

the point here is that training is ultimatively just learning. yes, even lora finetuning.
it only becomes an issue if it is learned to the point of memorizing, or if someone deliberately finetunes it on another artists EXACT style (but even then it's arguable if style can be stolen).

but other than that, all of the training data is merely an influence. patterns for the AI to generalize from.
even when i use artist tags, the piece does not actually end up looking like it was made by that artist.
there is no hokus pokus shit about AI stealing your soul.
the AI doesn't know how craig mullins would paint something. it can only use what it itself has learned from looking at mullin's art. and that will be mixed with everything else it has learned.

>>7000343
you only believe that because you think that every pro AI post is me :)
the last thread was from two weeks ago now that i checked.

>> No.7000358

you can't even train a lora with your garbage scribbles. could have invested all this time into learning how to fix your shit

>> No.7000366 [DELETED] 

>>7000347
came following a bunch of AIniggas I'll call ya a nigga

>> No.7000369 [DELETED] 

>7000347
no one cares go to /g/ you absolute fagget

>> No.7000371

>>7000369
Hey I care, it's pretty interesting.

>> No.7000373

>>7000347
There you go talking about it as if its a person and deserves the right to "learn from" people en masse without their consent. Its a machine. It plagiarizes and it doesn't matter if its taking two pennies out of every artist's wallet. Its doing it to many thousands and none of them are okay with it. Its labor theft and it devalues human creative efforts in general and pollutes the internet with fake low effort bullshit. And here you are spamming /ic/ again and again while calling others retards and just not "getting it". You're the one that doesn't get it. You're mediocre and lazy and apparently very pleased that some sociopathic tech assholes robbed our work so you could play pretend. Get fucked, honestly.

>> No.7000375

>>7000347
>but it would still be trained on other people's art and photos. do you understand that?
Thanks for explaining. That makes you an ultranigger then because it's trained on other people's art and photos period. Those other paragraphs are useless because you just made it clear you need things to train it on. This "just learning like a human" meme is just a semantic /g/ copeout. Shit was scraped to improve a piece of software in the way you explained.

>> No.7000378
File: 73 KB, 1188x523, 1698742711016675.png [View same] [iqdb] [saucenao] [google]
7000378

>>6999671
Threadly reminder that proompters are failed artists eternally seething at artists better than them, participation trophies were a mistake and Shadiversity is a weenie.
https://youtu.be/n1VybvjzaK0

https://www.youtube.com/watch?v=mTBCzH1UyNY
https://www.youtube.com/watch?v=5sNvD8ePFHs

>> No.7000387 [DELETED] 

>>7000371
>Inb4 samefag or came with AI wave spammers

>> No.7000388

>>7000378
Aren't there way more prompters than artists though?

>> No.7000394 [DELETED] 
File: 11 KB, 168x300, images (1).jpg [View same] [iqdb] [saucenao] [google]
7000394

>>7000388
sure anon but at time do you want to be in the same category as e-thots, because find this shit hilarious to see
>you'll buy anything

>> No.7000410

>>7000129
not really this is the first thread that i found about this AI shit actually

>> No.7000428 [DELETED] 

>>7000410
>first time
how new a fag are you than nigga

>> No.7000437

>>7000317
so like 5 or 6 hours later "you" just made a girl melting into a pool of melted cheese while wearing an oni mask and a corn husk hat while holding a nondescript white sphere?
is this some kind of weird fetish?

>> No.7000448 [DELETED] 
File: 143 KB, 656x1024, 1704148644793404m.jpg [View same] [iqdb] [saucenao] [google]
7000448

>> No.7000453

>>7000375
in the real world, information doesn't come from nothingness.
nobody can learn something without knowing what it looks like.
and if you can't follow the rest of the logic from there, then it's a waste to argue with you anyway.

>>7000373
>It plagiarizes and it doesn't matter if its taking two pennies out of every artist's wallet. Its doing it to many thousands and none of them are okay with it.
that is not how it works. it is not just sticking together existing shit.
everything you say is based out of ignorance.

once again: show me in any of the my examples how i plagiarized other artiststs.
>but it's only "two pennies out of every artist"
and how do you think that works?

look at >>6999714, do you think that suit exists in some form in the training data? in X arrangement, X color, X proportions, X type of the stature of the person wearing it?
even if it does, i can change every single one of those factors and the AI would still be able to work with it.
what does that tell you?
and if you can't follow the rest of the logic from there, then it's a waste to argue with you anyway.

>> No.7000457
File: 178 KB, 892x583, 1674929121923910.jpg [View same] [iqdb] [saucenao] [google]
7000457

>> No.7000478 [DELETED] 

>>7000453
>ain't reading all
and also keeps using himself as an example even though their a known shill. Literally keeps going------>>>
>source "my ass"
>look at my posts "blog" to see why

once again actually go to Twitter if you just want to blog post you fagget

>> No.7000484 [DELETED] 

>>7000453
I'm pretty sure tripfag you might get more people listening to you, if you just killed yourself and left your dissertation behind.
Maybe actually appealing to others emotion for sympathy would help your cause

If not though and most see you dying as an absolute win, than I wonder what that says about you and your need to be bitching all day saying use AI

>> No.7000489 [DELETED] 

>>6999671
I bet you drink piss for a living OP

>> No.7000493

>>7000489
Is that a career option? Professional piss drinker?

>> No.7000499

>>7000493
well t B h you prob find that in India if you know...

>> No.7000500
File: 87 KB, 2322x988, amuse.jpg [View same] [iqdb] [saucenao] [google]
7000500

>>6999743
ok, I installed it

I'm yet to learn and adapt to this app since I have a daily wagie to attend to, probably I'll try it later

>> No.7000502

>>6999853
It's certainly possible to make AI generations illegal on a commercial level at least.

>> No.7000523
File: 61 KB, 717x459, Janny and the GET failure.jpg [View same] [iqdb] [saucenao] [google]
7000523

Janny honestly scrubbing all traces of the GET failure
>truly they're the biggest trannoid to exist
and for what, an tripnigga's blog post

>> No.7000526

>half the thread gets deleted by janny
>let's the shilling avataring fagget stay

>> No.7000533
File: 340 KB, 1332x820, superspam.jpg [View same] [iqdb] [saucenao] [google]
7000533

>>7000526
there is clearly somekind of agenda to get AI accepted on this site because there is an AI general on almost literally every board (not welcome on /ic/ so they're trying a different angle here I think)
the mods/admins have already been pretty much exposed for allowing pay-for-play off-topic shilling (see fishtank spam on /tv/), and a lot of investor money is being funneled into AI shit so it isn't beyond the realm of possibilit

>> No.7000537
File: 14 KB, 1135x220, Screenshot (106).png [View same] [iqdb] [saucenao] [google]
7000537

>>7000533
honestly further reason not to trust shill spam considering how much this nigga wants to go on "convincing people" who rather honestly be left alone

>> No.7000573

>>7000533
People like it so they talk about it.

>> No.7000582
File: 309 KB, 643x720, 1667489626987316.png [View same] [iqdb] [saucenao] [google]
7000582

>>7000526
Did it ever occur to you that the tripfag has been on topic, and posting content from implementing Ai into his art? Meanwhile /ic/ is collectively losing their minds screeching and throwing temper tantrums. Please stay on topic.

>> No.7000586

>>7000582
whats up mikufag

>> No.7000589

>>7000582
AI usage for making art should never be normalized and most people here are on the same page. If the mods/admins actually cared about the community here they'd make it reportable but instead they seem to be "in on it".

>> No.7000601

>>7000589
>AI usage for making art should never be normalized
Why not? We see Ai being used to create art from /biz/ to shill various assets Ai is a very useful tool to get messages across. Do you honestly believe that other industries wouldn't take advantage of Ai just because some frogposter from some backwater art community nobody has ever heard of says so? How are the Mods going to moderate each and every image posted on /ic/ and determine whether Ai was used at all? Are we going to start counting fingers, and ban anyone who hides hands? I'd like to see that desu.

>> No.7000608

>>7000601
it's not "art", at least not the kind of art this board is about
there is a sticky
is anything referenced in that sticky applicable to generative AI?
this shit should stay on /g/ , it's literally off topic here

>> No.7000612
File: 8 KB, 465x160, Screenshot (108).png [View same] [iqdb] [saucenao] [google]
7000612

>>7000601
first off...
>nigga
>pic related
>AIfags can literally take any pic generated and claim it as their own
>your just looking at filters at that point
>tripfag needs to be told when his shit fucked
>"it will get better" but he doesn't apperently
>critique is apart of the board
>"Are we going to start counting fingers"
you're still fucking up hands than you braindead motherfucker

>> No.7000613

>>7000608
That sticky needs updating, you know this. The sticky is from another decade hell even the Loomis link with the red fox being thrown books at is dead it's the end of an era. Just because the Sticky doesn't accommodate Ai doesn't mean a thing except how it needs updating.

>> No.7000618

so just to be clear, janiggers are completely fine with AI slop thread spam?

>> No.7000619

>>7000613
updating for no AI than

because on /g/ their threads are basically image dumps and it basically falls more into /gd/ if you have to fuck around editing and photoshoping/bashing stuff with filters

>> No.7000621
File: 641 KB, 2928x4096, 1698516900654919.jpg [View same] [iqdb] [saucenao] [google]
7000621

>>7000619
If you insist on banning Ai as a tool, why stop there? Let's ban anybody who used 3d models who traced over because they used a 3d model. Let's ban those who used the warp and free transform tool because the software changed the proportion of a hand and the artist didn't redraw it instead. Hell let's go all the way and outright ban those who used a reference because they didn't draw from imagination. I'm all for banning you fags that use references.

>> No.7000625

>>7000621
>why stop there?
because no one disagrees about stopping there, that's a perfect place to stop

>> No.7000627

>>7000621
"Why stop there" because they want to stop there, are you stupid. "Why stop parking before the curb when you can park over the curb" idiot

>> No.7000630
File: 14 KB, 225x225, bc356b289a9b8f1df671b15d5107e31e.jpg [View same] [iqdb] [saucenao] [google]
7000630

>>7000582
tripfag made the thread to shill and practically blog post.
>On Topic
>made an AI when /ic/ getting raided with AI threads
>tripfag given at least a 50pg dissertation to use AI
>keeps (You)ing himself as an example
and for (YOU) honestly
>anime posting

>> No.7000735

>>7000630
>>anime posting
anime website, niggerman.

>> No.7000834

>>6999671
prompt alone is enough to be an artist, but that's a childrens depiction. prompt + workflow is more than 90% of anons - the workflow is more important than the words you feed into it.

>> No.7000836

>>6999671
neck yourself
>6999675
>ignore the autists
you're in a bubble of your own making

>> No.7000839

>>6999868
>in case people don't understand what is happening in that webm
you trying very hard not to do the work

>> No.7000840

>>7000342
this

>> No.7000841

>>7000582
stfu mikkunigger you're in every thread sucking his dick because you're a fucking beg unwilling to work

>> No.7000842

>>7000342
>>7000840
>it wouldn't be a thing without them in the first place
and where would you be without other peoples art?

>> No.7000845

>>7000834
you niggers keep touting "worflow" as some magic get-out-of-jail word

>> No.7000847

>>7000842
learning from life. but you're just using that overused false equivalency again.

>> No.7000848

>>7000845
if thats how you view it, whats your response

>>7000847
look, anon was on the debate team

>> No.7000850

>>7000848
seethe tripfag

>> No.7000853
File: 2.35 MB, 1000x1200, lowres, low quality, worst quality, b&w, bone, --_00035_.png [View same] [iqdb] [saucenao] [google]
7000853

>>7000847
>i have never seen another piece of art in my entire life, i have been completely uninfluenced, my only inspiration is existence itself

>>7000850
i would never steal tripfags valor

>> No.7000912

>>6999671
you and your two butt-buddies who might or might not be you need to rope. stop assuming you have support beyond less than a handful of permabegs

>> No.7000918

>>7000853
>I am a retarded ai niggers who shifts the goalpost and purposefully misreads
why yes (You) are mr false equivalence

>> No.7000926

>>7000847
but the fact of the matter is you are not only learning from life. you are learning from other artists all the time.
people don't have massive folders collecting real life photos or videos. no, they mostly view other creative's works for inspiration and reference.

so the answer to
>and where would you be without other peoples art?
is: you would be set back. massively. centuries worth of technique and exploration.
and yet you take that for granted and consider it natural to incorporate other's art like this.
visual library is the term we all use for this.

the same argument applies to reality too. because for AI, there is no separate "reality" to draw information from. only the training data. what matters is how it is used. just like how it matters whether you use an artwork/real life for inspiration versus tracing or painting on top of the piece.

again, if you don't understand this, there no point in arguing.
i'm pretty sure that most artists who aren't shitters understand this idea perfectly. but AI is a blind spot for many of you and you're willing to delude yourself in order to keep propping up the shitty arguments.

>> No.7000932

>>7000926
It's been well over a year you brain dead AI shitters need to come up with a better argument than the nonsensical
>it le is inspired by other art just like people are!!

>> No.7000938

>>6999671
I don't know what to say to this. You don't like drawing, but I wish you did.

>> No.7000940

So the tripfag is constantly shilling his AIslop because he has nowhere to put his "work" in?

>> No.7000943

>>7000932
it LEARNS from its training data.
meaning it can then APPLY what it has learned, faulty or not. as is apparent with every single example i posted in every thread.

i know it's hard to believe that a machine can do this, but that's just the reality of how it works. these aren't arguments, they're explanations. but you'd rather keep living in your delusion.

>>7000940
as far as i'm concerned, these are PSAs.
if i wanted to shill, i wouldn't be doing this here where nobody appreciates it. nor would i censor my own art with masks and shit, would i?

>> No.7000944

>>7000453
>>7000842
>in the real world, information doesn't come from nothingness.
In the real world humans are not software. What even is this comparison you're making. AI is not a human. You are not learning jack by downloading programs or using images to improve software. You are not AI. It legit sounds like you have made part of your work AI and you are using this as an argument that you are learning by using AI when it's just false equivalence and cope making you think you're better than you actually are. AI is software that /g/ scraped images off the internet to improve it that happened to use the buzzword "learning" that you cling to and argue semantics like "AI is like a part of brain" to justify that false equivalence.

>> No.7000959

>>7000943
>if i wanted to shill, i wouldn't be doing this here where nobody appreciates it. nor would i censor my own art with masks and shit, would i?
Yet you are here trying to make your case despite the influx of anons telling you to fuck off. Go stick to your AI shitty forums and wank the rest of your kind there. Let the rest of us /begs/ struggle here, a thread died so your shitty thread could stay.

>> No.7000960

>>7000943
nigger, there's nothing else to say

>> No.7000962

>>7000926
the fact of the matter is you keep peddling your false equivalence to justify the use of your crutch. you have a fucking savior complex to boot and don't realize how much of a clown and counterproductive you're being.

>> No.7000967
File: 808 KB, 1363x1524, They're-The-Same-Picture.jpg [View same] [iqdb] [saucenao] [google]
7000967

>>7000926

>> No.7000974

GUIs, hear me out.
AI is real and human. Melon and Double-Scam will lead the world into a technological utopia in joint dictatorship. The future is bright and digital. The gamification of life will bestow us with the hottest anime harems imaginable. It will be a true paradise in the brainmachine.

>> No.7000975

I understand AI, I use it daily. I don't understand wanting to have threads for AI on /ic/. we don't have photographers here sharing photos, why would we want AI users to share AI images? even if your lines remain and you direct it to an outcome, it's a different category of art, as is photography. It's more like googling for an image that perfectly matches your lines in a universe where every image that could exist has already been made, and you can find more of those images by modifying the input(the search).

>> No.7000977

>>7000975
Considering AI will always be a thing now, a scenario for anyone using AI would be to talk about the parts they drew here, then going to countless /sdg/ threads to talk about generating whatever else they want. Legit no reason to have threads about it here when it's just /g/ beyond that point.

>> No.7000987
File: 165 KB, 1098x710, Screenshot 2023-10-04 173415a_proca.jpg [View same] [iqdb] [saucenao] [google]
7000987

>>7000944
>you are using this as an argument that you are learning by using AI
lol......no. i don't know why you would read my post that way.
i'm telling you the simple truth that AI is learning from its training data. you're so misinformed and confused when it comes to AI that you can't even understand a simple english sentence.

when the AI takes in a thousand images of circles, it doesn't "learn" a thousand circles. instead it learns what circle GENERALLY is like. the concept of it. and that representation is what it then uses to create its images. and the same happens for every concept the AI learns, going from "cat" to more complex ones like color relationships and reflections or lighting.
the AI is not human, but the AI is learning despite it. this is what makes it work in the first place.

i find it funny how even now, you people cannot read this simple sentence as it is without adding some random baggage on top of it due to your own lack of understanding.
when i say that the "AI is learning from its training data", i mean literally just that. not that it is human, not that it learns exactly like a human, not that i am learning, none of that.

this is not some cute approximation, or some justification, it is just how the technology works.

>>7000975
>It's more like googling for an image that perfectly matches your lines in a universe where every image that could exist has already been made, and you can find more of those images by modifying the input(the search).
>in a universe where every image that could exist has already been made
but they haven't been made, so your analogy isn't quite the same.
those images couldn't even exist at all without said line input.
and lines weren't my only input. you can use your whole image as input, down to the % of it that you allow to change.

>> No.7000988

>>7000977
interesting...but you are a fag so nothing you write has value lol

>> No.7000995

honestly, these threads just convinced me ai is nowhere near a usable level
this retard literally spends more time fucking around with ai than it takes to actually paint the thing and the end result still looks like melted nonsense

>> No.7000998

it's interesting how AI "learns just like humans" do when it is convenient to ignore the fact that this technology was built on top of the work of people who did not consent to have their work used this way and may just have their lives negatively impacted by it

but it's "just a tool" when it comes time to claim it's actually your own work and you really are an artist

>> No.7001001

>>6999671
i'm kind of interested
post like a short general guide to start me off

>> No.7001006

>>7000987
You don't want to admit the simple fact that images have to be scraped aka "stolen" to train software without anyone's consent. And you are equating it to "it's the same as if you looked and learned from it" when it's false equivalence in every sense but the name. AI is software. If a picture is taken to enhance software, it's taken to enhance software under the buzzword "learning". It doesn't matter how it does it. Fact is IT does it. And that's the problem. Of course using it further for your own gain at that point is essentially using other people's work now to generate yours, because again, without it the software wouldn't exist. It's was an essential component, training, learning, copying or not. And equating it to me looking at the picture and learning from it is false equivalence.

You keep going in circles about HOW it works when that's not even the point because that's all you know. You keep writing these "how it works" paragraphs and discard anything else because you have no answer to that and comfort yourself by saying "you're misinformed" and "you're confused" when you have no idea what the conversation is about. You just grab a sentence you can write your essay in response and use it as gotcha.

>> No.7001016

>>7001006
>You don't want to admit the simple fact that images have to be scraped aka "stolen"
stopped reading there, saving images is not stealing

>> No.7001018

>>7001016
Saving isn't.
Using them is.

>> No.7001031

>>7001018
here, let me clarify your argument for you: an artist's style is intellectual property, therefore creating a distilled representation of that style via software and using it to create imitations is intellectual property infringement.
unfortunately, artistic style is not intellectual property and is not legally protected. it doesn't matter if software is doing it when the core info isn't IP.

>> No.7001036

>>7001031
a straw man. how nice for you.

>> No.7001037

>>7001036
how is it a straw man? be specific and articulate.

>> No.7001038

>>7001031
It doesn't matter if style is or isn't intellectual property. The image was downloaded. Put in a folder. Then the program was told to use that folder to train itself. The image was not just saved. It was used by the program for training. Or do I have to physically drag and drop it into the program for it to count as me stealing the image? Isn't it convenient that you can't see the datasets of models and loras? For sure I trust the model creators personally asked every artist if they wanted their work used as part of a training program.

>> No.7001039

>>7001031
here, let me blow your mind: human laws aren't natural laws. they weren't carved into stone by jeovah with lightning bolts; as the world changes, we come together and discuss things so life can be better for all of us. it's not shocking or surprising that our current laws aren't prepared for new technology. we can update our laws. we often do. we may come to an agreement about how to deal with generative AI and we may find out that it is better to ban it or regulate it or maybe even to do nothing about it. it's not set in stone. I'm not taking a stance one way or another yet, I don't know enough about this to have an informed opinion, but don't act like it's some sort of fool proof argument that old laws don't apply to new developments so we don't need to do anything about it.

>> No.7001043

>>7000066
I like how you did her hat.
I'm still not convinced "integrating AI into your workflow" will be a regular thing any time soon unless you're an overworked industry guy who really needs to hit a deadline.

>> No.7001051

>>7001037
not about style.

>> No.7001069

>>7001043
>did

>> No.7001072

>>7001031
It's not the style that's the issue you neolithic browed retard, it's the fucking thousands of images used without consent or compensation that's the issue.
I thought you tech slaves were supposed to atleast be somewhat smart but you're all unfathomably daft it seems

>> No.7001078

>>6999671
>ask questions
I have one. Why are you not learning how to draw?

>> No.7001079

>>7001072
The usual response is that if it's okay for artists to reference and learn from other people's work without permission, then it's fine for AI to do it too.

>> No.7001081
File: 764 KB, 3000x3600, lowres, low quality, worst quality,--_00103_.jpg [View same] [iqdb] [saucenao] [google]
7001081

>>7000987
that picrel is a nice way to explain it

>> No.7001082

>>7001081
neck yourself
>>7001079
the usual response doesn't fly

>> No.7001086

>>7001079
Explain to us how the process of le diffusion is similar to how the human mind interprets and produces new art

>> No.7001089

>>7001038
you were just whinging about the other guy getting hung up on the "how" and yet you do the same thing without making any argument. I know how it works. my point is that you're using this "stealing" verbiage when no property is being stolen.
>>7001039
of course laws can change. maybe they'll make owning a computer illegal. I think it's a reasonable basis to start things with regardless. so if we're designing a new law, how should it work? should generating images be banned outright? seems like a real luddite mistake to me. what I think would make sense would be to make commercial use of generated images which are explicitly marketed as imitations of specific artists copyright infringement; this may already be the case if the artist can demonstrate directly lost revenue.
>>7001051
>>7001072
style imitation is the strongest case artists have, though. using publicly available data (images) to train commercial AI models has a lot of precedent at this point, good luck making feature extraction illegal.

>> No.7001090

>>7000987
>>7001081
so it doesn't learn like a human, it's not human and shouldn't be afforded the same leeway as a human. stop it already with your false equivalency

>> No.7001091

>>7001086
Ofc it isn't you absolute retard my god how are you all so stupid shaking my head. It's just learning duh. Isn't learning what we all do?

>> No.7001092

>>7001089
>using publicly available data (images) to train commercial AI models has a lot of precedent at this point
well about that...

>> No.7001094

>>7001089
>my point is that you're using this "stealing" verbiage when no property is being stolen.
So by your logic that drawing doesn't belong to anyone because "style can't have copyright"? That drawing I posted online is free for anyone to use for whatever purpose because it's not literally on paper in front of me so it can't be stolen?

>> No.7001095

>>7001089
>luddite
he said it! he said the thing!
simply put, ban use of AI generation for commercial purposes. bam, solved the whole thing! you're a nigger btw

>> No.7001099

>>7001079
you know nigga that doesn't work if your going to sell stuff than "referencing" can be iffy. The mass production part is something most people can't do as well. So your argument like a human would doesn't exactly fly well.

>> No.7001101

>>7001094
of course the drawing belongs to the artist. they have the copyright on it by default. transformative works are fair use of copyrighted material
>>7001095
>muh tribal shibboleth
>no argument
>ban all commercial image generation
lol good luck buddy

>> No.7001102

>>7001089
>if airplanes are allowed to travel at 500mph how come I'm not allowed to drive my car at 500mph?
>fucking ludites trying to halt technological progress hmph
1. kill yourself
2. hurry up

>> No.7001104

>>7001101
it has been establish it isn't transformative work
>no argument
I gave you an answer, you not liking it because you can't confuse people in an endless semantic bickering is irrelevant. you're just saying "just give up bruh"

>> No.7001107
File: 1.10 MB, 5712x2616, lowres, symmetrical, ugly, frame, paint, linework, b&w, flower, green, fur , circle, complex, pyramid, yellow, illustration, render, earth, purple, sketch, intricate, 1girl, 1boy, human, person, figure, _00240_.jpg [View same] [iqdb] [saucenao] [google]
7001107

you guise just havent seen one you like yet, its okay itll happen at some point
>yours is shit
thank you

>> No.7001108

>>7001101
don't use buzzwords if you don't want people to mock you.

>> No.7001110
File: 139 KB, 614x714, 1000001372.jpg [View same] [iqdb] [saucenao] [google]
7001110

>>7001089
>though. using publicly available data (images) to train commercial AI models has a lot of precedent at this point, good luck making feature extraction illegal
well anon I know certain people who on that case. All I can think of when it comes to that is people going "you can not stop me" mentality

>> No.7001119

>>7001107
>you guys just haven't seen one you like yet, it's okay it'll happen at some point
https://www.youtube.com/watch?v=YMPQdaq-qqk

>> No.7001120

>>7001110
fags think "publicly available"="it's on the internet I can just take it"

>> No.7001121

>>7001101
>transformative works are fair use of copyrighted material
Because tracing and photobashing are looked at so brightly, right? That's why the original used "stealing" in quotation marks. If you resort to fair use to wiggle out, you know very well how much of a faggot you're being and you're ok with it.

>> No.7001122

>>7001119
not really interested in that anymore. i just do it for myself.

>> No.7001126

>>7001107
what does that have to do with anything?

>> No.7001129

>>7001102
>no one should ever use tech that lets people do more with less
>>7001104
it is transformative work. how is using an image as input on a model's weight adjustment not transformative of the original?
would love to see where this was "established"
>>7001108
>don't use my twitterbrain trigger words or I'll get mad and that's your fault!
no
>>7001110
how is your panic article screenshot relevant? if anything, this kind of thinking just creates a moat for corporate entities who will be the ones with greatest financial impact on artists
>>7001121
this isn't about whether or not it's "looked at brightly", obviously illustrators don't like it. fair use is fair use.

>> No.7001130

>>7001101
>transformative works are fair use of copyrighted material
tell us what is transformative use is, in your words. There is zero chance your words describe how courts use that term.

>> No.7001133
File: 277 KB, 723x1008, 1703381669169571.jpg [View same] [iqdb] [saucenao] [google]
7001133

>it's ok, I transformed it with linear algebra.

>> No.7001137

For ease of argument let's say before AI generated images there'd be like around a thousand nice to look at drawings/paintings uploaded to the internet every day, and then when AI became widespread it rose to around ten thousand nice to look at "drawings/paintings" being uploaded daily across the internet.
What about this is good?

>> No.7001145

>>7001129
>it is transformative work. how is using an image as input on a model's weight adjustment not transformative of the original?
would love to see where this was "established"
just using filters can be considered transformative work
>using twitter words
fuck off fag
>how is your panic article screenshot relevant? if anything, this kind of thinking just creates a moat for corporate entities who will be the ones with greatest financial impact on artists
Seeing how Youtube is and other media sites are you'll always be dealing with corporate entities retard

>> No.7001146

>>7001129
>no one should ever use tech that lets people do more with less
So true why go to law school and get a degree when I could just make a fake one in a few hours, work smart not hard

>> No.7001149

>>7001129
>don't use my twitterbrain trigger words
well, you fags kept misusing it. so whenever you mouth it you just paint yourself as a retard, what can I say?

>> No.7001152

>>7001130
transforming something into something different, lol
check out how generous the transformative definition is in the google cases
>>7001145
are you agreeing with me? I can't even tell what you're trying to say
>>7001146
please try that

>> No.7001154

>>7001129
>this isn't about whether or not it's "looked at brightly", obviously illustrators don't like it. fair use is fair use.
Did you not check what board are you on, /g/bro? If you just blanket "fair use" on these unethical practices and expect to be treated well, you are really out of touch with reality. You'll be called a faggot to hell and back on any self respecting art community and are not going to have a fruitful discussion. If that's what you're after, be my guest.

>> No.7001155

>>7001129
>it is transformative work.
you repeating it won't make it so. no human is enacting the transformation

>> No.7001159

>>7001129
>fair use
corpos can already show you their dick suck if you want to try and use it at times anon
it's more so a case which ones are more tolerant of that stuff

>> No.7001160

>>7001152
>transforming something into something different, lol
nope. wrong. check out how those generous case were lost by the person who ripped shit off.

>> No.7001161

>>7001129
>do more with less
Do what exactly?
Churn out generic Twitter schlop?
Hell yeah man #ifuckinglovescience

>> No.7001173

>>7001152
>are you agreeing with me
well anon seeing as some can put this video below on top right of a screen or how react videos can be. Is this the transformative media you want to see
https://www.youtube.com/watch?v=nbWOuI8K-jI
but use any sec of a copy written music song and you'll get be punished

>> No.7001174

Ai is fine as long as they know they're doing less work and accept its reducing workload. It's not comparable to digital art nor traditional.
I think ai art is art as I do not believe in a human soul. However it is assisted by an automated and generative process. Though I don't understand why you'd do it for fun rather than any practical reason when you have the time to do so.

It is also using the works and efforts of others without their consent for the ai, so there's that too.

>> No.7001180

>>7001174
I think both sides can see how idiotic this statement is.

>> No.7001184

>>7001154
I didn't make the thread, I just think the overlap of drawing and generation is an interesting topic to explore and I don't see many people doing it since they just want to proompt and hit the generate button.
>>7001155
it's objectively being transformed, being done manually is not part of the definition
>>7001159
the fact that some people will dump a case in your lap and some won't doesn't make it not fair use
>>7001160
google won the case bro, what are you talking about
>>7001161
it's enabled a lot of people who have no artistic skills to "create" ("capture" is maybe more accurate, I think it has about the same artistic input as photography but without historical value). the predictable result of this is a bunch of tasteless garbage, but it has creative potential for sure.
>>7001173
this is really a reach man, youtube's content policies are a preventative measure against legal cases, not an actually legal definition
if you're asking if I have an issue with react videos on youtube, no, I don't care

I'm going back to working through the proko course I'm doing, I never engage in these discussions because I think it's a waste of time but wanted to make an exception for this because I think it's neat.
best of luck with your experiments, OP

>> No.7001187

>>7001129
>no one should ever use tech that lets people do more with less
it's all garbage. every single AI image is the definition of slop.
I played around with stable diffusion. You need to put in a lot of time to tweak it to get something even close to what you imagine and in the end it's still just still soulless shit. Faster to actually draw it yourself than try to generate anything, you are just wasting your fucking life time. It's probably great for people whoes high of creativity is princess peach and ryan gosling together on a date. Literally goyslop.

>> No.7001192

>>7001187
>you need to put a lot of time into
Wait, you're telling me the computer doesn't do all the work?

>> No.7001195

>>7001192
The computer does all the (hard) work, retard, and the hard work was done by artists. What you do is trying to make the computer do the work you want to see. You do not create art, you lead a retarded machine in a direction in hopes it spits out something even close to what you have in mind. fucking waste of time for pathetic people who wish to be an ""artist""
It's like commisioning an artist who is retarded and can't speak any human language. You aren't creating shit.

>> No.7001198

>>7001180
>no refutation
Do i make you upset, aifag?

>> No.7001203

>>7001198
You're the only one who seems upset, well you and 90% of the anons calling each other retard. I'm having fun, why aren't you?

>> No.7001207

>>7001174
>less work and accept its reducing workload
they now have the work of everything placed on one person on far less and still would ship shit half finished or cut up for profit.

I don't believe you'll find many projects caring about your satisfaction only money, if you don't care about people or the human soul

>> No.7001209

>>7001203
Projection won't work. Nice try, aifag.
Why not draw normally instead of making everything instant gratification?

>> No.7001216

>>7001209
Do you think you're able to argue without throwing names?
>instant gradification
You know nothing about imggen lol

>> No.7001221

>>7001216
You literally started with an insult, aifag. You brought this upon yourself.
You will NEVER be a true artist.

>> No.7001229

>>7001221
You can do better next time, but good try

>> No.7001266

>>7001184
>since they just want to proompt and hit the generate button.
Because that's what it's for. If you're an artist, you take pride in what you draw and delegating your would-be work to a machine is just using a crutch for the sake of it and not you actually indulging yourself into a drawing or wanting to draw which is the point. Not to mention you create this greyhole of ownership where you did X% of your drawing? Not to also mention ethics of the tool and how it was trained on your peers? Are you really proud of something you didn't do for the sake of the dopamine hit you get when software spits out an image from your sketch you then photobash? You can argue
>Not everyone is the same retard
find me respected artists that admitted to dropping work for ai and knowing all this. It's not just about production. The experience of making it is the foundation. The "assistant" is a meme because in reality assistants are a mutual benefit for the artist and assistant. If the artist could, they would do it all themselves, but livelihood is at stake. The need for production is not out of boredom of drawing things you don't like or that you can't draw them, it's out of necessity for it to be put out asap.

/g/ on the other hand is all about production. You drawing shit is inefficient and pointless because let's be real - a stickfigure with some shapes will get you 99% of the way. Even if not now, it will soon with some 10 minute photobashing. You don't need to be an artist for that. Knowing prompt engineering at this point is more efficient than wasting time drawing.

The in between doesn't work. You're either being an inefficient retard for one, or unethical retard for the other. Pick your poison. Basically you never had it in you to become an artist but you want to be seen as one. Your best bet is /g/ for this "assisted art" because /g/ doesn't give a shit about ethics. Any drawing community will slander you for being a hack, because you essentially are.

>> No.7001269

>>7001266
not reading this

>> No.7001270

>>7001269
Unroll

>> No.7001284

I still don't understand this weird obsession with trying to force AI into a place that clearly doesn't like it. Give it a rest already

>> No.7001285

>>7001284
We just want to pee

>> No.7001286

>>7001184
>it's objectively being transformed
you've said that exact same shit a while back you nigger. again you resort to semantics equating transformation and artistic transformation. abusing technicalities is begging for more legislation

>> No.7001292

>>7001284
you aren't the absolute authority on what a community should decide to like.
either way, healthy communities are capable of talking about things they don't like. zoomers might think otherwise, but it's entirely possible to not have the urge to cancel everything you dislike.

>> No.7001294

>>7001292
/mu/ could really use a few threads about trigonometry you're so right

>> No.7001300

>>7001294
false. ai art is art related, so it belongs on /ic/. simple as.

>> No.7001302

>>7001300
you aren't the absolute authority on what a community should decide to like, bro

>> No.7001304

>>7001292
I know I'm not but look at the thread it seems no one else likes and wants it either. It doesn't make sense to have it on a board made for learning how to draw. yet AI lovers insist of forcing themselves upon /ic/

>> No.7001315

>>7001292
>healthy communities are capable of talking about things they don't like
But we are. It's just the 4chins way of talking about things "we don't like" if you haven't noticed. If you don't agree, r*ddit might be up your alley. The absolute authority is calling you a retard for wanting to talk about unethical /g/ art.

>> No.7001324

>>7001302
i never said or implied that, but what you said applies to all members of this community. i agree.
trigonometry doesn't belong on /mu/ unless the thread's topic is about using it to make music or possibly music software. it's exactly the same case in /ic/.
>>7001304
>look at the thread it seems no one else likes and wants it either.
irrelevant. this isn't normal social media, so you just force away relevant topics just because the vocal people don't like them.
>It doesn't make sense to have it on a board made for learning how to draw.
why wouldn't it make sense? are you learning by being on 4chan in general?
people are capable of dropping a thread and then going to draw if they want. i don't see why looking at an ai thread for a few minutes is going to harm everybody involved.
i also disagree that this is a board exclusively for learning how to draw traditionally. the board is more than capable of also having topics that concern other forms of art or drawing that impact the entire artist sphere just the same.
>yet AI lovers insist of forcing themselves upon /ic/
i doubt it. you might have gotten an influx of cross-board shitposters when stablediffusion was first a thing, but currently having the occasional threads about ai isn't "forcing themselves upon /ic/.

>> No.7001325

>>7001300
We could do a poll.
I count 4-5 aifags out of 52 posters.

>> No.7001326

>>7001325
okay, if you can get the site staff to agree to making board rules based on anonymous polls, then i'll concede to your arguments.

>> No.7001328

>>7001324
I could claim trigonometry sings to me, the same way you claim a piece of software randomly generating images is art to you.
Both are equally schizophrenic.

>> No.7001329

>>7001292
How about you go back to /g/ and stay there, Pajeet. There is no artistry involved in dragging and dropping loras, typing in word prompts, and hitting generate until you get something that doesn't look like a complete abomination.

>> No.7001330

>>6999671
Which generator are you using? Is that available within Krita? I'm only really interested in using AI for background/building ideas. I like otherworldly shit and AI is perfect for that.

>> No.7001331

>7001324
it's not relevant just because you say so...where is your trip btw?

>> No.7001335

>>7001324
>currently having the occasional threads about ai isn't "forcing themselves upon /ic/.
Nigga what are you talking about. Janny deletes bait threads from fags almost daily. Just this morning there were 3 doomposters and falseflaggers that got nuked. You let one of them settle and in a year /ic/ becomes the last resort option for /sdg/fags when they get bored on their own boards and get the urge to bumplimit yours. We don't need that shit everywhere. Relevant or not. Petition for your own board and have threads there.

>> No.7001336

>>7001324
there's literally a tripfaggot constantly shilling these AI threads while saying DO YOU ACCEPT AI YET /IC/?? Clearly we fucking don't so fuck off. As the other anon said 4chan is practically self moderated if you don't understand what that means fuck off to reddit

>> No.7001339

>>7001300
my little pony is /tv/ or /co/ related, and yet they had to contain that shit in its own board

the only reason these threads survive is because they're controversial. 90% of the posts are just people insulting the handful (?) of AI evangelists, the other 10% is that verbose autist copy pasting the same "argument". there's very little actual discussion. there wouldn't be even if the threads weren't derailed. there's not much to say about it and not enough people come to ic for this anyway

that and that trip autist likes the attention, or he wouldn't identify himself and absorb the constant abuse. so he keeps coming. again and again. this is spam and should be dealt with as such

yes, I know I'm part of the problem now. but the moderators can end this very quickly.

>> No.7001340

>>7001328
>I could claim trigonometry sings to me
yeah, then it would be /mu/ related; there's no problem there. even /ic/ has that one schizo that makes incomprehensible scribbles most of the time.
>Both are equally schizophrenic.
subjective and not a factor into what makes a thread on-topic.
>>7001329
>There is no artistry involved in dragging and dropping loras, typing in word prompts, and hitting generate until you get something that doesn't look like a complete abomination.
did you read the thread? it's not about just generating things, it's about adding a human touch and using ai in a photobash-like manner, and discovering other methods of using ai harmoniously without depending too much on it. all art related topics.

>> No.7001342

>>7001329
Oh nice you're learning more! Good to see. Let me know when you understand the unet or what the model blocks actually are / where concepts hide in them. I've seen trying myself for a bit.

>> No.7001346

>>7001326
You could just concede anyway. It would save you a lot of trouble. You'd be free. Free to discuss the merger of ai and drawing on forums that want it. Imagine the freedom, taste it. This place doesn't bring you joy, it's clear from your attitude.

>> No.7001351

>>7001340
>it's not about just generating things, it's about adding a human touch and using ai in a photobash-like manner, and discovering other methods of using ai harmoniously without depending too much on it
>ITS SUBJECTIVE THEREFORE IT FITS
Photobashing has never been a thing of discussion because it's you being a hack of an artist. Did you even read the thread yourself? It's literally schizo scribbles to make AI produce. 99% of the rest. And then you photobash. Might as well start tracing too. You never cared about drawing you just hate seeing your "tool" being slandered.

>> No.7001355

>>7001335
>Janny deletes bait threads from fags almost daily.
do you realize how many bait threads get deleted from literally all the other boards? that's not a symptom of ai being a bad topic, it's just /ic/ being too gullible to avoid taking bait from genuine shitposters.
>you let one of them settle and in a year /ic/ becomes the last resort option for /sdg/fags when they get bored on their own boards and get the urge to bumplimit yours.
you can just ban imagedump threads, but still allow threads that talk about ai art adjacent topics that are unique and insightful.
>>7001336
>Clearly we fucking don't so fuck off.
who's "we"? it's an anonymous imageboard.
>>7001339
>my little pony is /tv/ or /co/ related, and yet they had to contain that shit in its own board
false equivalence. my little pony is a show, while ai art is an even more abstract and broad field. they're nowhere the same things.
>the only reason these threads survive is because they're controversial. 90% of the posts are just people insulting the handful (?) of AI evangelists, the other 10% is that verbose autist copy pasting the same "argument".
if they're controversial, it means there's still a topic worth discussing about. why would /ic/ keep responding to the same autist with the same "argument" if there wasn't something new to be said? it seems to me like /ic/ is the problem in this scenario.
>yes, I know I'm part of the problem now. but the moderators can end this very quickly.
then the moderators will end it quickly if it proves to be a genuine problem. you don't have to do their job for them.

>> No.7001357

>>7001355
>you can just ban imagedump threads,
How?

>> No.7001359

>>7001355
>ree false equivalence
The core of all the pro AI arguments, are you unironically this daft?

>> No.7001360

>>7001351
>Photobashing has never been a thing of discussion because it's you being a hack of an artist
i doubt it. photobashing has nowhere the same taboo as ai art. it's probably similar to discussing the referencing/tracing of 3d models.
>It's literally schizo scribbles to make AI produce. 99% of the rest
it's not 99% of the rest. you're capable of adapting how much effort you put in and how much effort you want to retrieve from the ai. it's only a tool in the end. like a paint program.
>Might as well start tracing too.
no, because you seem to think tracing is less taboo than ai art anyways.
>You never cared about drawing you just hate seeing your "tool" being slandered.
now you're just making things up. proof?

>> No.7001363

>>7001355
>anonymous imageboard.
Pretending you cant read the room is silly. It's not just ic, but the majority of artists everywhere that reject ai. ic is no different

>> No.7001364

>>7001357
the same way you're implying how ai threads should be banned
>>7001359
sorry for using the meanie argument words, anon. i'll avoid using it next time.
doesn't change the fact that mlp isn't the same as ai art. they're two different things.

>> No.7001365

>>7001363
>It's not just ic, but the majority of artists everywhere that reject ai. ic is no different
yeah, but /ic/ is not the majority of artists, so they are not relevant to this discussion.
4chan has never been the majority of people, and it may never will be. why would you want /ic/ to be like all the normalfags as well?

>> No.7001372

>>7001365
>why would you want /ic/ to be like all the normalfags as well?
Keeping it ai free is a compromise I'm willing to make if that was the case.

>> No.7001373

>>7001372
so you want /ic/ to be a reddit community? it already exists.

>> No.7001376

>>7001373
>if you don't like AI you're now automatically reddit
Anon, this is an intentionally absurdist take and you and I both know it. Anything else you want to add? Are we trannies now too? Liberal SJWs?

>> No.7001377

>>7001373
Shut the fuck up ai-nigger

>> No.7001380

>>7001365
what a pointless argumentation. "just because it's different let me shit everywhere". don't be surprised if you get spit on like the faggot you are then.

>> No.7001386

>>7001376
just as absurd as banning all ai discussion because an unidentifiable amount of regulars on /ic/ dislike it.
>>7001377
whoa there anon, calm down. if you want me to shut up, you can just stop replying to me.
>>7001380
alright, then we're in agreement. we can post about interesting ai art related topics and you can throw tantrums in return?

>> No.7001387

>>7001360
>photobashing has nowhere the same taboo as ai art.
There are no "levels of taboo". Shit is unethical. No, there is no discussion to be had other than some hack trying to justify it or posting memes about it. You vastly underestimate how much work goes into a complete "traditional" drawing. Hereby goes my accusation that you don't draw much. And tool has ever been creatively autonomous and has never used works of others.

>> No.7001388

Put your trip back on your narcissitic faggot, ironic you call it an anonymous image board yet you feel the need to trip in the first place to shill your slop

>> No.7001390

AI-generated pictures has no artist it thus isn't art
>"bb-but muh prompt"
no artist
>"bbububu muh workflow""
stick to the human-drawn part then, any output tampered by AI doesn't not belong. and anything tripnigger has proven is that any vision he might have had (and I'm being overly generous attributing even a speck of it to him) get tampered by the AI.

>> No.7001396

>>7001387
>There are no "levels of taboo". Shit is unethical.
there are indeed levels of taboo. it's social based.
ai art is simply more in the social presence than tracing is, so it is more taboo. cannibalism is also more taboo than tracing. you wouldn't disagree, would you?
>Shit is unethical
irrelevant. unethical things can still be discussed.
>You vastly underestimate how much work goes into a complete "traditional" drawing.
i don't. i'm just saying this doesn't immediately warrant being against all forms of ai art discussion.
>And tool has ever been creatively autonomous and has never used works of others.
a tool is a tool. you choose what outcomes you want to use. the tool is never the one at fault.

>> No.7001399

>>7001386
>whoa there anon, calm down. if you want me to shut up, you can just stop replying to m
It was a reddit joke. Have a gold kind stranger.

>> No.7001401

>seething and coping intensifies
it never ends bros, when will it end?

>> No.7001404

>>7001386
>>7001386
>alright, then we're in agreement. we can post about interesting ai art related topics and you can throw tantrums in return?
no more than you "can" post porn on a blue board. also
>tantrum
you're an autistic little twat that keeps wanting to shove your shit here and act candid you get pushback. there is no ai art, it is a missnomer, it doesn't belong on a artwork critic board.
>>7001396
"blahblah let me shit everywhere"
this isn't up to debate and your sophistry isn't welcome. why do you think no one appreciates an "akshually" guy

>> No.7001407

>>7001401
When trip gets a clue. (So never)

>> No.7001408

>>7001407
Trip spergs out sometimes, but (you) just egg him on. Why not leave the thread?
>its contaminating /ic/
not really

>> No.7001409

>>7001396
Yeah I can just hear you jumping around repeatedly yelling
>fair use you can't touch me
Yeah you're the pompous fag arguing technicalities and not giving a shit about anything else. Thank you for the oh so fruitful discussion we'll agree to disagree here. Have a good day.

>> No.7001415

>>7001408
I get dopamine from (you).

>> No.7001419

>>7001396
you take us for fools thinking no one sees you coming from a mile away. you'll take allowing one type of ai discussion as allowing all of them.
>tool
you calling it so won't magically make it one. we loop back to the "a tool for what" and you'll answer "art!" like a toddler missing that this isn't an acceptable definition due to being overly vague and could be used to justify just about anything as being a tool.
your argumentation hasn't moved an inch. it keeps relying on accepting a premise because you say it is so.
>inb4 no u

>> No.7001420

>>7001415
At least you admit you're just trolling. Next thread, try to be more eloquent, okay?

>> No.7001421

>>7001404
>no more than you "can" post porn on a blue board. also
ai art is art related, so it belongs on /ic/, which allows discussion of art-related topics. it's not the same as posting porn on a blue board, which is usually disallowed.
>there is no ai art, it is a missnomer, it doesn't belong on a artwork critic board.
ai art is art if it is presented in an art-like manner. it's simple.
>this isn't up to debate and your sophistry isn't welcome.
we're still talking so apparently it is up to debate.
if i'm using sophistry, feel free to point out which parts of my arguments are disingenuous. i'm curious to see your counter-arguments.
>>7001409
>Yeah you're the pompous fag arguing technicalities and not giving a shit about anything else.
this is simply arguing. just because i'm not being hyper contrarian and calling people reddit faggots does not mean im being technical.
>Have a good day.
you too, anon. i hope you've learned something valuable from these discussions.

>> No.7001423

>>6999671
so when are you gonna learn to draw?

>> No.7001425

>>7001419
>you'll take allowing one type of ai discussion as allowing all of them.
i don't think this. it's extremely rare for me to even make threads on any of the boards.
>you calling it so won't magically make it one.
a tool is a tool because it is used like a tool. it doesn't matter what you think is a tool either. what matters is what the artist does with the tool.
>it keeps relying on accepting a premise because you say it is so.
the premise that what the artist does with his tools is what is most important? isn't that what you're trying to argue also?

>> No.7001428

>>7001420
No.

>> No.7001432

>>7001425
>it's extremely rare
Every week isn't rare...

>> No.7001435

Sorry, I just prefer to take the side that doesn't resort to petty name-calling. It's a good indicator.
>grow up
Exactly.

>> No.7001436

>>7001425
You do realize people clearly use it less of a tool and more of an instant image generator which is exactly why you try to force your own discussion here. In vain attempts to make it seem more like a tool. It's fucking sad. This isn't a tool for artists as much as it's a tool for lazy NGMI faggots like yourself either

>> No.7001437

>>7001432
i'm not who you are implying i am. i just post in threads, but rarely make them.

>> No.7001441

Ai turns you into a puppet. Doing it's bidding. Soon you will find yourself working within it's box, tiring of it interpreting things wrong you will learn not to push things that far. Already evidence of that itt.

>> No.7001443

>>7001437
Sorry. You smelled just like him.

>> No.7001446

>>7001436
>You do realize people clearly use it less of a tool and more of an instant image generator
the fact that a tool can be used as an instant image generator does not disqualify it from being used in other manners.
/ic/ isn't the same as other people, so what the majority of other people do with a tool shouldn't be used to disqualify that tool.
i genuinely believe /ic/ is capable of using such tools in manners that show the skill and effort of the artist.
> In vain attempts to make it seem more like a tool.
it doesn't have to seem like a tool. it's appearance as a tool is only relevant once an artist uses it like a tool.
i simply refer to it as a tool because it's convenient to call it a tool in these scenarios.
>This isn't a tool for artists as much as it's a tool for lazy NGMI faggots like yourself either
it can become a tool once it's used like a tool.

>> No.7001448

>>7001421
>ai art is art related
nah
>ai art is art if it is presented in an art-like manner. it's simple
no, no more than a sunset is or a painting by a monkey is.
>we're still talking so apparently it is up to debate.
goddamn you're at the kindergarten level of "nu-uh" and try to win through attrition
it isn't up to debate, it has been discussed at nauseam and you've been told at length why you're wrong. look through an archive site if you want.
> just because i'm not being hyper contrarian
you're a fucking contrarian, have no illusion about it
>>7001425
>i don't think this. it's extremely rare for me to even make threads on any of the boards.
not fucking fooling anyone
you're not using it like a tool, checkmate, get wrecked. all your argumentation is based in it being a tool and it's output being art. the former you have yet to prove and the later isn't possible by definition
>>7001437
again not fooling anyone but even if that were true to participate in shitting things up so you might as well have.
>>7001435
you do realize acting like the reasonable party doesn't really add weight to your argumentation? being an autist and irritating people to no end is bound to get you backlash

this is going to be the last (You) from me. I am right, you are wrong and a fag. I am right because I have a soul imbued with divine knowledge and you don't. Yes I used the word on purpose facetiously because it baffles your autistic ilk. I don't need to elaborate anymore, you have proven time and time again you won't budge on any point while failing to convince. now go cry about the mean words being used.

>> No.7001449

>>7001448
So much seething anon just take a break from 4chan, okay?

>> No.7001450

>>7001446
you haven't at any point used it like an art tool in those threads. you use it to avoid learning key concepts because you are lazy and no amount of relying on your toy will give you the necessary knowledge to do something similar to its output.

>> No.7001453

>>7001449
I accept your concession on his behalf

>> No.7001454

>>7001446
I genuinely believe /ic/ is capable of using such tools in manners that show the skill and effort of the artist.
There you go again forcing yourself onto a community and trying to bend them to your beliefs. Do you know how egotistical you always sound? Have it ever occurred to you that a community of people that love drawing aren't interested in a "tool" that takes away a lot of that effort from them? Why do you keep persisting otherwise. Seriously, why keep on persisting on /ic/ instead of discussing it in actual places that welcome it? You're running on nothing but pure egoistical narcissism

>> No.7001455

Commending tripfag for getting anon to seethe this much. They just won't stop replying.

>> No.7001457

>>7001448
>nah
i disagree.
>no, no more than a sunset is or a painting by a monkey is.
i don't disagree. a sunset can be art related. a painting by a monkey is genuinely art related.
>it isn't up to debate, it has been discussed at nauseam and you've been told at length why you're wrong. look through an archive site if you want.
if it's not up to debate, then why are you debating? i'm merely stating the facts. if it's not up to debate, then you don't have to reply to me.
what happens in the archives is irrelevant. i'm not going to stop discussing things because somebody might've discussed something similar in the past.
something new can always be brought up that past discussions haven't covered.
>not fucking fooling anyone
im not trying to fool anybody. i'm not the tripfag.
>you do realize acting like the reasonable party doesn't really add weight to your argumentation?
why do you keep thinking i'm trying to play the reasonable party? i'm just not devolving to mindless insults.
i know it's unusual to see in a post-contrarian board culture, but this is genuinely what most internet arguments were like before normalfags took over.
>being an autist and irritating people to no end is bound to get you backlash
i'm not trying to irritate people, i'm simply responding to their replies. if they don't want my replies, they should stop replying to me. it's that simple.
it must be so frustrating to not be able to see simple solutions to simple scenarios.
>I am right, you are wrong and a fag
i disagree. saying you are right does not mean you are right.

>> No.7001458
File: 697 KB, 1572x1887, GCxEfopXgAAbpQj.jpg [View same] [iqdb] [saucenao] [google]
7001458

might as well post something relevant unlike 99% of this thread

>> No.7001462

>>7001458
You realize the only reason this case exists is because they are jealous that AI company is making money while they aren't, right? They do not care about morals lmao. Good luck to them arguing against the people who created the thing.
>it contains compressed copys of copyrighted work

>> No.7001463

>>6999671
>assisted
lol OP is mentally handicapped

>> No.7001467

>>7001450
you are also doing the same by responding to me and not grinding your art, no?
>>7001454
are you disagreeing and implying /ic/ is too incompetent? how does that help your argument at all?
i have respect for /ic/'s skills. that's why i think it's strange how they feel threatened by a simple tool.
>Have it ever occurred to you that a community of people that love drawing aren't interested in a "tool" that takes away a lot of that effort from them?
irrelevant. as i've said before, you can still discuss the usage of a topic that can bring harm to your manner of living. do you think people avoid discussing wars because war kills people?
they discuss these things because it's better than being in the complete dark about them.
>Seriously, why keep on persisting on /ic/ instead of discussing it in actual places that welcome it?
you mean hugboxes? i just don't think it's interesting that way.
it's good to talk about controversial topics sometimes. they aren't going to kill you.
>You're running on nothing but pure egoistical narcissism
that's simply how you perceive contrasting opinions. not everything that disagrees with you has to come from the ego.

>> No.7001470

>>7001463
It's ok, I use waterwings at the beach too.

>> No.7001473

>>7001467
>by a simple tool
Disingenuous
>you mean hugboxes? i just don't think it's interesting that way.
So you're just in it for the (you's)
>not everything that disagrees with you has to come from the ego.
But you clearly have a massive one
As everyone has seen by all your previous threads and I'm sure you've more to admitted to on more than one occasion.
You've done more damage to AI's credibility here than anyone or anything else has. It's ironic how you can't take a step back and look at the bigger picture of your actions

>> No.7001477

>>7001473
>Disingenuous
okay, a complex tool; but a tool nonetheless, and a tool that /ic/ can still overcome.
>So you're just in it for the (you's)
no, i'm in it for the discussion.
>You've done more damage to AI's credibility here than anyone or anything else has. It's ironic how you can't take a step back and look at the bigger picture of your actions
doesn't matter to me. what do you think my goal even is?
i just want to discuss ai art here, even if other posters don't seem to like it.

>> No.7001484

Let's discuss ai art
>I use it to make ai render parts of a picture
>Actual discussion already exhausted after 10 posts
>Try again next week with the same shot

>> No.7001517

>7001477
go to reddit then for an discussion tripfag

>> No.7001525

>>7001467
>I have respect for /ic/'s skills
writes an near 50pg dissertation to use AI and want to seem reasonable after niggas wanted to spam "it's over" posts
lurk more you fagget or stop sucking your own dick
>inb4 it's just trip/mikufag
the biggest NGMI niggas /ic/ ever seen

>> No.7001532

>>7001477
>what do you think my goal even is?
not that anon but you sound like a marketing firm in which case nothing you say is so value, you basically are the same as seeing chinese bots or vatniks
it's very a much a (YOU) problem that your hated

>> No.7001614

>take a break
>thread fills up immediately
jesus, i don't even care to argue anymore at this point lol. can't you tards take this somewhere else

>>7001006
>And you are equating it to "it's the same as if you looked and learned from it" when it's false equivalence in every sense but the name.
but it is the same.
if you look at how training actually works,
it is eerily similar to how studying works. the raw images aren't kept, what is kept instead is the ABILITY to make the images using the AI's chosen method (denoising).

what you're saying would apply if the AI was a giant collage machine that puts together pieces using people's literal work.
but as i'm saying over and over again, this is not how the tech works.

if you had two braincells to rub together, you'd realize that no collage machine could possibly make anything coherent.

>>7001090
it has nothing to do with leeway.
before AI, this wasn't even something people thought about. that you'd need to ask permission to learn something that is out in the open is ridiculous.

your argument goes both ways. because you'd need a reason for disallowing its use too. it not being a human isn't actually an argument for anything.

>> No.7001623

>>7001614
>can't you tards take this somewhere else
Ironic. No apparently it has to be here!

>> No.7001627

>>7001614
>can't you tards take this somewhere else
https://www.youtube.com/watch?v=LyZv4EghX_I

>> No.7001629

>>7001614
>can't you tards take this somewhere else
i chuckled

>> No.7001632

>>7001614
different rules are applied to things that are not human. Most of of which of which goes back to how people actually use it

>> No.7001633

>>7000582
>mikufag

>> No.7001634

>>7001614
>take a break
maybe take a longer break of about a couple months or year retard

>> No.7001636

>>7001436

>> No.7001644

woah

>> No.7001645

>>7001614
>that you'd need to ask permission to learn something that is out in the open is ridiculous.
confidentiality is a thing tripfag, just because you see something doesn't mean you should have the means to recreate it on mass

>> No.7001646

>>7001644
oh

>> No.7001648

>>7000533
huh...

>> No.7001650

>>7001614
>jesus, i don't even care to argue anymore at this point lol. can't you tards take this somewhere else
doesn't want o argue contentiously you's multiple people still

>> No.7001652

>>7001646
yeah

>> No.7001654

>>7001650
hey man it's okay

>> No.7001659

>>7001355
>false equivalence
if AI that abstract and needs multiple threads about it, then it needs it's own board than more not less than retard. AIniggas couldn't even keep it to one thread even, and the tripfag making this thread when AI fags began spamming again is his fault honestly

>> No.7001662
File: 10 KB, 220x167, Kool-Aid_Man_breaking_wall_1978.jpg [View same] [iqdb] [saucenao] [google]
7001662

>>7001652
indeed

>> No.7001666

>>7001659
>needs

>> No.7001671
File: 370 KB, 1312x1226, 1691959090398359.jpg [View same] [iqdb] [saucenao] [google]
7001671

>>7001650
i just mean it in the sense that i won't put as much effort as i used to.
to me, it really only comes down to understanding a few things, and if you don't get those, you won't get it.
and as it happens, most people refuse to understand it even when explained to.

the best course of action is just showing you exactly how AI can be used and what i mean by all this. and then you can ask your lying eyes whether i'm telling the truth or not.


>>7001645
again, it is not recreating the training data.
do you know what "curve fitting" actually means? it means that it is trying to create something that statistically falls into place AMONG the existing data.

once again, your language suggests that you think that AI is copying shit.
how do you reconcile that point of view with the reality that AI is creating images that are not and cannot be copies of existing art?
contextual understanding of lkighting, reflections colors, none of these are static things that can be just copied. in your mind, how is this supposed to work?

>> No.7001672

>>7001467
>i have respect for /ic/'s skills
want to be able to ignore most of them, and seeing how AIniggas use it they don't exactly want to be critiqued

>> No.7001674

>>7001671
>it really only comes down to understanding a few things, and if you don't get those, you won't get it.
I could tell you to fuck off yet you don't seem to get that

>> No.7001675

>>7001671
>i just mean it in the sense that i won't put as much effort as i used to.
Ah, and there it is. You're just lazy. "Me want art now."

>> No.7001677

>>7001675
>reading comprehension
i meant effort in trying to argue with people.

but leave it to retard kun over here to jump on it like a vulture before actually underestanding what i wrote.

>> No.7001678

>>7001671
I don't care about you though and honestly I don't consider the AI part even yours to begin with. I might as well image search on Artstation or other places because I could give less a fuck about your "creations"

>> No.7001679

>>7001677
no one cares about what you write besides to mock you because you wanted to be that annoying

>> No.7001681

also you do relize some of those reflections are siuper fucking off the fuck you wanted to show.
the top middle has long hair but the refection has a pony tail and the top left is all fucked up zooming in
>looks good enough ship it
I could you actually get your eyes cvhecked dumb ass

>> No.7001684
File: 45 KB, 258x361, reflections.jpg [View same] [iqdb] [saucenao] [google]
7001684

>> No.7001685
File: 43 KB, 300x489, reflections retarded.jpg [View same] [iqdb] [saucenao] [google]
7001685

>> No.7001694

>>7001681
It's clear he's just more enthralled by tech than actually drawing. This anon is right
>>7001675 and the tripfag has admitted to it multiple times in past threads. We don't need more people like him on /ic/

>> No.7001700

The more you tell me I'm not an artist and this work is not mine, the more it fuels me.

>> No.7001705

>>7001679
yes. i call it "seething"
that's all you can do at this point.
you refuse to understand.
but you also refuse to educate yourself. so all you have is blind seething rage.


>>7001681
>>7001684
>>7001685
yes yes yes, same with wobbly hands, and all other errors AI can make. i already call them "erroneous" into the pic in big red letters in case you missed it.
but do YOU understand what it means that it even attempted it in this way in the first place?

if you believe that it plagiarizes and copies from the trainingdata, do you think it copied both the character and the mirror/reflection? who happen to be the same?

>> No.7001712
File: 3.24 MB, 1824x1248, 1678928353795712.png [View same] [iqdb] [saucenao] [google]
7001712

in case you don't understand, i keep bringing up mirrors and reflections, not because the AI is any good at it, but because things like this have CONTEXTUAL dependency that you can't just copy from another image.
to me it is irrefutable proof of its ability to generalize. as opposed to copying.

and this is just the obvious example. in reality lighting and color itself are also contextually dependant and can't just be ripped from images.

>> No.7001728

>>6999671
https://arxiv.org/abs/2301.13188
>Extracting Training Data from Diffusion Models

>> No.7001733

>>7001700
ok

>> No.7001738

>>7001677
you don't get to ask for fair argumentation when you have made yourself thoroughly insufferable

>> No.7001740

>>7001462
sounds like cope :^)

>> No.7001741

>>7001728
maybe read the actual paper.
see how much they managed to extract, percent wise.
and what they had to do to the model to even be able to do this much.

>>7001738
ah ok. is that how you justify being an insufferable retard who seethes at me blindly for all of eternity?

>> No.7001747

>>7001741
>act like a flaming autist over several months
>wtf anons are mean to me :'((
lol

>> No.7001748

>>7001741
>blindly
the victim complex on this clown

>> No.7001749

>>7001740
I accept your consession :^)

>> No.7001761

>>7001748
>>7001747
i'm saying the same shit i've been saying from the beginning.
the only difference is i'm not insulting you people as much anymore.
which i only did in the first place because i was returning insults.
you people simply have no self awareness.

>inb4 no u!
nah. i'm aware when i resort to insults. and i at least deliver my points WITH the insults on top.
i'm not sorry for any of it.
the difference between me and you is that you think you're angels. as if you didn't argue in bad faith and insults from the very beginning.

>>7001747
you are literally crying right now about me being mean to you for months.
like i said, no self awareness whatsoever :)

it's a complete joke, and this is why i focus on showing rather than telling now with these threads.
keep up the seething.

>> No.7001766

>>7001761
People don't care how it truly works, /ic/ still doesn't want to use it. So good job for arguing for over half a year the same shit over and over again

>> No.7001782

>>7001766
that is absolutely fine.
but then you don't get to talk about it plagiarizing or stealing from artists without me talking back.

you don't get to say something dumb and when someone corrects you just go
>i don't care lol

yes, you don't care. most of you would rather live under this delusion than accept the simple reality. the reality that the AI can legitimately make art on its own that is not its training data.

>> No.7001787

>>7001741
Just like the diffusion algorithms improve over time as research advances, so to will the adversarial algorithms for extracting the effectively lossy compressed data out of them.

>> No.7001790

>>7001782
If only you could deal with that fact and stop trying to shill it here then lol

>> No.7001800

>>7001787
you know what lossy means, right?
you seem to be under the impression that you could extract whatever you want eventually.

but this is more of a security issue than a plagiarism issue.
again, look at >>7001728 this paper and see how much they actually extracted, and what they had to do to even extract that much.

of course overfitting will always be an issue. but it's not the goal to overfit, it's the exact opposite, so it will always be a minor thing.
but then again midjourney does show that there is overfitting is a tricky thing, especially when met with movie frames in slow scenes where a lot of the frames are samey looking.

>>7001790
deal with what fact? do you still not understand that i'm posting here BECAUSE artists would rather stick their heads in the sand?

>> No.7001807

>>7001800
Yes, I do know what lossy means. You won't create a perfect replica of the input data, but nor does saving a picture as a jpeg. I am under the impression that they will be able to extract more training data from networks eventually, specifically. People said the same thing about image generation itself in the first place, and even seemingly unbreakable encryption before that. Ultimately, it doesn't really matter how much they extract, just that they could get a non-zero amount, regardless of what preparation was necessary, which can help bolster support for a legal precedent on plagiarism against the people who generate, distribute and use models.

>> No.7001818

>>7001800
>stick their heads in the sand
typical techbro. you're doing a lot of head burying yourself trying to convince yourself you have any artistic legitimacy. you are too lazy to do the legwork and chose the easy path. heck you don't even admit using such a software is ethically
if not morally dubious nor do you entertain the far ranging consequences this technology might have. nay, you'd rather paint yourself as the martyr missionary eager as you are to drag us into the same depths and lows as you.

>> No.7001820

literally not reading any of this

>> No.7001821

>>7001800
There's your ego again. Keep trying to be the messiah of AI for /ic/ while talking down on others. It's not pathetic at all

>> No.7001825

>>7001782
>talking back
by all means, keep doing your mental gymnastics to think using those is A-okay. keep comparing apples to oranges
>n-no I'm not!

>> No.7001834

>>7001787
True. This is just the first wave of AI and it's not that good. In 5 years an ai not being able to spit out training data verbatim would be laughable.

>> No.7001835
File: 115 KB, 1351x638, EwA4jzxVoAAeCSU.jpg [View same] [iqdb] [saucenao] [google]
7001835

>>7001761
>you are literally crying right now about me being mean to you for months.
goddamn you're stuck in an alternate reality. I'm telling you the autistic retard that your clueless behavior has lead to the consequences you see today and the only thing you can spew is a feeble "no u". you were not mean to us but an annoying mule headed autist.
you're the embodiment of picrel

>> No.7001838

>>7001821
desu I'd be more willing to listen where he, you, know, good at drawing?

>> No.7001859
File: 2.12 MB, 1000x1200, lowres, low quality, worst quality,--_00063_.png [View same] [iqdb] [saucenao] [google]
7001859

>>7000285
if you're still around, have you ever tried plasma sampling? picnotrel but the last time i tried it, it was very keen on giving just a character with a plain colored background.

>> No.7001974

>>7001787
That's not how it works.

>> No.7001991

>>7001339
we should have an /ai/ containmenboard

>> No.7002021

>>7001825
like i said, if you don't get it, you don't get it.
but it's more that you don't want to get it.

you call it apples to oranges but you couldn't tell me what the difference is if your life depended on it.
it's just based on the blind belief that it is copying.
and you think my stance is based on the same blind belief like yours, instead of knowledge of how the AI actually works.

>>7001821
it's not my ego. i'm talking down to you precisely because you're living in a collective delusion.
you feel fine about it because the other retards next to you are spouting the same dumb shit.
and collectively you get to feel righteous and superior about something that simply isn't true.

>>7001835
the consequences of you being in a seething rage everytime you see me post?
that's you anon. that's just you being angry over something you don't understand.
that's why we're in this stalemate. you have nothing other than anger. because you refuse to educate yourself.
just blind belief.

>>7001818
i do think it has far reaching consequences. but since it's not as ethically dubious as you say, i don't consider the the entire thing unethical.
and after that, the only thing left is the inevitable change that the tech will bring about.

>>7001807
again. look at that paper. i don't remember the exact number, but it was a tiny amount they managed to extract, and that is within a changed set up that was deliberately trying to find those.
without doing what they did, it's practically impossible to get extractions like that.
so outside of overfit examples, it is pretty much irrelevant for day to day use.
because the people making these models don't WANT this. they WANT generalization. that's the "intelligence" part of it all.
it's more likely that people will get better at training rather than retrieval. because that's what people actually want.

>> No.7002026

>>7001834
AI spitting out training data IS laughable, do you even get that? it's considered a failure.
why would the field move towards that?

>>7001859
never heard of it before. i'll look into it.

>> No.7002035 [DELETED] 

>>7002026
>AI spitting out training data IS laughable, do you even get that? it's considered a failure.
>why would the field move towards that?
to stop hallucinating when it is needed. multimodal AI knowing exactly what x looks like is ideal. "design a character wearing clothing from 1785"

>> No.7002040

>>7002026
>AI spitting out training data IS laughable, do you even get that? it's considered a failure.
>why would the field move towards that?

to stop hallucinating when it is needed. multimodal AI knowing exactly what x looks like is ideal. "show me a person carrying x painting into a museum".

>> No.7002058
File: 2.11 MB, 1791x1910, 1697947663003773.jpg [View same] [iqdb] [saucenao] [google]
7002058

>>7002040
>to stop hallucinating when it is needed. multimodal AI knowing exactly what x looks like is ideal. "show me a person carrying x painting into a museum".
you don't realize how limiting that is. i find it funny how you think copying something is actually a boon.
but it would just be photobashing. and we all know how fucking shit that is without manual overpainting.

if i want the AI to make a "cat that is leaping", i'd want it to make ANY cat that is leaping, not one specific cat in one specific angle that exists in one specific training image.
same with your example. i'd want to make ANY person carrying a painting into ANY place. what good is it if there is one image that happens to meet your criteria, what if i want
>a woman carrying an painting into a pizzaria
instead, and that image doesn't exist? then i'm just shit out of luck???

you make the same mistake that everyone here makes, you think the AI works by recalling its training data.

look at pic related for example. you think AI plagiarized something to make this?
guided examples like this make it blatantly obvious that AI isn't simply copying its training data.

>> No.7002078

>>7002058
I make no mistake, that's your overactive imagination.

>making a film
>make lots of pre-production work
>feed it to the 2030 AI
>request something specific from the data we fed it to appear in a scene

>> No.7002095

>>7002078
you're still making the same mistake.

for example, if i wanted to train on a specific painting in order to have it appear as an object in a movie, then i would still want it to be understood GENERALLY. instead of it being understood pixel by pixel, which is what it means for the AI to spit out training data.

for example i want it to look right in different light conditions. and not just the way it looks in the training data.
in no way is the model giving you back the training data ever a good thing. unless the training data just happens to be exactly, and i mean EXACTLY what you want, down to the lighting and angle and everything else.

>> No.7002101

>>7002078
I kind of get what you're saying, but this is not the same as it giving back the training data.
it's more akin to when the AI learns the face of a celebrity. it doesn't learn specific photos, but memorizes the face in a more general sense.

>> No.7002102

>>7002095
>unless the training data just happens to be exactly, and i mean EXACTLY what you want, down to the lighting and angle and everything else.
if it can, that's a sign of perfect understanding of the training data. powerful AI should be capable. You don't want it to make mistakes under different lighting.

>> No.7002106
File: 2.38 MB, 1000x1200, lowres, symmetrical, ugly, frame, paint, linework, b&w, flower, green, fur , circle, complex, pyramid, yellow, illustration, render, earth, purple, sketch, intricate, 1girl, 1boy, human, person, figure, _00223_.png [View same] [iqdb] [saucenao] [google]
7002106

did i miss the part where a model spitting out its training data is any different than copying someone brush strokes verbatim?

>> No.7002110

>>7002106
did i miss the part where a model spitting out its training data is the same as copying someone brush strokes verbatim?

>> No.7002121

>>7002110
>>7001728
why do we care if it can, then?
>that means it's copying
no but really tho

>> No.7002152

once again pointless arguing around technicalities because a handful of begs just don't want to learn. you can babble all you want about technicalities of your toy it matters little as you miss the bigger societal picture. get off your high (tin) horse and lose that savior complex.

>> No.7002155
File: 1.93 MB, 1000x1200, lowres, bad quality, b&w, fabric, fur, avian, torch, complex, pyramid, yellow, illustration, render, earth, purple, sketch, intricate, frame, man, woman, human, figure, person, body, face, _00006_.png [View same] [iqdb] [saucenao] [google]
7002155

>>7002152
>technicalities
what do you think they talk about in court?

>> No.7002157

>>7002021
>do think it has far reaching consequences. but since it's not as ethically dubious as you say, i don't consider the the entire thing unethical.
because it is convenient for you to think so as you are a lazy twat. you have brought absolutely nothing of artistic value and would rather argue endlessly why it's okay for continue that dubious path. had you actually considered the consequences you'd have realized how it induces a rot in art.

>> No.7002162

>>7002155
why do you think it is in any way relevant? you're still not doing the work nor does it improve your artistic skills.

>> No.7002164

really lose that self-righteous attitude. you have no high ground.

>> No.7002166
File: 575 KB, 3037x2482, Scene.jpg [View same] [iqdb] [saucenao] [google]
7002166

>>7002162
anon brought it up as if to say ai is done for, yet once proven wrong its "you're just arguing technicalities"
>you're still not doing work
lol okay

>>7002164
there are certainly those in these threads with a high and mighty attitude and it is not me, if that's what you're suggesting. lots of anons positing absolutes while knowing little about the subject.

>> No.7002170

>>7002166
>anon brought it up as if to say ai is done for
no Im saying you're arguing meaninglessly. there's no "ai is done for" there's just you not being artistic and wanting some anonymous people on the internet to acknowledge that you are.
>picrel
look at all that non art that you want people you say is art that you did not do but want people to say you did.
>there are certainly those in these threads with a high and mighty attitude and it is not me
that'd be somewhat believable had you not accused artists of collective delusion and put yourself up as a higher authority.
>b-but other mean anons are doing it
maybe stop trying to make people purple is blue because they both are born from light interacting with a surface.
but we're back to this again you'll keep arguing and keep refusing to learn the craft.

>> No.7002171

and in case you don't get it I am calling you a non-artist and this needs no further debating as your numerous rants damned you as one.

>> No.7002173
File: 2.30 MB, 1000x1200, lowres, low quality, worst quality, b&w, bone, gold, --_00034_.png [View same] [iqdb] [saucenao] [google]
7002173

>>7002170
>wanting some anonymous people on the internet to acknowledge
i really dont care what you think about the images i post
>accused artists of collective delusion and put yourself up as a higher authority
can you link me to where i said this? you are most likely confusing me for another anon.

>> No.7002174
File: 522 KB, 1280x720, SteamedHams.jpg [View same] [iqdb] [saucenao] [google]
7002174

>>7002171

>> No.7002178

>>7002155
you're just begging for new laws to be socially tailor made for that case just to stamp out the behavior.

>> No.7002181

>7002173
>i really dont care what you think about the images i post
>said the attention whore as she attention whored and begged for others to dig through threads upon threads so she can keep argue

>> No.7002184

>>7002181
nice reply

>> No.7002187

>>7002184
as of we needed even more proof you just want (You)s

>> No.7002189
File: 2.25 MB, 1000x1200, lowres, symmetrical, ugly, frame, paint, linework, b&w, flower, green, fur , circle, complex, pyramid, yellow, illustration, render, earth, purple, sketch, intricate, 1girl, 1boy, human, person, figure, _00078_.png [View same] [iqdb] [saucenao] [google]
7002189

>>7002178
it will be interesting to see its progression for sure. a non-copyright lawyer told me he estimates laws being on the books in the matter of three years. i wonder how well that prediction will hold up all things considered.
>you're just begging
yeah, we could all be wrong about how it plays out, but i like to hedge my bets anyway - but im sure at heart we all hope it benefits us regardless of how we think the benefit will come to fruition

>>7002187
ive just never been so upset i went out of my way to make my quote post not quote the anon im responding to

>> No.7002218
File: 2.00 MB, 1000x1200, lowres, 1girl, 1boy, body, human, person, figure, face, skin, head, subject, symmetrical, ugly, frame, paint, linework, b&w, flower, green, fur , circle, complex, pyramid, yellow, illustration, render, earth, purple, sketch, intricate,_00007_.png [View same] [iqdb] [saucenao] [google]
7002218

>>7001672
>they don't exactly want to be critiqued
i love it when people tell me what they feel im doing wrong - as a tame example, its common for images to fry because one gets boiled in the CFG like a frog. and thats important to realize when its happening. i think knowledge of the what the anon is using is helpful, and i dont mind being called a fag as long as theres a critique laid in there.
saying the hands are bad, or a specific detail being off is meaningless because one could inpaint it away most times. that is an example of someone critiquing without really understanding - you should be telling that to the groups pouring loads of money into training commercial models, not the random anon asking for critique.

>> No.7002255

>>7002157
quite the opposite. i actually considered the consequences.
you cannot even consider them properly because you cannot assess what is happening. in fact most you retards are still in denial about the very nature of AI.

as for the "rot", i actually believe in art and artists. i myself started practicing again more due to AI. i know of one model maker who started painting for one reason or another. possibly because he was looking at so many paintings, both real and generated over the course of making his model.
because AI is limited people will always still have reason to learn manual art. and there is reason to believe it will always stay limited, simply becaue it can't read your mind.

it's easy to be a braindead doomer who has nothing to say and can see nothing but destruction in the face of change.
very telling posts in general because you cannot say anything of value. only throw insults and bemoan how oh-so-bad things will be.

>> No.7002265
File: 957 KB, 1024x688, 1681913792935440.png [View same] [iqdb] [saucenao] [google]
7002265

>>7002102
i don't think you understand what is meant when people say "give back the training data".

if it can adjust an object under different lighting and contexts, then it is not simply giving you back the training data. it has an understanding then, an ability to apply what it has learned in a "intelligent" way.

one is just copying an image of let's say a cat, the other is understanding what a cat looks like GENERALLY from seeing millions of cats.

one is just copying one specific photo of morgan freeman.
the other is understanding his facial features from looking at thousands of photos of him.

it's all pattern recognition.
one is copying, the other is learning.

>> No.7002287

>you cannot say anything of value. only throw insults
holy fucking shit. the constant projection

>> No.7002298

>>7002121
Selling access to a tool full of shit you have no right to

>> No.7002300

>>7002265
I'm saying what people could expect in 2030. The meaning would change with the new capabilities of advanced ai

>> No.7002320
File: 218 KB, 512x512, 1690010895332409.png [View same] [iqdb] [saucenao] [google]
7002320

>>7002300
it wouldn't change. there is simply no benefit to getting back the training data verbatim.
ironically what you're describing is just an advanced search for a database.
if you could use the training data verbatim, then you wouldn't need AI in the first place.

and that means it is doing things """"intelligently""""
even if currently that means that when you prompt for a "pile of coins", you get this pile of shit instead of a photo from the training data that happens to perfectly fit your requirements.
(not that the prompt can't work, this is just a funny example i found in /sdg/)

>>7002287
have some self awareness anon. do you even read your own posts?

>> No.7002341

>>7002320
>there is simply no benefit to getting back the training data verbatim.
verbatim is too strong a word.. but to test it is accurate, getting back accurate reconstructions of images in the training data would be useful in being confident it wont mess up when you are requesting things that require accuracy.
>Verify you are human

>> No.7002354

>>7000589
This, is honestly kind of disgraceful how twitterniggers can easily band together to tell AI "artists" to go kill themselves but here on 4chan we are forced to have a place for these parasites because mods are cock gargling faggots.

We have worse community standards than fucking Twitter and furaffinity, the absolute state of 4chan /ic/

Its great to see these fuckers being driven away from the rest of the internet tho, I just wish there was a containment thread on /ic/ of all places.

>> No.7002363

>>7002341
but it's not an accurate reconstruction, it's a literal copy.
a copy not of the entire object, but just the single image of it.
i think we're on the same page but you're just wording it wrong.

in terms of LLM, what you're saying is like you WANT the LLM to copy a phrase/text instead of gaining an understanding of the text and the ability to work with it in a natural way.

>>7000589
>>7002354
and why should it not be normalized? what do you think will happen? it's out there. it will only get better.
this is what i'm talking about with the "head in the sand" attitude.

>> No.7002370

>>7002363
Go to /biz/ with the nft retards or to /g/, this shit isn't drawing.

You literally spend time doing worthless "art" that nobody cares for because you didn't fucking do it, if I wanted to see slop I could go to prompt hero and see a fucking wall of it, WHILE also being slower and less productive than a fucking proompter paying Dall-E, since he is much faster than you.

Literally the worst of the two fucking worlds.

>> No.7002380

>>7002370
you can see exactly what i did or didn't do itt.
but keep repeating your delusional mantras.

>> No.7002479

>>7002354
Go back to twitter then lmao

>> No.7002512

>sees thread i do not like
>enters thread i do not like and tells anon to kill himself
>complains about thread i do not like and safe spaces
>continues to reply to things i do not like

>> No.7002525

>>7000285
You purport yourself to be so knowledgeable about its operation, yet you don't seem to understand that it doesn't have an attention to speak of and instead makes plausible adjustments at every pixel position simultaneously, based on its neighboring regions, for each iteration. You change the bias for the operation when you add a colored background most likely because a large proportion of the training data is with white backgrounds. Therefore it would be beneficial for you to understand what it's actually being trained on to have an expectation of its regional biases.
Given the results for this instance, I'm not sure what this achieves that you couldn't have done with a blending brush. You may as well let it go all the way like your prior examples and the paint over it to get the maximum benefit.
I'm also not sure that you guys will ever understand the satisfaction of creating something with your own hands. It's sad honestly, but any appeal to pathos will be missed despite the topic supposedly being about "art". I wonder though if this is a thinly veiled computer science thread.

>> No.7002554

>>7002525
fair enough about the attention thing, it was really just a wild guess. i was aware that it might even just be placebo, since the empty section makes the detailed sections stand out more.

although i wouldn't completely disregard the idea.
there is obviously a limited amount of attention when generating, otherwise you'd get perfect faces even when they are just a small section of the image, but that is never the case. and i doubt the attention is equally spread across the image. the attention on the prompt isn't equal either (with the first words being stronger), so i don't see why the attention on the image itself would be perfectly equal.

it's still a wild guess. actually i could test this out by making a full body gen with a background versus a white bg and checking if there's any qualitative difference.
can't do it right now but will try later.
it's probably just bullshit tho lol.

>You change the bias for the operation when you add a colored background most likely because a large proportion of the training data is with white backgrounds.
i highly doubt that. at least i don't think the large majority of artwork is something on a white background.
but yes, any changes would affect the bias for the generation, including a white background.
for example a black background bias itself towards chiaroscuro so you might get better values, who knows.

>muh joy
personally i find much more joy in drawing than painting. working like this is unironically more fun.
although i agree that there is something to be said about setting everything down yourself and accepting the quirks and accidents you make. but really it doesn't feel that different.

i already painted over stuff and then fed it back to the AI. also every mask is completely painted over since the AI could only get it half of the time.
the joy is still there once i refine it. for example, once i painted over the mask, the only thing the AI contributed to that section are hints for the color and values.

>> No.7002560

>>7002554
>otherwise you'd get perfect faces even when they are just a small section of the image
isn't this because latent space is just compressed pixel space as in a supreme loss of detail? not anything to do with attention

>> No.7002566

>>7002560
yeah probably.
still, attention does exist for the image as well. it checks the image with the input as it generates. and if one input section is a pure color then obviously there is less to do compared to those sections that still look like blurry undefined blobs.

but whether other sections actually benefit from that is dubious at best. so eh.

>> No.7002575

>>7002566
ive never heard anything about attention being spread evenly or unevenly, so i just presume attention is spread somewhat evenly (as even as the noise or image you feed it? probably also how you choose to sample?) but it sounds like you've encountered something to make you believe thats not the case. i think the test you describe in your other reply is a good idea and i would be curious to see the results.

>> No.7002589

>>7002575
well, the idea of attention itself is that it selectively focuses on things that are more important.
like in a sentence, a LLM wold focus more on the nouns and verbs or something like that, because they matter more (it's probably not actually nouns or verbs, but some words definitely get more attention).
the real question is just if attention is limited, or limited enough for something like this to matter.

i'll give it a try tomorrow or next thread since i'm using my PC for other work currently.

>> No.7002673

>>7002554
bruh you can't even draw well

>> No.7002926

I havent been on /ic for a while but why are AI threads like this still up

>> No.7003120

>>7000317
>>7000437
KEK the absolute state. If you actually knew how to draw and practiced it, you could shit out a fully rendered good looking piece in 6 hours, but you decided to settle on melted dogshit

>> No.7003157

>>7003120
what if he likes "melted dogshit"? it probably looks better than your art

>> No.7003218
File: 1.22 MB, 765x2365, 1688455072638644.png [View same] [iqdb] [saucenao] [google]
7003218

>>7002926
mods are trannies and not actual artists

>> No.7003272

>>7003157
>t.mikufag

>> No.7003339

>>7003218
you're right, she doesnt even have ".artist" in her handle

>> No.7003361

>>7003120
I don't get it, it's really not a good way to use AI at all. And this guy has supposedly been on /ic/ for 15 years?

>> No.7003386

>>7002298
>you have no right to
why? because you have the potential to generate a preexisting IP?
>it steals from artists
no but really tho

>> No.7003555

>>6999671
this is a product
go to /biz/

>> No.7003672

Test test. Is this thing on?

>> No.7003680

>>7003672
Test complete

>> No.7004082

>>7001110
>article
why would anyone care? Victimless "crime".

>> No.7004193
File: 139 KB, 400x600, 1665328050695038.png [View same] [iqdb] [saucenao] [google]
7004193

okay so what are the steps to actually using this thing?
i know it's stable diffusion because of how shit it is but what other tools do you use? controlnet? a krita plugin?
also post some models please

>> No.7004219

>>7000437
>>7003120
pyw
>inb4 "erm you first sweetie? (:"
i'm not using ai. pyw, i want to see if it's better than the slop you think you're above.

>> No.7004306

nigger

>> No.7004431

>>7004193
ill give you the goods once /ic/ settles down, sorry

>> No.7004523
File: 710 KB, 2241x2140, 1695950913893274.jpg [View same] [iqdb] [saucenao] [google]
7004523

some more slop for you guys! all from new sketches save for the sword one.

>>7004193
unironically check the OP in >>>/g/sdg/
feel free to join here once you have it all running tho.

it is the krita AI plugin. although thre are other tools like it. don't worry about the live painting for now. it's fun but it's not that important.

>> No.7004537

>>7004523
Thanks, though i have sd installed already, i'm just interested in model/lora recs and if possible like a short description of the workflow.

>> No.7004561

>>7004537
model is exquisitedetails, i also use a bunch of others but that's probably the most relevant one for here. if i'm not mistaken no loras were used in anything itt. but that's more out of lazyness than anything else.

also this >>7004523 was not made using krita live painting but comfyui only. most of everything else were krita live experiments.
live painting is worse but it's more interactive and you get more direct feedback. also you get layer integration and inpainting similar to adobe firefly.

>> No.7004570

>>7004561
thanks ill check it out tomorrow