chromatocloo reblogged
vampcass and wolfzo
chromatocloo reblogged
chromatocloo reblogged
Anonymous asked:
I always avoided using references, but couldn't point out what exactly made me avoid it. I have finally realized that due to being unfortunate enough to be surrounded by people boldly claiming "real artists draw everything from imagination without any references" [now i know it's bullshit] in my early years of artistic journey, I developed fear of being too dependent on references and fear of being completely unable to draw anything good without "copying". Any ideas how to overcome that fear?
It's not something I hear at all in professional spaces. We've straight up traced weird angles of our arms before. I use lots of references. You will naturally get better at drawing intuitively with the more references you use since we learn by example. I draw intuitively much more accurately than I did many years ago, but I still use references because my idea of a polished drawing is a moving goalpost.
Jason Rainville is one of the best of the best, paints MTG cards and he often posts his references. You can see how closely they mirror the final image:
If you want to get good at art, you have to look at things and that's all it boils down to. Those who say it's bad to use references are either 1) never getting good at art, or 2) are insecure and are posturing in front of other artists. Hope this helps!
chromatocloo reblogged
click this..
then click your blog(s)
scroll all the way down and click this..
profit?
On mobile it’s under visibility
This is better than the how-to I was about to write.
So yeah. Do this.
And let's see if we can get "fuck ai" trending, shall we?
chromatocloo reblogged
Tumblr is doing some stupid AI shit so go to blog settings > Visibility > Prevent third-party sharing.
also use Nightshade and Glaze to protect your artwork and images!!!!
best combination is usin Nightshade then Glaze on your images. Glaze creates a protective layer and Nightshade poisons the A.I. software.
chromatocloo reblogged
chromatocloo reblogged
Remember all you. THIS is the post that triggered Rita's ban.
This is the #2 trending post sitewide right now, just below an outline of the ways tumblr staff has historically targeted trans women. Keep it up.
chromatocloo reblogged
Happy late Valentine everyone!
I bring you Mchanzo/Yeehan as my first artwork of the years
chromatocloo reblogged
ah yes. han solo. han solo, so suave
so cool under pressure
so calm in a crisis
great at handling difficult situations, for example, can get his own gloves off WHILE talking to a cute girl AT the same time no problem thanks for asking
so great at witty comebacks
definitely has slept with MANY a lady because, again, Han Solo is a cool guy, and not a grumpy hermit who, were he a person in the world, would spend all his weekends alone in his apartment with his phone turned off watching Ice Road Truckers
definitely not a weirdo with a shitty haircut who talks to his car
no. mister cool guy. always looks so cool. so cool in a fight
so cool. never panics about everything all the time constantly.
people trust him cause he’s got that cool guy charisma
always knows what he’s doing. han solo. an expert.
in conclusion: han solo, a cool space scoundrel, not a nerd. maybe you’re the nerd around here. hmm. looks like it. check and mate
chromatocloo reblogged
chromatocloo reblogged
Begging people to stop reblogging this AI trash from “The Phantom Painter” on Instagram (instagram.com/phantom.painting). I’ve been seeing it on my dash more and more often from people who are otherwise anti-AI and either can’t tell it’s AI or don’t care because it looks cool.
This is the kind of shit that is VERY CLEARLY trained on the works of existing talented artists’ with distinct styles and this asshole is selling prints and making a profit off of stealing other people’s hard work.
Don’t give people like this money or attention and they will go away.
Please, if you’re going to buy art prints, buy them from an actual artist.
@thegnat That’s the problem, it’s getting really hard to tell what’s AI and what isn’t. This phantom painter person at least says on each post on Instagram that it was created with AI, but when people re-post it on tumblr, it isn’t specified, and people end up reblogging it.
It’s not reasonable to think everyone should be vetting every single art they reblog to make sure it isn’t AI, which is why I made this post to let people know this artist specifically is AI, and I see it reblogged a ton.
While we’re at it
and so is everything by that person, xis.lanyx on Instagram. They also sell prints of their “original images”
That last one is for the people saying “you can always tell by the hands.” You cannot. The whole point is that it’s getting better and better at it. That’s what it does.
As an art curator on tumblr I now have to spend a considerable amount of time trying to decipher whether or not the art I want to share is AI or not and I still sometimes get it wrong.
To actually answer the question of how one can determine something is AI, particularly as it grows more sophisticated with anatomy: you have to train yourself to recognize artifacts.
There is no one single unifying giveaway beyond a strange sense of uncanny that you will eventually begin to recognize the better you attune yourself, and certain models have their own unique “styles” you can begin to recognize (midjourney and stablediffusion produce very different looks, for example). There are however a few things on which one can tend to focus.
- Edges: AI, as of this post, still struggles with distinct edges of objects and figures and has a tendency to blend details together. Look for hair, ribbons, and other flowy details if present. Do they fade into other details? Look at how the hair fuses with the smoke:
- Edges 2: Sometimes they will also have the edges completely avoid each other, with a foreground figure slightly warping along the edges in a way that matches the background edges, like repelling magnets:
- Patterns: AI, as of this post, still struggles with patterns. Filigrees, mandalas, brickwork, scales, anything that involves a high level of intricate detail tends to get blurred together. This can be a tricky one, because a lot of artists will also fudge pattern details in looser renders, but usually in a way that makes sense as an impression and not…. this:
- Architecture: Are there buildings present in the image? AI has a tendency to make Escher-esque nonsense structures, with pillars in places they don’t belong, arches that go nowhere, bricks that don’t align, and support beams that start on one plane and connect to another. It also struggles with perspective, but, so do many humans so I would not consider it evidence alone. Check out the placement of this pillar, and also the detail on the… window? Candle cage?? Thing?
- Resolution and quality: AI cannot make high-resolution images. It just cannot. While most artists aren’t posting their full resolutions, generative images can’t be enhanced, and the “artist” will not be able to provide proof of work. You should be able to zoom into work by an artist and admire their strokes, relate to their errors, and appreciate their process at every skill level– zooming into generative images somehow makes them even less clear, a mess of pixels that are somehow both blurry and also look like they have been run through a sharpen filter:
- Text and signatures: AI struggles with legible characters in any language, and the result is a simlish-looking approximation of characters at worst, and hilariously misspelled words at best. Since these models are trained off real artists, they will also often have artifacts of a signature that oopsed its way into the image. These signatures are always illegible or, if “legible”, are not actually the names of real accounts.
Things like this can be tough to spot at a glance if you’re not actively keyed into looking for them, but they’re the type of uncanny stuff that once you see it will start gnawing at the back of your mind. You’ll be scrolling your feed and suddenly take -1 psychic damage and you have to scroll back up to see why. Stuff that goes beyond inconsistent lightsources and bad anatomy.
This uncanny feeling is such a good helper! Just the other day, I saw an AI generated pic on Instagram. Looked nice at the 1st microsecond, then something at the back of my brain went 'hang on a sec this is fishy'. Scrolled into the tags; the post was literally tagged as AI made. The person was not trying to hide it at all.
Looking back at this peculiar piece with this helpful list, I can see what gave me this feeling: nonsense folds in clothing, one arm vanishing into undefined ""stroke"" splotches, hair mixing in with other edges, nonsense patterns (in lightening of all things), and an ersatz of signature almost cut out from the frame.
Another thing that helps is checking through the OP's posts. If the art style radically changes or is full-on inconsistent, that is also a good clue that doesn't require eye training:
chromatocloo reblogged
://www.dreamstime.com/ ://www.freepik.com/ ://www.craiyon.com/ ://stock.adobe.com/ ://storybird.ai/ ://www.dinosaur.org/ ://pngtree.com/ ://creator.nightcafe.studio/ ://www.123rf.com/ ://lumenor.ai/ ://neural.love/ ://www.vecteezy.com/ ://openart.ai/ ://www.artpal.com/ ://generativeai.pub/ ://promptbase.com/
Block these sites in your uBlock Origin so you won't see that shit in your searches
💌 send this to the twelve nicest people you know or who seem to have a good heart and if you get five back you must be pretty awesome 💌
chromatocloo reblogged
octobuddy
chromatocloo reblogged
Fuck fuck fuck fuck fuck google with a 10 feet pole.
Seriously, fuck them. They are breaking the internet BADLY.
Everyone needs to get out of Chrome ASAP. Use duck duck go or any other alternative too.
Jokes on them too, we know how to be petty bitches in reaponse to this. Sit through the fucking 5 second delay and continue to use adblock AND Firefox to spite them.
They realise that 5 seconds of waiting is still better than waiting for, what is it now, 30+ seconds of unskippable ads right?
Friendly reminder that you can add this to your ublock block list to get rid of the five second delay on firefox:
www.youtube.com##+js(nano-stb, resolve(1), 5000, 0.001)
chromatocloo reblogged
one of the best academic paper titles
for those who don't speak academia: "according to our MRI machine, dead fish can recognise human emotions. this suggests we probably should look at the results of our MRI machine a bit more carefully"
I hope everyone realises how incredibly important this dead fish study is. This was SO fucking important.
I still don’t understand
So basically, in the psych and social science fields, researchers would (I don't know if they still do this, I've been out of science for awhile) sling around MRIs like microbiolosts sling around metagenomic analyses. MRIs can measure a lot but people would use them to measure 'activity' in the brain which is like... it's basically the machine doing a fuckload of statistics on brain images of your blood vessels while you do or think about stuff. So you throw a dude in the machine and take a scan, then give him a piece of chocolate cake and throw him back in and the pleasure centres light up. Bam! Eating chocolate makes you happy, proven with MRI! Simple!
These tests get used for all kinds of stuff, and they get used by a lot of people who don't actually know what they're doing, how to interpret the data, or whether there's any real link between what they're measuring and what they're claiming. It's why you see shit going around like "men think of women as objects because when they look at a woman, the same part of their brain is active as when they look at a tool!" and "if you play Mozart for your baby for twenty minutes then their imagination improves, we imaged the brain to prove it!" and "we found where God is in the brain! Christians have more brain activity in this region than atheists!"
There are numerous problems with this kind of science, but the most pressing issue is the validity of the scans themselves. As I said, there's a fair bit of stats to turn an MRI image into 'brain activity', and then you do even more stats on that to get your results. Bennett et. al.'s work ran one of these sorts of experiments, with one difference -- they used a dead salmon instead of living human subjects. And they got positive results. The same sort of experiment, the same methodology, the same results that people were bandying about as positive results. According to the methodology in common use, dead salmon can distinguish human facial expressions. Meaning one of two things:
- Dead salmon can recognise human facial expressions. OR
- Everyone else's results are garbage also, none of you have data for any of this junk.
I cannot overstate just how many papers were completely fucking destroyed by this experiment. Entire careers of particularly lazy scientists were built on these sorts of experiments. A decent chunk of modern experimental neuropsychology was resting on it. Which shows that science is like everything else -- the best advances are motivated by spite.
This is a good reminder that in order to prove *something* scientifically speaking, there are processes and protocols to respect. At the very base, you have to actually understand the technics and statistics you work with.
When those steps are skipped, at best the people involved are seen as stupid clowns or lazy asses. At worse, they actually get people killed by affirming wrongful shit that has Big Consequences.