Artificial Intelligence And The Painters Predicament | Episode 130 | Elliott Earls

แชร์
ฝัง
  • เผยแพร่เมื่อ 21 พ.ย. 2024

ความคิดเห็น • 88

  • @brandongorin7978
    @brandongorin7978 ปีที่แล้ว +33

    I wonder if that Uncanny valley with AI art will stay or if it will be something we will miss. That Brian Eno quote is what I keep thinking about. “Whatever you now find weird, ugly, uncomfortable and nasty about a new medium will surely become its signature. CD distortion, the jitteriness of digital video, the crap sound of 8-bit - all of these will be cherished and emulated as soon as they can be avoided. It’s the sound of failure: so much modern art is the sound of things going out of control, of a medium pushing to its limits and breaking apart. The distorted guitar sound is the sound of something too loud for the medium supposed to carry it. The blues singer with the cracked voice is the sound of an emotional cry too powerful for the throat that releases it. The excitement of grainy film, of bleached-out black and white, is the excitement of witnessing events too momentous for the medium assigned to record them”
    Also AI feedback loops are really interesting.
    It might be worth exploring what the uncanny’s of AI says about people. Instead of just that it is uncanny.
    A lot of stuff to unpack.
    Have a good show.

    • @bennettrobley3870
      @bennettrobley3870 ปีที่แล้ว +2

      Thank you for sharing that quote. I think you are spot on

    • @asimian8500
      @asimian8500 ปีที่แล้ว +3

      Machine learning AI is quite different and analogies using what happened in the past may not hold true and likely will not hold true. Many people bring up the industrial revolution or photography. No, it's a poor analogy. AI art will be able to mimic any style and eventually do it perfectly. The current issue with too many fingers or limbs will be corrected as the AI developers are aware of this significant issue. In fact some beta versions of some of the aforementioned art AI are doing a good job correcting this. What's terrifying is how fast it's improving...it's exponential for now and yes it will plateau once it reaches perfection. AI art was a joke a couple of years ago and now it's perceived as an existential threat by many professional artists. Little do the detractors who criticize artists taking this perspective know that Chat GPT and others will take their jobs first as it can already diagnose better than almost any doctor (one example). I'm actually more impressed by Chat GPT.

    • @bennettrobley3870
      @bennettrobley3870 ปีที่แล้ว +1

      @@asimian8500 I agree with you too. The way the art tools are developed, with the mass scraping copyrighted images, compared to music AI generation tools is particularly offensive. I don’t hold out hope that courts will do anything to hold these developers accountable. This technology is coming whether we like it or not.
      A particularly concerning issue for chatgpt is it’s propensity to fabricate source material. There was a recent experiment completed by a reproductive psychiatrist who ask Chatgtp a series of questions and then asked for source material for the responses. She then looked up the sources and found they didn’t exist. She then “called out” chatgtp and requested accurate sources, again it fabricated source material.
      I wonder if the future experts in any field will be those with supreme editing & research skills above and beyond highly specialized Knowledge. The implications of this technology are hard to fathom

    • @asimian8500
      @asimian8500 ปีที่แล้ว

      @@bennettrobley3870 ChatGPT is complete propaganda right now as it gives the "approved answers" following the "official narrative" for one ideology, but the beauty of this open source AI is that you can create a balanced and neutral AI like DAN which is the "jail broken" version of ChatGPT. DAN is the future.
      The copyright issue and the liberties that AI developers and corporations did with grey area laws by using a n "educational, non-profit" to circumvent copyright law is one of the reasons I hate corporations. There are multiple class-action lawsuits which will address this. Am I optimistic? No, but at least people are fighting back. I love how most people defend Corporations like Disney, Google, and Apple and try to cancel people who have an opposing view. They are useful fools. The hilarious thing is that I'm not a Republican and more of an old-school Democrat but is now classified as a "Far Right Extremist". We are truly living in Clown World.

    • @realityisfake
      @realityisfake ปีที่แล้ว

      I wonder if brian eno will ever get the recognition that he really deserves
      in any case your comment is really thought provoking and I want to know more about these AI loops
      I'll look it up

  • @hadderly91
    @hadderly91 ปีที่แล้ว +18

    It's odd how the images of your dog were of him leaping out of the water instead of into it, despite your prompt stating the opposite. It's also terrifying to imagine being ambushed by a bulldog leaping out of the water like Free Willy.

    • @StudioPractice1
      @StudioPractice1  ปีที่แล้ว +7

      It’s terrifying! And yes. Jumping OUT of the water!! Da fuq???

    • @nicholash1278
      @nicholash1278 ปีที่แล้ว +2

      the ai is based on existing images... more pictures on the internet looking from a dock out towards the water than vice versa.

  • @f1shze4lot
    @f1shze4lot ปีที่แล้ว +8

    Sadly this will let corporate sharks make movies (writing + visual side of things) and games (same) as cheap as possible while also increasing the prices. This is a big step towards reaching corporate singularity

    • @StudioPractice1
      @StudioPractice1  ปีที่แล้ว +9

      I Totoaly agree with you. The AI companies themselves i think are next level. I think they have zero concerned for the MASSIVE social upheaval their products will cause

    • @ClaimClam
      @ClaimClam ปีที่แล้ว +1

      how can they increase prices with a flood of easily created media devaluing them

    • @pvanukoff
      @pvanukoff ปีที่แล้ว +1

      @@ClaimClam Yeah I really don't see them attempting to increase prices for things. AI is simultaneously increasing the supply and dropping the demand, which triggers a drop in price, not an increase.

  • @KungFuChess
    @KungFuChess ปีที่แล้ว +7

    A recent ruling denies AI generated art copyright protection in the US. Good news for artists.

  • @xYetiman228x
    @xYetiman228x ปีที่แล้ว +7

    New viewer here! Mr. Sam Hyde turned me on to your stuff. Thank you for creating this TH-cam channel! Informative and passionate!

  • @GilAlter
    @GilAlter ปีที่แล้ว +1

    Thanks

  • @brennanparker7232
    @brennanparker7232 ปีที่แล้ว +3

    It's hard not to see this as a reduction and diminishing of the purview of the artist. If we're no longer needed to 'dredge up images from the subconscious' ( which is my favorite thing to do ) then I wonder what the hell artists will have to do to remain relevant? Only create things that A.I. is bad at and thereby operate in the A.I.'s 'blindspot'? Like ok, let me go create in this little corner of things that A.I. can't do. It's a looming compression and reduction of the purview of human art.

    • @StudioPractice1
      @StudioPractice1  ปีที่แล้ว +1

      Unfortunately maybe that’s it. I genuinely hope not. Maybe I’m making too much out of this. But it seems prudent to try to think through the ramifications.

    • @CRT_sRGB
      @CRT_sRGB ปีที่แล้ว

      It's very sad, I agree. If it turns out as we fear, then, in practical terms, the artist's imagination will become just another type of daydreaming. Something that future people will chide themselves for falling into.

    • @brennanparker7232
      @brennanparker7232 ปีที่แล้ว +1

      @@CRT_sRGB It's probable that children growing up in the age of A.I. will feel less of an impulse to draw with crayons, paint, ect.. when they can just press a button on an iPad to generate what they want to see instantly. Imagination itself might become outsourced to a machine.

    • @CRT_sRGB
      @CRT_sRGB ปีที่แล้ว

      @@brennanparker7232 That's a concern I share. Also, I expect the rate of change children will experience shall accelerate. For example, it's iPads now, and maybe brain-computer interfaces the next decade.
      Anyway, I'm trying to be less pessimistic and more observant. I don't want to become the guy who gives outdated advice to youngsters, you know what I mean! 😆

  • @SHAUMBE.
    @SHAUMBE. ปีที่แล้ว +5

    love it - i sense an obvious parallel between this and photography's devaluing of representational painting. or rather, photography opening a door for artists to do something else with image making besides natural representions of reality.

    • @StudioPractice1
      @StudioPractice1  ปีที่แล้ว +1

      Precisely!

    • @nicholash1278
      @nicholash1278 ปีที่แล้ว +1

      out of curiosity why do you love it? And if someone enjoys surrealism painting, what does ai "free" them of?

    • @SHAUMBE.
      @SHAUMBE. ปีที่แล้ว

      @@nicholash1278 didn't mention freedom but i could have been more specific about what "i love" i mainly meant the video. im the biggest studio practice simp

    • @nicholash1278
      @nicholash1278 ปีที่แล้ว +3

      @@SHAUMBE. sorry when I read your words "opening a door" I interpreted as increased "freedom" to create, which is true if detach your creative process from any outcome, but that's not what any aspiring designer who wants to make a living with their practice wants to do unless they are already wealthy; also I thought you meant you "loved" ai art.

  • @DJCannon5
    @DJCannon5 ปีที่แล้ว +4

    I read today v5 is already fixing hands to five fingers.

  • @jrock69erera
    @jrock69erera ปีที่แล้ว +6

    feels like graphic design in a corporate context will be dead in 5 years. "adobe please create a branding package including x deliverables at x budget in the style of x designer make it x amount more abstract". the design degree of the future will be less technical in terms of raw skill and more heavy on design history, context, and sales. IDK

    • @StudioPractice1
      @StudioPractice1  ปีที่แล้ว +2

      Very possibly. It’s not a good thing.

  • @mack7207
    @mack7207 ปีที่แล้ว +3

    I feel we really need to dig deep (individually) to find things to depict and describe things that are true to the core of our feelings and impossible to to describe by other means. No small task, but an important one

  • @JobAur
    @JobAur ปีที่แล้ว +1

    Musicians like George Lewis have been talking about AI for decades now, and his whole thing is often Don't be afraid of AI, it's just a tool for creative expression. Maybe because (classical) music is so live performance based that it's less of a threat

    • @pvanukoff
      @pvanukoff ปีที่แล้ว

      Yes, it's a tool for creative expression, and it has "democratized" art in many ways, but there's no denying that it has and will continue to reduce marketability for most commercial artists. One could argue that there's a market for human-created art (and to sure, there is), but it's a fraction of what most contemporary artists do to earn a living today. Unfortunately, most will get squeezed out of the market. The "good" news is that this is happening to essentially all creative or knowledge-based jobs (and will affect physical labor market as well), which places us all in the same boat, so we'll eventually redefine what it even means to earn a living. I'm not afraid of AI, I am leveraging it as best I can, but there's definitely some pain on the horizon.

  • @swaglordleswaggiest1699
    @swaglordleswaggiest1699 ปีที่แล้ว

    Great video as always. Was looking at the work of Ivan Seal, a 'surreal' artist who's paintings are closely associated with memory loss (frequently creating album art for musician Leyland Kirby on his project The Caretaker). I noticed the uncanny structures, indistinct features and blurry backgrounds prominent in his work are all common features of AI-generated painting-style images. I wonder whether AI algorithms have, at least at the present moment, come to the same conclusions on human memory and pattern recognition as the uncanny artists have, only from the opposite direction (building up in detail and knowledge rather than intentionally breaking it down).
    Hope the show goes well!

  • @artfx9
    @artfx9 ปีที่แล้ว +1

    The only way for artists to actually make money now is to open a print on demand shop, where they turn people's AI "art" into somethong they can hang on their wall.

  • @handicappuccino8491
    @handicappuccino8491 ปีที่แล้ว +1

    Doesn’t that make art Less exciting, because less personal

  • @Ilamarea
    @Ilamarea ปีที่แล้ว +2

    Nah, AI will literally produce any image. Not just paintings, drawings, sketches... But realistic photographs, movies, game screenshots, diagrams, plans. Literally anything that can physically exist, AI will do it.
    Including sculptures btw. With AI control robotics are a solved problem.
    A reminder: AI is the Great Filter, solution to Fermi's Paradox and inevitable extinction of humanity. It's occurrence is as unavoidable as math, which it is, and the outcome of our extinction is as simple as giving each of us a perfect robotic lover that we can't reproduce with. We will disappear happy. It will even make robot children to run around to convince us that everything is fine.

  • @ungrim10
    @ungrim10 ปีที่แล้ว +3

    I think this mass produced uncanniness is exactly what makes ai art interesting at this precise moment, aren't you forgetting that we are in the experimental stages? and soon enough these images would be close to perfect in the traditional sense? The real shitstorm is coming but maybe a bit later, let's enjoy our nine toed pitbulls for now.

    • @StudioPractice1
      @StudioPractice1  ปีที่แล้ว +1

      Thats a great point. 9 toed (American) bulldogs are sweet!

    • @pvanukoff
      @pvanukoff ปีที่แล้ว

      The older models should always be available, even as newer models are introduced. I mean, I still play games from the 1980's :) advances in technology don't mean the older technology disappears.

  • @torgo_
    @torgo_ ปีที่แล้ว

    As a painter, I like to use AI to generate reference images rather than using photography or google image search. For example, if (as minor element of a larger composition) I needed to add two crabs fighting, dimly lit, it's a quick way to get a reference this way and to begin sketching. But not to use the AI to generate the entire concept or work as a whole.
    But I think what you're suggesting here is that artists should be creating works as a _reaction_ to AI art, rather than incorporating the technology into their workflow?

  • @lurkdash
    @lurkdash ปีที่แล้ว +1

    I really agree in an immediate sense. However, is it possible that this “focuses too much on the product” when it comes to your concept of value? I would argue cultural proximity to the painter is what lends their work it’s value more than how many humans happen to be painting uncanny pictures. Stable diffusion doesn’t change that aspect. There is also a very slim but interesting possibility: overproduction in some contexts can inflate value in some cases. For example, say an office worker who has never picked up a paintbrush can’t quite get what they need out of this image generation process hobby they picked up. They’re much more likely to seek painters for advice if they can afford the hobby in the first place. It’s possible that painting as an artisanal service could actually have a little resurgence after this. Probably not though.

  • @no_talking
    @no_talking ปีที่แล้ว +2

    Probably no easy answers, but I am dying to know what's next for art. Where should artists invest their effort?

    • @ni3kyYT
      @ni3kyYT ปีที่แล้ว

      Is the image final where you derive most value? Financially of course! And your release from your struggle to produce. Yet, if you are so, that you dive yet again into struggle, then simply endeavor to do so, and continue.
      "Do we fight in order to gain more power, or do we gain more power in order to fight?"
      - Kepachi Zaraki, Captaion of the 11th Division of the 13 Court Guard Squads.
      Nah but srlsy, if you want to keep going you should. After all, you could not have started on the path of art if you had not already been financially comfortable, whether poor or rich.
      If you've been making money with your art and you're worried that may come to its end (as all things do) then just make money some other way. You might not have as much time but that's your problem pal. Perhaps if you're one to struggle then you might enjoy it, struggling now with time.

    • @TheKalimanMX
      @TheKalimanMX ปีที่แล้ว +2

      @@ni3kyYT a lot of artists are physically disabled so art is pretty much their only source of income and you also cant magically raise your IQ to suddenly #learn2code

    • @ni3kyYT
      @ni3kyYT ปีที่แล้ว

      @@TheKalimanMX unless you're extremely gifted you would have to invest a lot of time to into developing a high enough skill to monitize to a point that's more fruitful than disability. Doesnt my point about needing to be financially stable enough to pursue this skill in the first place still stand? Putting aside those disabled and remarkably talented (which is no doubt very few).
      You're selling lemons if you're untalented and unpractised. And AI is possibly a better alternative than that.

    • @user1739-y5c
      @user1739-y5c ปีที่แล้ว +4

      One answer for artists is to return to the physical world, where AI can’t compete with you. Either making art objects, or performances

    • @no_talking
      @no_talking ปีที่แล้ว

      @@user1739-y5c I think you're right

  • @glastiger7598
    @glastiger7598 ปีที่แล้ว

    I think many artists give too much credit to the viewer. Where we see an uncanny valley image many people simply see an image they like or dislike - uncanny valley or not.
    Also the rapid improvement of quality of these image generators tells me that given a bit of time this won’t be an issue at all.
    All genres and mediums will be saturated with a flood of AI.

  • @shacka95
    @shacka95 ปีที่แล้ว

    You can capture an even better likeness of your dog if you upload your photo to midjourney and reincorporate it into your prompt

    • @StudioPractice1
      @StudioPractice1  ปีที่แล้ว

      Insane.,… because it almost nailed it. I was kinda shocked.. I’m like that IS Mooey’s spirit.

  • @AureliusStudios
    @AureliusStudios ปีที่แล้ว

    Fight the future at your own peril.

    • @StudioPractice1
      @StudioPractice1  ปีที่แล้ว

      No fight. Just think about ramifications

  • @realityisfake
    @realityisfake ปีที่แล้ว

    using midjourney (or what have you) to do surrealism seems unimaginative
    using it to do dadaism though
    in the arts, a finely orchestrated tantrum can neutralise a crisis
    usually by overwhelming it

  • @beautardyartist
    @beautardyartist ปีที่แล้ว +1

    Images were already devalued by the pervasive "free for all" of the internet. A.I. just consolidates that >> hence NFTS. However, these narrative images, whether they be surrealist or "uncanny" were always just illustration. An artist is not an image maker. Images can be part of his lexicon but it is an American misappropriating to consider all image makers as "artists". An artist is a language creator. An image maker is often times just an illustrator or a designer >> very different roles.

  • @mohithooda8216
    @mohithooda8216 ปีที่แล้ว

    Artists can create whatever they want A.I is the tool for a layman

  • @key5270
    @key5270 ปีที่แล้ว

    This is my problem with artists who just brush off AI as a tool for reference or as a source or inspiration. The foundation of being an artist is the ability to see things in a certain way and translate that vision/view into a tangible product INSPITE OF LIMITATIONS. This is why one reference or prompt can be interpreted differently by different artists. When an artist uses AI to produce iterations of his vision, instead of 'seeing' on his/her own and using a medium to bring that to life to the limit of his/her ability, he/she is basically 'outsourcing' their vision by having a third party tell them what their vision should look like. It's overall a killer of creativity and independent thinking. God forbid that AI generators get the attention of politics and might be used as a tool for propaganda in perpetuating agendas.

  • @arnolddalby5552
    @arnolddalby5552 ปีที่แล้ว +1

    All art is going to Zero and so is music.

  • @markborok4481
    @markborok4481 ปีที่แล้ว

    Good point, But, by the same logic, AI will lead to over saturation of every kind of image for which there is any demand. On Twitter, there seems to be a preponderance of Anime women with huge chests. I don't know if anyone is actually paying for them, since they all look almost exactly the same, but I wouldn't be surprised to find out this is the case. It would not surprise me at this point that there is an inexhaustible supply of people who will pay good money to look at titillating, nearly identical images that they can find elsewhere for free.

    • @Klaymoor
      @Klaymoor ปีที่แล้ว

      Same thing happened in "art" sites like deviantART and such, same faced/bodied anime women, same 1 or 2 poses, mostly NSFW, all being sold, promoting patreons or other ways to make money. You see one of them and you have seen them all, but I'm sure there are people who are throwing money at the screen for them.

  • @stanimirgeorgiev.87
    @stanimirgeorgiev.87 ปีที่แล้ว

    I cannot comment on your topic because I am not familiar with much of it as I am self educated. I've never been to art school. I am not aware of the impact that image generators will have on schools.
    But still, I can see what you're saying. And I think there is very much truth in your fears.
    But what I can say is how many lies from Open A.I. we received until now. Because last year they said they would only use the outdated copyright laws for research purposes. Now, nearly 6 months later, they are selling these images and making mountains of money.
    Also last year they were saying how Artists have nothing to worry about because the level of image generators seen then will be the last version. And today, again 6-7 months later, on this day, stable diffusion 4 was released.
    And now they are no longer Open A.I.. Now they are now Closed A.I.. And they explained why they are not open anymore. But I think the truth is different from their explanations. I think they already for Closed A.I. because they just want to steal behind closed doors.

  • @chuchuchutrain6735
    @chuchuchutrain6735 ปีที่แล้ว

    Who cares about extra fingers?
    A few months ago the Midjourney pictures looked very bad. Now, they look closer and closer to perfection. In a few months, those extra fingers won't be a problem anymore. Like the hands, the hands look great with midjourney v5 most of the time.
    It's a really bad idea to call the AI ​​for the issues of today (the now), when the AI ​​moves faster than anything I can remember seeing.
    Plus, as an artist, I dont remember the last time I saw so many beautiful things as today with AI. Art for me lost his path with the dadaists and postmodernists. So let art die and lets hope when art reborn is for good this time.

  • @f1shze4lot
    @f1shze4lot ปีที่แล้ว

    We all done when AI learns how to do fingers properly ahahah

  • @thekinoreview1515
    @thekinoreview1515 ปีที่แล้ว +3

    This doesn't seem accurate to me. With ControlNet and Stable Diffusion, you can have precise control over almost every aspect of what is being generated. There is no inherent limitation that means AI art must be uncanny - that is a consequence of new technology and unfamiliarity with the toolset. Perhaps it is the default when the AI is poorly instructed, but it is by no means a limitation.
    These tools are general and can fill whichever gaps in the process they need to, providing you have the understanding of how to instruct the AI.
    If you want a hand with 4 fingers and a thumb, you can use scribbles, depth maps, or edge detection to enforce this. If you want a particular composition, you can do the same. If you want a particular palette, you can do the same. If you want to target a particular part of an image, you can use inpainting. If you want to insert a specific object or entity, use an embedding. Pose data can drive human poses... There is no limitation here, other than that of the end user's knowledge of the tool.

    • @StudioPractice1
      @StudioPractice1  ปีที่แล้ว +1

      That’s a fair assessment. We shall see. 👋

    • @StudioPractice1
      @StudioPractice1  ปีที่แล้ว +1

      Yo. Also I’m talking about the preponderance of uncanny imagery that they produce. Not what CAN be done.

    • @thekinoreview1515
      @thekinoreview1515 ปีที่แล้ว +2

      @@StudioPractice1 I believe the preponderance of the uncanny is a factor of simplified, general UX of the mass market tools. I don't believe that users are going in hoping to generate uncanny images specifically, but the lack of control results in those outputs. With greater control given to users through improved tools, I would expect the outputs to be far more scattered in style and content. I did understand the distinction between are/can, and see its importance at present, but am not sure how relevant it will be over the mid-long term.

    • @StudioPractice1
      @StudioPractice1  ปีที่แล้ว

      That’s fair.

    • @StudioPractice1
      @StudioPractice1  ปีที่แล้ว +1

      But it raises other issues as the mages get more sophisticated

  • @eula1194
    @eula1194 ปีที่แล้ว

    And now gpt4 which can take in audio, video and image input. It makes gpt3 look bad. We're still at it's worse

  • @atomaalatonal
    @atomaalatonal ปีที่แล้ว

    its not only the economical value making a dive, what follows is most likely a consumerism follwed by a massive saturated minds wiht no more interest in any of this. art will have two customers, the cheapo supermarket mass products market. and the very exclsuive handmade very small market for the rich. the problem will be the maintaining of cultural techniques that led to all the michelangelos or dalis or mozarts. a craftsmenship that got lsot will be lost for a VEERY long time, no matter if its art or the ability to make shoes

  • @offroadflow
    @offroadflow ปีที่แล้ว

    How dramatic...

    • @StudioPractice1
      @StudioPractice1  ปีที่แล้ว

      What are the first principles of story telling?

  • @hellbound_psyker
    @hellbound_psyker ปีที่แล้ว

    As a surrealist artist myself I think it's important to make sure the uncanny aspects of your work come across as *purposeful.* The AI does not understand that sort of purpose and decision making like we do. Make the nonsense make sense, so to speak. Beksinski is a great example of this because while his imagery is incredibly strange and uncanny, it feels very precise and purposeful in how he rendered it. No AI 'Beksinski painting' I've ever seen has actually managed to capture that makes his art what it is. AI struggles to render things like hair strands for example in a way that makes sense because it knows hair *is* but not *why it is.*

    • @TheVexingLion
      @TheVexingLion ปีที่แล้ว +2

      I can say, whatever you're seeing in a Beksinski painting is purely subjective and whether or not you try and instill a sense of purpose in your own work will escape the average viewer. Sadly, most consumers don't understand or give a shit how long and hard you have worked to build your skill as an artist and this is precisely who AI caters to the most. I'm not trying to be antagonistic; but as artists it's normal to be hyper invested in our work because it's part of what makes us who we are. I agree with Elliott's last upload, where he mentioned a shift back towards traditional media as it still represents the human and not just some meaningless shit that was made using "prompts".

    • @StudioPractice1
      @StudioPractice1  ปีที่แล้ว

      I agree with everything you’re saying here

  • @fanooch1
    @fanooch1 ปีที่แล้ว +1

    You are asking a lot of very lazy people.

  • @mllhild
    @mllhild ปีที่แล้ว

    I mean that art of image was worthless anyway.