A lot of fantastical claims made in this article with little citation.<p>I definitely don’t think the author is wrong here. It is very likely AI animation tools will have a huge impact on the industry.<p>I have friends who work in animation and illustration professionally. Based off their predictions, the truth is likely far more mundane. These AI models will likely just become tools that automate or simplify existing animation workflows for larger projects. This would definitely result in less labor needed for any given project.<p>For smaller projects. You might see folks use AI for the entire thing<p>> “the AI art movement is already causing a sizeable reduction in artist commissions, and that effect will soon trickle down into B2B and wider world.”<p>This line specifically made me lose confidence in the author’s ability to predict the future. Cursory Google searches return nothing. This is also a very difficult thing to measure at a population level. In my friend group of illustrators (which is a very small sample of the population) all of them have used TikTok and seen their commissions drastically increase in the last few months. That doesn’t mean ALL artists are making more money though. Neither does a few mediocre artists losing commissions due to AI either.
I’m a little skeptical - at least of the view that we’re a few years away from all animation jobs being dead and ML models producing entire movies or TV episodes e2e from text prompts.<p>What’s likelier is that ML-based tooling becomes a key part of the animation workflow. Used to generate assets, animations, backdrops, characters, etc, which are then combined by an animator/editor. The examples he cites in the article all fit this mold - the anime character generator he cites in the article uses separate models to generate the character then rig it from facial data.<p>After working in the self-driving car industry, I’m really skeptical of any claim that rapid advances in one modality or task mean we’re “just 3 years away” from all related tasks being done via ML models. Alexnet et al completely revolutionized perception in self driving cars - between 2014-2017 it was really common to hear predictions that we’d have end to end models driving our cars perfectly in “less than 5 years”. That reality never arrived because ML just wasn’t capable of handling more complex tasks the way it could with object detection. Lots of articles similar to this one talking about what we were going to do with all of the out of work truckers and Uber drivers.<p>And yeah obviously generative art is a different beast than autonomous driving. And I’ve seen examples from WIP text2video models. But I just want to caution that tons of progress in image and text doesn’t necessarily mean we’re just a year or two away from all related tasks being conquered.
I love these “AI *will* replace [x] job in [y] years” genre of posts because it’s a fun thought experiment. Taken to its extreme in writing, visual art, coding, conversation bots etc.<p>It’s like the future will be everybody living in NvidiaBlackrockDominoes housing, eating TysonIntelJ&J gruel, and the only employer left will be TSMCOnlyFans — which will be a staffing agency for test subjects in GarnierRevlon’s automated anti-aging research facilities.
Television will kill movies because people don’t have to leave their homes. The entertainment industry is doomed because people now have social media. Recruiters won’t have jobs any more because we have LinkedIn. Restaurants are obsolete because anyone can learn to cook with YouTube.<p>I work in an animation adjacent field and this is hogwash. What happens when the person responsible for the creative content says “I don’t like the way this part of the sky looks in the background”? Or “The way this character falls down the stairs is too violent and not funny enough”? Who’s going to sit there with the AI and make it do the thing that someone has asked for? It’s not going to be the director or producer because they have too many other things to do. They will delegate that task to someone. Is that person an animator? Does it matter that they’re working with an AI instead of moving keyframes around?
I find this kind of writing so unbelievably annoying, it's unreadable.<p>People who write as though not only can they predict the future, but they're so arrogantly certain of it that they write long, scary sounding articles about it.<p>Guess what? This random guy has no idea what will happen in less than 5 years, no one has any idea. I'd bet AI won't, in fact, dominate the animation industry. Am I certain of it? No, because at least I have some humility.<p>Anyone who thinks they can predict the future is full of shit.
This is undoubtedly spot-on. I saw this happen with print design. Many job categories like typesetting, pasteup, layout, archiving, plate production, and other prepress and press work simply WENT AWAY as pages were designed on-screen and spit out of printers.<p>At first, there was great resistance and many cries of "It's not as good!," which were, of course, true. But it didn't matter. Desktop publishing as it was called (now something your child does in elementary school) was good enough and opened the door to many uses of "the press" which were simply unaffordable before.<p>The same is coming for media production.
The big question is what happens when the human training set for art stops being generated? Will AI art stagnate, or will there still be enough humans making art without financial reward to continue to supply models with better data?
This falls into the same category as the other efforts to replace human crafts with AI: Too much uncanny valley.<p>Generated text looks like the person crafting it is having a constant stream of micro-strokes while writing in their third language.<p>Deep Fakes have a proportion problem. Not to mention little things like skin tone and lighting.<p>Images look like one of those “one person draws one part, and hands it off blind to someone else” activities.<p>And animation looks like the characters on screen are also having micro-strokes.<p>“It will get better,” but will it? Or will we just continue to live in the valley with a ceiling of “it’s good enough”? Getting better will be a huge cost, after all.
“DALL-E 2 and other AI art models can now produce a near-infinite variety of illustrations using a simple text prompt. By 2025, they'll outperform human artists on every metric”<p>jesus this is so stupid on its face its laughable. this is like an artist saying by 2025 that artists will have replaced coders. you have no idea what you’re talking about and how stupid you sound
I guess what freaks me out is that if this is true, how can't it be true for people that write code? And how much longer do I have, than my artist sibling?
I can see this being true, but it won't be because it's good, it will just be because companies lower their production standards in the interest of saving money. There were (maybe still) lots of really crappy CG animated shows that came out as soon as the tech allowed (20 years ago). The bits I saw were much worse than anything drawn, but they were for kids shows that nobody really cared about the quality of.
> Just look at the recent V-tuber craze as an example. AI generated animation allows people to multiply the utility of a single drawn frame, increasing accessibility and lowering the barrier to entry for animators considerably.<p>It's not accurate to call it "AI generated animation". Face tracking is an input into the software that animates it. Those inputs feed into a bunch of manually configured math to move the thing around.<p>It's also not a new idea. The idea of having a human puppeteer a computer animation goes way back. What's new is streaming.
My 70 year old mother’s new favorite past time is making ai art. She’s using nightcafe and has no knowledge of the internal model workings. I strongly suspect ai startups are about to get their moment.
A few years ago when ML could only dream creepy dog faces, I would not believe that it could generate coherent images with decent framing and composition, correct proportions, realistic light and shadows, depth of field, excellent taste for palettes, and even good faces without uncanny valley — and hands and text being the hard parts.<p>So this time I'm willing to believe that the small morphing clips and blurry NeRFs are going to grow into whole realistic 3d scenes.
That's what they said github copilot was going to do for coding, but I haven't heard of anyone using it to replace developers.<p>Prediction: It's going to run into all sorts of copyright issues, wouldn't be perfect and make mistakes that a human would never do and end up being just another tool for Animators to use to create rough drafts that they can refine further.
I wish every time someone writes a "AI will kill X in 5 years" they'd take me up on a 500$ bet that it won't. If they did I would have been able to retire by now. Anyone remember the "I wouldn't train radiologists any more" arguments from like, ten years ago?<p>computer tools already dominate the animation industry and have for a long time. Lots of that already is AI if we hadn't started to redefine what the word means. It's not going to cause unemployment, it's going to give animators more tools, and it will make animation more accessible.
I do wonder if all these AI generating content software will be the end to artists, animators, media creators etc. I guess for a lot of things it will, but I wonder if people will want an AI or organic created content. Akin to how hand-made crafts are still desirable as opposed to mass produced textiles etc. eventhough the mass produced stuff is cheaper.
I guess there is room for both. But likely much less organic v AI created. Maybe we will see labels placed on content with " natural made " or something a long those lines as a selling point in the future.
When someone is this confident, and they get anything about their prediction substantively wrong, you should question anything else they say with confidence in the future. So, on the off chance this gentleman's prediction is wrong, we should question his fundamental methodology and distrust by default any predictions he makes after this. It should be easy to test, because he provided a year, and outlined how the process would unfold. We could check back in late 2027, though I reckon we'd have a pretty good idea long before then.
I remember when AI was going to replace radiologists.<p>There's a big gap between "Wow, this is really good for being made by a computer" and "Wow, this is consistently better than human experts".
Sounds plausible. However, usually, software doesn't replace all jobs, but makes the top 20% of works productive enough that the bottom 80% aren't needed.<p>At the same time, large businesses are very slow, so many companies will likely continue with traditional methods for decades while newer startups use the latest technology. It'll take decades before all business use AI pipelines. Animators should start their own studios to take advantage of the technology wave.
The key to the success of AI models in animation production is going to be how "directable" they are. Even with traditional procedural techniques, this is an issue.<p>It's relatively straightforward to generate a procedural forest. Less so when the director wants "this branch" to curve more in a given direction.<p>Where I can see this having the greatest initial impact is in concept art and matte painting. And not in a good way for the artists.
When someone figures out AI that can take reference frame then a bunch of sketches and generate intermediate frames now <i>that</i> will be revolution.
<a href="https://qz.com/962427/what-its-like-to-be-a-modern-engraver-the-most-automated-job-in-the-united-states" rel="nofollow">https://qz.com/962427/what-its-like-to-be-a-modern-engraver-...</a><p>When the engraving industry was automated demand went up.
"Dominate" seems a little off. Feels to me like said AI tools are merely an augmentation, and will have a push on the spectrum from hand-animator to curator. In the end I think it will probably come down to taste, which comes down to whatever floats the boat.
I always compare these claims to self checkout. We’ve had technology to replace cashiers for 30 years now and they still just augment checkout cashiers a bit.<p>People are creatures of habit and really don’t want to change our routines. Plus large businesses are extremely risk averse.
I think there's a bit more inertia with animation here than the author realizes. Only the one-shot techniques currently allow you to animate a custom character and only within the confines of human faces and other things for which we have a lot of data.<p>For the vast majority of animation that isn't just humanoids talking, or bodies moving, rigged cg animation is going to be better.<p>In fact, try comparing the control you get from rigged CG animation versus blindly groping through latent space for "just the right character with just the right animation"<p>I think in 5 years AI will dominate the meme generation "industry", but animation will still be dominated by good ol fashion rigged CGI.<p>The rigging and animation tools are what will be an order of magnitude better + faster, particularly facial and humanoid body stuff.
The idea animations will get easier to make means more human time can be spent on the creative composition side and less on the production side. It's still ultimately a big win for art and consumers.
Aside: if you scroll to past issues, the author has all sorts of other articles like how to get into p0rn. Wonder if the author picked up AI recently, in which case, that's impressive.
Ah yes, in less than 5 years AI will completely dominate the fields of animation, art, programming, moviemaking, car driving, healthcare, etc, the sensationalist treadmill never stops.
It wont. Same as 3D won't kill 2D animation. Digital drawing wont kill classical drawing, etc.<p>Animation isn't just about moving static images.
"And there's really nothing you can do about it."<p>Typical neo-liberal ideology: "you can't do anything about capitalism".<p>Yes, you can. Humans controls AI, not the other way around (at least for now). Join your local communist party, organize to combat capitalism, discuss regulations to AI and corporations, worker ownership over the means of production, etc.<p>Another capitalist mouthpiece that thinks that AI, capitalism or human misery are natural events, like the rain or wind. There is nothing "inevitable" about this at all, capitalism doesn't grow in trees.
I’ve said this before, but this article is the view from someone who has no experience in the field who sees technology as the answer to everything.<p>I am someone who has both worked in the animation industry (both as an artist and a developer), and currently work on what I would consider the bleeding edge of graphics engineering with ML (with large teams of course), I feel very confident in saying that people who make articles like this are just spewing technocratic bullshit.<p>There’s no understanding of the craft that goes into making content. There’s no understanding of the iterations and creative control that goes into it. There’s also no understanding of the business of animation.<p>Saying “you’ll lose your jobs” is downright stupid. Yes, AI tools will help streamline jobs. But that’s what they are: tools. Some jobs may be displaced if people don’t learn the new tools (see 2D animation being displaced by 3D) or if you were already just a worker doing menial work. However there’s so much artistry involved in every phase of production that you’re not going to see job reductions for a few decades.<p>I believe the opposite, there’s such a desire for a glut of content, that you’ll see more jobs. AI may help accelerate those tasks, but there’s already so much work that studios are constantly at capacity.<p>The other fact is that it ignores human nature. We still value stop motion animation or 2D even if CGI is better for mass production. There’s a human element that is not purely rational that we crave.<p>Lastly, the reason I think articles like this are bunk is because the people who write them assume studios do little research and development internally. We do so much. Attend any SIGGRAPH session and you’ll see so much mind blowing content.<p>We spend years trying to make it so artists don’t have to do work. I’ve literally automated my job and the job of other artists on other projects.<p>Does that mean my department or those artists get fired? No! It means they spend time working on making the content better instead of spending time on the stuff that was just automated. When they were spending time working on menial things like crowds, they can now spend that time finessing great performances.<p>We often stay on top of the cutting edge of research outside of our industry and make good use of it. It’s not like the industry, made up of countless brilliant engineers and artists, are blind to what’s going on.<p>Anyway, my advice to people: don’t listen to charlatans who sell technology as the answer to problem spaces they’ve never worked in. Yes, every now and then you get an outsider who shakes up the world, but it’s rare without some deeper understanding of the problem space.<p>The people saying Diffusion art generators and nerfs are coming for the animation jobs are quite clueless in my opinion, and are easily hyped by “shiny”.