Sora is here: why you should not panic (yet)
ฝัง
- เผยแพร่เมื่อ 5 ส.ค. 2024
- Since Sora was announce I receive a lot of DM from artist, afraid that this tech could be apply to animation as well in the near future and put us all out of job, but.....
- ภาพยนตร์และแอนิเมชัน
It’s okay if we all find ourselves unemployed, get together and form an underground club of anonymous animators)))
Animators of the world unite
First rule of Animation Club : You don't talk about Animation club
Second rule of Animation Club: You don't talk about Animation club
Ahah love you guys😂
RIP stock footage companies
Shutterstock signed a deal to share their library with OpenAI, so they're also profiting from this. Sora wouldn't exist without Shutterstock
has shutterstock got permsission from its sellers to use their contents for AI or they'll be paying them too a cut? @@takahashierik
@@jkartz92 they rather hired a bunch of expensive lawyers to cover their @zzes
More like Stock Photographers
@@IamSH1VA The boots on the ground creators will be the ones with the most liability here.
I am not worried about ai, I am worried about Greedy Companies & CEO 😅
Nice video…
my fear is not about ai taking people's jobs, but mostly about flooding the market with garbage ai products to the point that submerges good quality works
That's already todays world, it's flooded with content, yet we consumers use rating systems to find the good work, and the producers pay money for commercials and ads to get our attention, which you shouldn't do if your product is shit
Really good and inspiring final message Alessandro! Really liked your opinion on this matter.
I really appreciate Sir How you came and give motivation like mentors does. Hats Off to you Sir.
Much Love and Respect from Pakistan
Grazie Ale mi dai speranza così
The problem with all the people saying "don't panic" is that the logic, much like this video, is that humans have "soul" and are able to make something more interesting than robot. But you forget two thigns: 1) these AI-techs makes what humans ask for, so animators can still be replaced by a human who tells the A.I. the details that will give it life e.g. to "open the door, take a brief pause first", and 2) they forget that any feature you say is lacking from the AI, migth be added in a patch in a couple months or years from now. Listen closely; it's unlikely that A.I. will stop improving, and ultimately be able to replace many animators, because it's speeding up the work process. HOWEVER, I have two positive thoughts to share; 1) the world will probably start producing MORE content, so we might not see a lot of people losing jobs, we might just make "MORE" instead of firing everybody. 2) you animators are the most fitting people to use this tech, so you can become self-employed, use the AI, and rather than losing your jobs... you start competing with the companies that would fire you, and since you're alone, you can afford to sell smaller volume of videos.
That's very reasonable, but still I can give you that there will still be demand for human-crafted products, what Alessandro said about patterns that will start to show up is true, there are visual concepts that doesn't exist in the real world, things that are created exclusively by humans, such things the AI is very unlikely to come up with own it's own (Cthulu from the lovecraft universe comes to mind for example).
In the end it's like you said, the best people to use this tech are the same people it is competing with. An animator is MUCH more capable of using the tech to it's full potential than an intern at a advertisement agency. Knowledge about the craft will still be relevant when tweaking prompts.
Yeah but we have to talk about how many humans are needed to work. In your example about k'thulu, we still need only one human, and you could say "before you needed a team of animators" for the cinematic k'thulu movie, whereass in the future you might need only the one guy. So sure, AI can't replace every human... but it can replace 9 out of the 10 humans that work in a company.
@@AugustoXRock
The main problem people to this day cannot figure out. It doesn't really matter if it's going to replace all of the animators or if it's good enough. The problem is if it's replaces even a small amount. It's already hard to have that job you don't need even more competition and problems. And this AI can create animated character that you need a month to create animate and render in seconds and it might be a couple of dollars for a client or $3,000. What do you think he will choose? I think this is a problem but thank you for your video.
Honestly as an artist, even the generative AI is good for rough ideas, but for specific concept that actually has good design that fits a specific purpouse,style and world... the AI is basically useless.
As for animation I wouldnt be worried about the video generative AI, but the one that animates. But I feel it's the same as the mocaps. Mocaps are cool, but are useless for a good animation, even for a realistic character.
thats what you think now. just like how the people in the past didnt even think it was possible to fly, or talk over distances in the hundreds of kilometers
@@Hyldassims well yours is a dumb reply. People in the past didn't think it was possible to fly because it would take thousands of years before there would be enough technological advancement for a man made object to have enough power to lift up it's ass from the ground. You can't blame someone for not having a crystal ball. It's easy to judge when you have the comfort of looking back.
So yeah, maybe AI will be able to do it tomorrow, or maybe it won't be able to do it in the next 100 years. Who knows...not you obviously.
At the moment AI is good but not good enough.
@@Hyldassims Mocap is 95% useless, it's a fact. AI trained on datasets made with Mocap...you guess the outcome :)
@@Hyldassims In the same way, people in the past imagined a lot about the future that did not happen. Because progress is not linear and therefore all your interpolations are false.
No one gets the point here. All I hear is everyone saying that because AI is not good enough now, it will not be a problem, now or in the future. Well that's clearly not the case. How hard is it to understand that just because something isn't good enough now, doesn't mean it cannot be good enough in the future.
The gaming industry isn’t going anywhere so y’all don’t need to worry. I just feel bad for concept Artist. I have a startup gaming company and I never felt the need to bring on a concept Artist because AI has given me great images
Interesting, I find AI to be a bit limiting in the styles it produces, I only use it to pitch an idea to client. There's just something missing about AI art. some element of creativity.
The end is near
I agree 100%.
In 3d u get super pixel detail and flickering but with ai not u cant and u cant reveal large perspective
100 % agree
Hello sir, Great video ❤
As per my observation there is still jobs in future also but numbers are definitely decreasing. And competition is increasing. So as a artist may be we have to work on low wages. I am also working as a game designer. My company don’t need team of concept artist and graphic designer and rigger also due to ai. Only small team can produce great results. Please correct me if I am wrong.
Thanks again for sharing your valuable observations ❤
I would know untill I see how the game look... Some company are dropping ia because the art look very generic and derivative
I LOVE U
To your point, I think the issue here is with the lower-end tier of the media market. High-end production will still have its place, but how do you get into a high-end market? You must cut your teeth on years of lower end work to build your skills whilst still feeding yourself. When those opportunities are gone-replaced by AI-there will be far fewer professional creators out there who can survive long enough to build expertise and recognition to be marketable. The segment of the market exposed to automation will grow as this technology improves. My heart and my brain are telling me different things here. I only just started learning Blender about a year ago and definitely don't want to stop the journey. I'm conflicted. At any rate, I'm searching for reasons to be optimistic, so thank you for your video. You have a good message regardless of where the future steers our collective circumstance.
Well first, there are student that get their first journey by in big studios... They work very hard to achieve that.... Also big studio, will always need new junior or new people to replace the old one, is part of the cycle... So they will need to adapt too.
Another thing... There is plenty of small studio that are dropping ia image already because look too generic...
Damn Alessandro I am so afraid right now 😔I haven't even entered the industry yet and I'm so scared about the future
It's not too late to change, are you younger than 25? Become a programmer and never have a fear for job the rest of your life.
@@Exsulator2 I'm 24 and I've been studying for a year now. I made the decision to change careers to animation but now I don't know what to do anymore.. completely lost and terrified.
Don't be terrified, life takes some turns and we don't know yet how things will go. Worst case is not the worst... you're looking at a career shift as the worst case. It's not like you'll starve or anything, you'll manage. You should play it cool for now and wait a year or two, before making rash decisions, but if you start seeing hard proof that your future is in danger, keep an open mind about pivoting to an adjacant career. Even consider being self-employed. Also if you're really talented and feel like you've got confidence, then I think you'll be fine staying in animation either way. I don't know man but good luck. @@Supinori
@@Exsulator2 Thanks my friend, really appreciate your take on this
@@Exsulator2 I'm so tired of seeing this suggestion. First of all, you really think that programmers are safe? Maybe some of them (the 'elite' ones), but there are different levels to this stuff, not everyone is a software engineering genius. And honestly, why waste my life doing a job that I have zero passion for? I'd rather keep doing what I love, try to adapt my workflow when needed (and staying optimistic, using ai as an advantage). Because if you don't pursue your passion in life, what's the point?
I think keyframe animation will become less popular. It is like using a hoe instead of a tractor. Video Mocap with AI will become even more viable for character animation. Only cleaning the data and implementing it to the engine will remain.
Mocap exist for 10 years already... Plenty of studios that still use key frame still
Ok but this describes text to video AI. What about AI such Motorica which uses mocap data to create walk/run cycle. The outcome can be also customised and they seem to be supported by Epic Games. How this will impact animators? Or AI assisted physically based 3D software such as Cascadeur which helps with posing.
Mocap was meant to be a replacement of animators, thing evolved and every iteration had more fidelity compared to what the actor in the suit does. Now go and work on a production that uses mocap...half of the time you remove 99% of the keys and only leave 4 or 5 poses and that mocap data is only used as a check of what the director wanted to see, specially when it's not a human character. Sometimes it works and with a minor cleaning and some layer on top you can do something decent but it's not the norm, don't see how this is gonna be different, specially if the model is trained based on mocap data who's not production polished and has no aware of it's surroundings
@@carlesv7219 I’m not sure about cinematic animation. I think Motorica AI focuses more on gameplay animation specifically cycles such as walk/ run cycle ..
HEY ALESSANDRO GREAT TEACHER,, According to you, animation jobs are safe in the future RIGHT SIR...?
they will still be,
There are a lot of professions that people said the same: how can a robot or a computer do this? Only a human can do it. And they are unemployed now.
Your "yet" says everyfhing. AI development evolves in geometric progression.
Machine can do a lot of things... But some stuff are better when human emotion are involve... Art and animation needs that to transmit something to the audience...
@rota The problem is what a machine do is very similar to what most humans do when are creating something: they get references and transpose those references to a new stuff. Most of the work of artists is based on references (including animation) and machines can do it as well - maybe better than a human because they can use big data.
Few artists are creative enough to present a completely new work based only on their imagination.
I think sooner or later, machines will "understand" emotion and will simulate it as good as a mid/good animator. I hope I'm wrong.
Love the fact that you said (YET)….cos y’all will still panic soon. Esp when you just a specialist and not a creator. Cos this technology has made a level plain ground for everyone that’s creative. Not everyone can make it Hollywood to work for big studios. So why not make our own stories easily with available tools that’s cheaper and easier. My thoughts tho
Artistc are all creator for definition... The beauty of art is create somenthing base on your emotion.... Sure you can use ia to make video.. But so will be able to do the same 100 millions of people... And 99% will be garbage...
@@acamporota Yes you are right. As we speak AI cant be used to pass emotions. But I’m hoping when it gets to that level, it should be limited to by a creative law or something so that it won’t disrupt the industry too quickly. The reason why I believe AI will get to level that we can’t even tell the difference is this, the first images AI created years back was trashy, but see where it is now. Same as AI video only God knows what the ability will be like in the next few years, cos the developers will want to get a raise by making it better by solving the problems pple complain when using AI. Also competition will drive them to be better amongst their counterparts. (Sorry I’m just stating the sad reality of the future, my hope is that a strict law should be set for how pple have access to these AI tools)
@@joeharmony6949 of course will be indistinguishable... But that's not the point... I give you an example... 25 years ago if you want to appear on TV and speak on an audience... You had to go on TV.... And only few could do that... Now with intent.. Any idiot can have a TH-cam or tiktok account and create content with a phone...
But.. 99% of those are garbage with no following... Also we are still watching movies and TV series done by professional artist.... Because people who study a craft for 10 years will always have an edge....
What if it comes to the 3d field and starts generating 3d models? Could it be the end of 3d modelers? (Theres some AIs that are currently doing this but they're horrible atm)
Not sure a studio can use ia to generate a 3d model, because there will be copyright issue
There are a couple of significant problems with using neural networks for generating 3d models: 1) neural networks work well with so called regular data such as pixel matrices, but polygonal meshes are not such data, you need to use 3D voxel matrices, and this is very expensive. 2) data sets for training, there are not as many free 3D models for training as 2d images and videos.
@@ComradeHugo thank god
@@ComradeHugo But what about AI generating a video of an rotating object and then theres another AI that will 3d-scan the video do generate the model?
@@higorss It would be nice if you provided some links, but as I already said, neural networks can be applied to voxels, and the voxels can already be translated into a polygonal mesh using existed algorithms. In theory, such a neural network can be trained on scans of real objects, so that it can think out and complete the invisible parts of other scans. But what about non-existent objects? What if you need to get a model of something specific and not random? And I have not yet seen an example of applying neural networks to polygonal meshes directly, so we all still do UV unwrapping , retopology and a bunch of other procedures manually.
Not sure about it..midjourney first versione 1 years ago was terrible and now reach the phorealism..in a couple years we will have animation expression and prompt that could reproduce perfectly any character or acting from actor or cartoon from any studios or past animation..Just need time..the trick is use AI to make your work easily and integrate with your job..we need to surf the wave, as in the past we integrate some tool that help us in post-production and replace technical stuff.
Photorealism has nothing to do with craft, art, direction, and emotion. I saw plenty of those extremely photorealistic picture.. After a while they look all the same... Nobody deny that ia can be use to improve our work... But not replacing it by writing a few prompt....
For sure there will be a market for those too.. I mean.. Tik tok is full of garbage content extremely popular.. Mc Donald idem...but there will be always space for artistic content
@@acamporota what artist does? When U do animation you use reference.. pose, when U create art you use reference, past artist. Imagine a AI can have access at infinite possibility based on our cultural artist background. It will not take too long to said to it, " create an animation based on blue sky studios where scrat do some movement jumping on the wall and interact with other character, while is doing an weird expression, stopping in the door before go inside,and create multiple example of this, like 100 version.based on principle of animation" It will use reference based of thousand source and video and movie to create the perfect animation in less of 1 minute. Photorealistic or not. Probably the new tool instead Maya will begin the AI. Who know..change the tool it's only that the difference. You give the command to the AI. Nothing different to give a command to a tool to move a rig.
Just use the tools for your advantage. Horses are still around after centuries of using other vehicles. If i have a software that gets me a cool render fast or things like that, why not use it? lets face it, part of the process is pretty tedious and if there are tools to speed things up then good
Do people really like doing stuff like UV mapping? really? lol i dont think so. Some tasks should be sped up, no one is gaining anything from this
there are things that would be nice if they were speed up, but taking away the creative process won`t bring any good,
@@acamporota that wont happen, machines at the end of the day are just machines. People will get tired pretty soon if there isnt humanity involved , and of course this things cant think
yep, but for some reason, despite all the AI hype, we still do UVs by hand, hehe.
@@ComradeHugo for now, im sure it will get better. And its a boring task, people could be designing more creatures and characters but they have to deal with technical non sense . Its good to know the basis always, but if its a task that can be automated then i dont see it as a negative for creatives
@@cosmozapallo4530 What are your positive forecasts based on? I haven't seen any results of applying neural networks to polygonal meshes so far, neural meshes only work with regular data like pixel matrices and I don't see why that will change. I'm afraid this is just wishful thinking on your part.
I've been trying to incoporate AI into my work from 2018, they're so underwhelming. I've even tried inbetweening with AI. In the end creativity comes from life. Both animals and human beings have it. AI is too logical and generic.
AI is a gimmick guys. With more technology, the human touch will be more valued. We see it with animation movies and the shift away from polished computer animation, to films where directors want finite control. AI could be used for texture generation, that's the best use case now
I think we should ban IA in the industry
That would be just the same like the big oil states and companies, trying to hinder transitions to clean energy. there is nothing inherently bad about Ai, it just competes with you, like how clean energy competes with oil. sooner or later you will probably need to understand how to use ai to your advantage, to have it in your workflow
I think it shouldn't replace a human completely but if you have A great animator or even A great writer write a very complex prompt and have AI outcome be iterative we should be able to have similar animation that we have nowadays, with the emotion needed. And maybe the animator can touch up a couple of frames. It's just a matter of time for these AI tools to evolve to that point.
It still is faster to rewrite part of a prompt than to keyframe or mocap shoot and integrate.
So the animator workforce should be severely impacted by these (in some years) if regulation doesn't come in a timely manner (which I doubt it will). But one can only hope 🙏🏻
I see so many videos arguing why AI won't affect their profession and just pointing out all the flaws. All those people will be left behind. AI wont stop and will only get better. First movers on new tech reap the rewards. Look at what happened when digital cameras came out or when CGI exploded. 2:39
One mistake I think many are making though is assuming Sora will be available to everyone. I really think, at least for a few years, it will be invite only or maybe access to corporations for millions in fees. The energy consumption and server usage would be too overwhelming if everyone had access. Also OpenAI is for profit now with VCs involved. They were a much different company when they dropped Chat GPT to the public.
I dont think there is anything special about SORA or OpenAI. They got a headstart on the transformer technology and billions from microsoft to scale above competitors. And most of their breakthroughs come from external opensource research. So with time the others will catch up to them, likely in the near future.
SORA is not an innovation. It is based on diffusion models. The difference is that they have scaled it, because they have the necessary capital for it.
If you have seen what videos AI can make 1 year ago, and compare it to SORA now, and you understand that it can make such an improvement in about 1 year, can you imagine the improvements it can make in the next few years to come? Sure, AI may not be a significant of a threat as of now, but that is here and now only. It will not be the same couple years down the road. AI improves at a much faster rate than humans possibly can. You mention that there are things that us humans can do, that AI, machines cannot do. But years ago, did anyone even imagine cars driven autonomously by software? Centuries ago, most people did not even think it possible to fly in a tin can, or reach the moon in a tin can, yet here we are. Just because you think it not possible, does not mean that it will never be possible. Nobody (me included) likes and wants to admit that AI will be here to replace you, but it is a fact, that AI is inevitable, and there is no stopping it whether you like it or not. A year ago, I refused to believe that AI could replace me, or my skillsets for that matter. But now, I have come to accept the harsh truth and reality that AI can and will replace anyone and everyone. It is only a matter of time.
You are extremely pessimistic, man. Do you think that human beings will sit back and let AI take over everything? Obviously not. Just as AI should not be underestimated, we should not underestimate humans either. We will end up self-improving ourselves. The human brain unconsciously operates at a much faster rate than the best supercomputer in the world. It performs extremely complex mathematical calculations all the time, implicitly. And you, as a good human being, make the mistake of basing your hypotheses on nothing and, most likely, they will fail, as has happened throughout the history of humanity. Human beings can self-improve themselves. We are not at that point yet, but it may arrive at some point in history. Everyone talks about super intelligence. But no one talks about a superhuman. As far as I know, a human of the stature of Von Neumann or Sidis is capable of becoming an expert in a subject in less than 2 weeks. Sidis was able to learn several languages in a single night, in addition to being an expert in many scientific fields at the age of less than 25. Now, imagine if all human beings were capable of reaching those levels of giftedness, or even beyond. If you believe it's not possible, you fall into the same naivety as the people you criticize. Regards.