What's All The Fuss About Sora?
Вставка
- Опубліковано 6 лип 2024
- Thanks for watching. Let me know what you think about Sora in the comments.
Discover other AI tools, tutorials and creative talent at aianimation.com
This video is a quick walkthrough on what Sora is and my initial thoughts.
------------------- -------------------- ---------------------
Times
0:00 - Sora
------------------- -------------------- ---------------------
Tools Used:
Sora:
openai.com/sora - Фільми й анімація
This is so cool, thank you! What a time to be alive. We are living in a Renaissance.
I think it's one of the things I've waited for the most ever. I'll finally be able to bring to life all the projects I had in mind. For each one, there will be the opportunity to expand one's creativity, and technology doesn't necessarily have to supplant the quality that the industry brings to the table thanks to technicians and artists. The combination of these two entities will evolve cinema and art in general.
I can see Sora working really well for advertising and maybe for making music videos for Tik Tok, but an actual movie might be a long way off. I tried making a series of images for a fantasy style illustrated story using Dalle E 3 on Chat GPT4. As I developed the text for the image, Dalle/GPT really seemed to understand the idea and even expanded on some of my themes. Yet when it came to producing images, I could very quickly get one good image, but when I tried a follow up image in the story, consistency of character was a nightmare. Even using the exact same text description of the character, there was always something that changed , like a hair style and item of clothing, or a location. Even the most simple locations had significant changes in them and when I had more than one character or animal, Dalle just seemed to go crazy. The odds of getting a reasonable follow up image were about 50 to 1. In the end after using hundreds and hundreds of images I gave up. I can see exactly the same thing happening in Sora unless the images actually consistently obey the text input and have a form of memory. The images we see in the Sora examples show no evidence of memory. Saying all that it is an exciting breakthrough even if it's just 1 minute
I have your same problems but all can be fixed and open Ai it's working for agi not just for the ai the we know so be patience and develope your skill in production after will be a joke for us to make our work, compare with what need to do it right now off course....
I disagree, I was in the same line of thinking as you about 6 months ago thinking ai could never even make video because of how incosistent it would be btween each frame; but look at Sora.
Have you tried using the same seed? I can duplicate somebody else's images by using their prompt and identical seed. Currently, I use NightCafe.
Sora will be really fantastic. Probably it will be used in science also as a physics simulator, after being really precise. Seeing that kind of evolution make me imagine the unlimited potential for its usage. We are just in the beginning of 2024...
I can not wait for it!
Wow .. looks amazing. Do we know at this point if it will be able to use an image as a starting prompt?
Thanks for keeping us up to date!
So I'm seeing this, and if they're prompting this to get video like this...are we witnessing the first holodeck engine? like plug this in vr and talk to it in vr and describe the landscape or scene you want to walk in. I didn't realize we were potentially this close to this kind of tech. Add realtime audio, and combine it with gpt writing capabilities and you'll have fully interactive immersive stories.
Yeah... my mind keeps stumbling into similar trains of thought. If OpenAI (and who ever else) is able to create Physical World Models (PWM) rather than Large Language Model (LLM) to generate content as they strive toward AGI ... the holodeck ideas (along with evolving VR/AR tech) seem more achievable.... also it seems far more than my brain can comprehend. 🤣
Maybe solipism is true. AI shows us what is possible.@@AIAnimationStudio
We have no idea when it will come out or who will have access though, but indeed groundbreaking
Chances are, this model won't be available for the public anytime soon, and when they release it, my guess is that it will be an enterprise level product since this would require massive compute. We can only hope the open source caching up and cheaper GPUs 😂
Is Sora live?
has anyone you know got access to it?
Not yet... but if anyone at OpenAI wants to say hi... I would'nt complain 😃.
where i can see sore?
at chatgpt like dall e?
Link in the description. You can just see the previews at the moment whilst they further test and put in safety aspects
How can this affect video games?
You could for example have it generate a 360 video rotation around an object scene. Then use Gaussian splatting to create an accurate 360 model. Cut scene generations. R&D of ideas/concepts. Quickly iterate on art styles.
Plus perhaps eventual impact on advancements in rendering (?) and simulations. As compute power advances... would Sora be capable of instant interactive video generation in what are essentially world simulations?
this impact a lot of game procedurals, animation, vfx, you never give up , adapt always, i'm a game director of procedural games with ai, i need it like sora
@@AIAnimationStudio I know feel bad for 3D Artist/Modelers. 3D scanning tools combined with this AI Tech is going to be the norm when creating assets and even complex characters… I’m glad I started my own studio
Each person's essence is etched in their dreams and daily pursuits, especially in their chosen profession. Yet, the looming shadow of AI replacing human roles casts doubt on our future. It's confounding why some fervently champion innovations that jeopardize our fundamental right to earn a living. Beware, the allure of novelty can exact a devastating toll. Remember.
It's baffling why those in power don't halt technologies that threaten lives, and why courts or human rights groups don't step in. Ordinary folks uphold the world's wealth structure, yet their job security is at risk. It's the duty of every government to protect their people's livelihoods.
Bet adobe are cr@pping themselves. Hope 100k workforce is safe , ,but only for now.
Yeah... no doubt. It will be interesting when Adobe starts rolling out their own latest video generation tools, which when combined with their existing tools add some extra levels of control and will make it easier to make it part of a wider production process. But will they have anything close to Sora.... hmmm..
Also looking forward to see what MidJourney, Google and Meta bring to the table. Plus what Runway, Pika and Stable Diffusion have in the works.
@@AIAnimationStudio would be funny if they were all as shocked as us.
I think this is more exciting than when mj started to turn because this is more of a shock. It's gone from 'meh' to 'soooooo my gawd" in a flicker. Roll on 2024. Bring it on.