OpenAI Sora: Next Level AI Videos! AI News
HTML-код
- Опубликовано: 7 июн 2024
- OpenAI Sora shows the BEST Ai Videos we have ever seen. Up to 60 seconds of totally realistic AI Videos. Gemini 1.5 can now use up to 1 Million tokens. Gemini is multi modal and can use audio, video, voice, text, code, photos, sketches and much more. Chat with RTX gives you are super fast LLM that can also use data you provide, like videos and pdfs. Face Fusion brings a new AI Video Face Swapping update. ComfyUI to 3D can turn 2D images into 3D models. Forge Ui is faster than A1111 and Stable Cascade is really good with Text and Hands
#### Links from my Video ####
Forge UI Guide: • Forge UI - 75% faster ...
Stable Cascade Local: • Stable Cascade - Local...
openai.com/sora
deepmind.google/technologies/...
Chat with RTX www.nvidia.com/en-us/ai-on-rt...
ComfyUI to 3D github.com/MrForExample/Comfy...
Face Fusion: github.com/facefusion/facefusion
#### Join and Support me ####
Buy me a Coffee: www.buymeacoffee.com/oliviotu...
Join my Facebook Group: / theairevolution
Joint my Discord Group: / discord
AI Newsletter: oliviotutorials.podia.com/new...
Support me on Patreon: / sarikas
00:00 Intro
00:08 OpenAI Sora News
03:45 Face Fusion 2.3
04:22 Forge UI
04:35 Stable Cascade
05:01 ComfyUI to 3D
04:45 Gimini 1.5
08:57 Nvidia Chat with RTX
10:07 Outro - Хобби
#### Links from my Video ####
Forge UI Guide: ruclips.net/video/Th-jfath02E/видео.html
Stable Cascade Local: ruclips.net/video/Ybu6qTbEsew/видео.html
openai.com/sora
deepmind.google/technologies/gemini/#gemini-1.5
Chat with RTX www.nvidia.com/en-us/ai-on-rtx/chat-with-rtx-generative-ai/
ComfyUI to 3D github.com/MrForExample/ComfyUI-3D-Pack?tab=readme-ov-file
Face Fusion: github.com/facefusion/facefusion
👋
Sora marks a new era. Nothing less
It's the new age. Totally amazing
Wow ungrateful *****
@@OlivioSarikas the new age that doesn’t actually exist yet because it hasn’t been released yet, and it will likely be worse after all the safety stuff they end up doing to it.
Now I can believe we might be living in a simulation...
Apple has an app for that ;)
The Holographic Universe explains how you could be right. Stopped me worrying about being on a ball hurtling through space!
Ai gaining speed again... nice :)
Thanks for keeping us up to date, Olivio!
I have 98 pictures generated months ago by midjourney that were waiting for the comfyui 3D model generator to be used, thanks for letting me know!!!
Reflection in the window is so awesome
That one blew me away
Can't wait for more details on Sora. That video output in seriously impressive. Wonder how pricing and video generation times will be.
Sora's capabilities are absolutely mind-blowing! The visual realism, especially with its understanding of physics, takes AI-generated video to unimaginable new heights. I'm incredibly excited to see the stunning, creative experiences people will dream up with this technology.
Matrix world is coming...
@@iangillan1296 Feels like we're not too far from having AI create our dream sequences for us. Just hoping it doesn't mix up left and right in our dreams, or we might all wake up thinking we can moonwalk like Michael Jackson!
Thanks!!
Awesome
The way this is going is computer games may end up being a scaffold for interaction and rules, then an AI real-time skins the 3D world in the desired consistent art style.
Exactly my thought. Remember when rendering in a 3D software took ages? Now with graphics cards you can render a scene very quick. It will happen to Ai As well.
Sora looks great but as always with openAI products,its gonna be heavily censored.
What i loved the most about this demo is that now we know what AI txt2vid will be capable of, and now i want the opensource models even more
Cant wait to finetune my waifu model UwU
yep, gonna be super woke. But it will inspire open models
@@OlivioSarikas You're too intelligent to be using lame terms like 'super woke'.
@@Elwaves2925 You're not... maybe you should say it....
@@generichuman_ It
@@OlivioSarikas OpenAI (among other initiatives) has to walk on eggshells because ALL eyes are on them, kinda dissapointing to see you reduce the likely consequences of that to "super woke"
Tried to install the 3d thing, obvioulsy couldn´t make it work at first try, that´s the thing that makes me nuts the most...
In 10-20 years (probably less) kids are going to look back at us and laugh thinking why would you watch a movie or play a game someone else made up. You prompt your own movie or game and you're in your own imagination
or having movies that only have a single plot line in it
Was very excited to see "we will include provenance metadata in outputs" in their marketing blurb. Finally, tech companies acknologing where the data for their models comes from and providing a way to statistically plot its contribution for each output. I thought that was very "open" and even brave of them. But, no... . they just mean it will tag the video as AI generated or not. 😢
What you're asking is impossible and shows your lack of understanding of how AI image generation works.
In your mind, when you ask AI to generate "a funny cat image" you assume AI is using some specific images (which you'd want tagged) to draw the new one, which is wrong. AI simply learns how a cat looks like after having watched tens of thousands of cat pictures. It doesn't copy paste portions of images. It would be dumb to expect all those thousands of images to be tagged on your resulting image simply because they would be having nothing in common but the mere fact that they show you a cat.
you can made video about infinigen landscape generator based on blender and python
this is like going from gamecube to apple vision pro
I'm running to create movie scripts and test it.
maybe this can do good videogame, at last😮
I can for sure say that you need to post more
I'm one of the most active AI Video creators in the scene to be honest ;)
Sheesh 😮
That’s alarming 😂, just give it 3 months and you will see this in the hand of public or get shipped to animate diff 😅 and … that would stock videos end 🎉
If I was an actor right now, I'd be SHAKING in my boots!
smarter thing would be to put yourself in your own movie as soon as the tools give you enough control and ways to do custom training
It’s amazing but only if I can run it locally and off-line. :-)
absolutely. getting tired of all these break throughs being only for advertising plans.
That'll come. Maybe with a 7090?
@@peacefusion openai says it will not go public. hahaha
I dont care about openai. I care about those that do open source. @@Yue4me
@@Yue4me Ah i think it will become public in the future, but behind a paywall. Like RunwayML. Let's hope that Stable Diffusion catches up quickly. Sadly Stable Video Diffusion showed that even this is no guarantee for not ending behind a paywall. It has a non commercial license ...
I've seen that exact wolf video in archive stock footage. Impossible that its not AI stolen IP
So you have seen a wolf video before, and nobody else is allowed to do a wolf video anymore since it is stolen then? Tell me more :)
It already starts with the fact that it is technically impossible for AI to steal. This is not how AI works. It would already be simply impossible to store all this "stolen" data in a weight. The bigger version of Laion 5 is 240 terabyte big. Now tell me how you would fit that into a graphics card ram with 8 gb size. The nobel price is waiting for such a genious compression method.
It can happen that you just have one training data available. And then there are of course not this much variations possible. Maybe this was even the case here. And then you could get the impression that it looks too similar and is stolen. Like with an image of the Mona Lisa. There is just one Mona Lisa as a matter of fact. But even here the AI does everything but a pixel perfect copy of it. Since, to repeat, this is not how AI works. It learns the concept. It does not use a single pixel of the original input. It does not steal. It can't. The way how AI is and works does not allow it.
great technology
ok, lets watch sora videos for the 3rd fking times .. but .. it is ok .. they are cool
1:50 no way in hell that is AI generated.... no way!
Why I have feeling that kangaroo is prompted by Olivio? ;)
It's too sexy for it's shirt ;)
Now I remember why I don't subscribe to this channel 🤣🤣💀💀💀
HOLY FUCKING SHIT
SORA looks amazing. I have heard it willhave tons of censorship of named people, NSFW, and artist names. A shame, so after a few months the local version of something similar will come out, thanks goodness. SVD is ok, perhaps SDV-something will be comparable. I only use SD for my prompt engineer job, except the quick MJ pic for something unspecific. I do hope I am wrong about what I hear about censorship. Much rather replace my 3070 with a 4090 than pay for subs that limit what I can make!
Also, I am thrilled that everyone will soon be able to make movies. I have always felt celebrities who are rich for movies, or music, or similar things.... WAY overpaid. Let's being democracy to movie making. Soon. Mmmm.
Would make sense for a commercial model. They need to make money somehow
In about 50 years the Ai will be able to update itself without the work of humans.
you mean 50 weeks?
When Sam Altman took requests on Twitter-X it looked less impressive, truth really remains to be seen. Cascade is underwhelming. What kind of cost or hardware requirements would this incur too.
true. not sure that was a good move. Then again, prompting isn't easy in any AI tool
Spelling error in the title.
Really? What's wrong?
Sora yeah, but can it run Crysis? 😂
the game?
@@OlivioSarikas ja, als es raus kam lief es so lahm selbst auf guten PCs und als es bessere gab konnte es nicht multi core skalieren, so kam der Meme auf bei jeder technischen Neuerung zu fragen "yes but can it run Crysis" 🤣
@@FunIsGoingOn I'm old enough to remember the meme 🤚
Openai ⚾️🦇language ⚾️🦇 vision ⚾️🦇 hearing ⚾️🦇 images ⚾️🦇 video ⚾️🦇
Is openai just creating genious marketing trick with cherrypicked materials?
inb4 scammers use gemini to write realistic fake comments about their own content
You were just made completely redundant and obsolete guys, sorry. Just move on. All you have learned in the past year was for nothing, sorry.
It just means that a large number of people who work in film and animation will be left without jobs.
Lol. No it won't. It oy means effects and movies will become more complex and new jobs will be created
@@OlivioSarikas I'm glad to see your optimistic outlook on this matter, and I sincerely hope you're right.
@@tanyasubaBg just look at history. New technology is always massively overestimate, until we see it's limitations but also how much more we can do with it. Look at vfx and cgi. Sure practical studio model builders largely lost their jobs. But now movies cost more and are massively more complex to make. Because the effects are way way more complex. Thus more people have a job and more money is being made. Plus even indie studios can now create movie grate films and shows for Netflix, making the movie landscape bigger than ever. 5 years from now tens of thousands of creators will make movie grade shows for RUclips in their home studio
@OlivioSarikas Thank you very much for this wonderful and optimistic response. I truly hope it will be so. I see that many people, like myself, are afraid of where all this will lead us, but we'll see. In any case, your attitude is encouraging.
Stable Cascade is useless for the majority of users who dont own 24gb vram cards.
You don't need a 24gb card
More jobs go poof
If jobs wouldn't go poof, we would still live in caves. Or would you subsidise a million horse cobblers, just because we once used to travel in horse carriages? There will be billions of new jobs created by AI. You can either learn how it works now, or retire early
Yeah jobs vanish but the money is generated anyways, so who will get it? 😉 Inequality is the real problem here, technology is fine.
@FunFreakeyy correct ...it's happening too fast for us to react. These are huge political social economic changes
I'm not sure if people realize that a tool like this released to the public will become a timer to doomsday. I Genuinely hope they will nerf this a little or prepare to live the rest of your life with the inability to distinguish truth from reality whenever you are connected to the internet.
DALLE3 is nerfed for the same reason so I expect OpenAI to handle Sora exactly the same.
Now I have to think that you mispronounce things on purpose because mispronounce in Marquez Brownly. Come on you must not watch this video content if. you don't know how to pronounce his name.