Filmmakers Test Sora + AI Actor Swapping
ฝัง
- เผยแพร่เมื่อ 8 พ.ค. 2024
- Find all of the links from the video here: curiousrefuge.com/blog/ai-fil...
Join Our Newsletter for the Latest AI Filmmaking News: curiousrefuge.com/
Our AI Filmmaking Course Starts Soon: curiousrefuge.com/ai-filmmaking
Check Out AI Advertising: curiousrefuge.com/ai-advertising
Be Part of the Biggest AI Filmmaking Event this Year: curiousrefuge.com/ai-filmmaki...
Time Code
00:00 Intro
00:16 SORA’s New Examples
00:56 Paul’s Thoughts on SORA
02:13 Open AI Takes SORA to Hollywood
02:34 Viggle
12:42 AI Filmmaking Mega-Party
13:15 Brooklyn AI Filmmaking Meet Up
13:35 Haiper’s Latest Update
14:56 Adobe Structure Reference
18:11 Enrollment to Curious Refuge is Open
18:36 New Podcast with Matan at Pika Labs
18:51 Nicolaus Neubert Prompt Share
19:35 CHAMP
20:03 AI Misinformation
20:48 MIT’s Fast AI Image Generating Method
21:21 Claude 3 Opus beats GPT-4
22:09 Midjourney Video Update
22:48 AI Films of the Week
22:52 Escape
23:12 But a Dream
23:38 Prince of Egypt
24:22 Outro
SORA Examples: openai.com/blog/sora-first-im...
Paul’s Short Film and Thoughts: / paul-trillo-544b268_ex...
SORA goes to Hollywood: mspoweruser.com/openai-to-sho...
LA Times Article: www.latimes.com/entertainment...
Viggle: viggle.ai/
Abel Art’s Example: www.linkedin.com/posts/abelar...
Max Einhorn’s Example: x.com/MaxEinhorn/status/17713...
AI Filmmaking Mega Party Tickets: www.eventbrite.com/e/the-ai-f...
Brooklynn Meet Up: curiousrefuge.com/ai-film-events
Haiper: haiper.ai/
Adobe Structure Reference: www.adobe.com/products/firefl...
Learn the Latest AI Techniques with Curious Refuge: curiousrefuge.com/
New Podcast Episode: • From Rockstar to AI Cr...
Nicolaus Neubert Prompt Share: / 1772654888696300019
CHAMP: / 177259201188479018
AI Misinformation: / 1772752077057364226
MIT’s Fast Image Generation Method: news.mit.edu/2024/ai-generate...
Claude beats GPT 4: / 1772759835714728217
Midjourney Office Hours Break Down: / 1770547499549294865
AI Films of the Week:
Escape: curiousrefuge.com/ai-film-gal...
But a Dream: curiousrefuge.com/ai-film-gal...
The Prince of Egypt: curiousrefuge.com/ai-film-gal... - บันเทิง
Curious Refugee, you do a great job rounding up the latest news in AI filmmaking, it's no small feat to keep track of it all!
We appreciate it - lots of news to cover every week :)
Cool, cant wait for it to release to public, i need that every day
Thank you - very informative and hence very helpful! 👏🏻
Glad it was helpful!
Wow Firefly style updates! 🎉
very complete information, I love your videos! they're SO useful. thanks!
Thank you so much for featuring us this week!
Our pleasure!
I would love to use these tools❤ thank you for the video!
Our pleasure!
Thx Caleb, this show is always a must watch!
Thanks so much! :)
Fine concept, I love it!
Glad you liked it!
Thank for this video ❤
You're welcome 😊
Oh the venue sounds awsome n broiklyn yay
The description is very detailed
Thanks!
U made my day w viggle n reminded m3 of runway remover in video
WOW!!!! Not going to lie... I've been getting a tad bit more annoyed with how slow and buggy my Adobe Software has been running lately, especially on high-rez work... I did NOT expect this. Midjourney has been great with Style Reference... But this looks just a little better for my particular needs considering my workflow. I am very impressed! Now, I can FINALLY integrate my Textures library into my process in a less manual way. Being able to have full control over comps and everything, without worrying too much about refinement, and then running my comp through this before upscaling and overpainting... AMAZING! Thanks for sharing, yall literally fuck up my entire workflow with every video. And I love you for it. Thank you for another awesome video!!! 🖤🖤🖤
Because many of us are utlising these tools not only based on quality, but on quality vs cost- could you consider throwing in more info on subscription costs, billing etc in these episodes?
Sure! The problem is that these figures change regularly, so we may be giving information that is outdated the day after it's released.
Отличное видео! Продолжайте делиться такими полезными советами!
Thanks for watching!
I'm struggling to see why Adobe can't make their rotoscoping as easy as this.
Thanks for your perfectly concise tutorial. Mind is opened.
You're very welcome!
Absolutely loving Haiper. I've been experimenting with it the past month and it might really be better than Runway.
We love Haiper too!
I want 3d before video. Sora already gonna knock it out of the park. Midjourney should be able to use 3d models that I can animate.
17:43 Would you get better watercolor results if you set the style weight to 1000 (--sw 1000) and left the image weight as default?
Thanks for showing us these incredible tools! Congrats on the LA Times quote!
Great suggestion! We'll do a bit more testing and see how it goes. We appreciate the kind words :)
idk, i sent you a few MJ watercolor birds from your images sources (admittedly screenshotted) and i think maybe your settings on the sref might have made it fall off track a little :) but i could very well be wrong. I have been known to occaisonally be wrong.
Let me check! We appreciate you testing it out :)
I wonder what the interface will look like for these tools. And I wonder how often the user prompts them, and how easily manipulated the results will be. More control needs to be given to the user to keep their creativity intact and even accelerate it.
For SORA? We will see! But very likely within the ChatGPT type of interface (at least in the beginning). You can start to see a bit how it may look by using DALLE and inpainting all within the interface.
wiggle is amazing!!!!
Thank you for your videos. The link for champ is not working or it may just be me?
We'll check! Thanks for the head up on the link
Great video!! - DJK
Hey! We've recently started creating our own AI Films. It would mean the world to us if you took a moment to check them out.🙃
Thank you :)
THANKSSS>......is there any way I can add two characters into one photo? For instance, I have an image of a girl and a photo of a cat. How can I use Midjourney to merge these two into one image while ensuring they fit the script? For example, if I want the girl and the cat eating rice.
You can prompt that and it will make that image just as you did in your comment :). You can use CREF to make sure the girl has the same face :)
@@curiousrefugea girl and cat are eating rice, --cref (link for girl) , --cref (link for cat) ,,,,i did this and get error, prompt error
Oh was just saying to huns abt see if ur course was open
Did u guys miss googles new video analisizer??? Hope thats in the next 1. Looks interesting ;)
We try to get as much as we can. We'll take a look and see what we can do!
When you generated the viggle you could of used "template" option for background.
Very true!
what if you used topaz to upscale the outputs so the frame rates could match ?
You could do that too!
amazing
Thank you! Cheers!
Thanks
What if you used mid journey and told it to produce multiple angles of a character design so that it actually had footage of the front side and back of a character based on what mid journey had generated.....
Check out my two new AI-made movies. th-cam.com/channels/TP8fHZ0FBkRAKjHq9thAWg.html
This is certainly possible! We chat about it in our discord server :)
Only BIG BAD THING is that all of these are subscription based and all content produced is not local but global so no one will ever own rights on their things without default implication of the company that handles the software. AI should be 100% Open Sourced and given to anyone for fair use and not give monopoly to companies. Such a shame to see towards what AI is going considering how it all started
Interesting perspective, but you are able to run (albeit it's quite technical) many of these generational tools locally.
Once the tech has been developed, you might be able to run it locally. Look at Stable Diffusion, they are open sourcing their AI model to be able to run locally.
@@cocreativeAI Yes but Stable Diffusion is also now seeing issues that I've mentioned, reason why their CEO Emad stepped down from his role as chairman of the company. People will try to gain the upper hand and gain money instead of thinking about the decentralized option. Open AI is also another fine example where they went from the whole OPEN to Not so Open (well almost completely closed) AI
@@curiousrefuge I know, I am running both SD and SDV locally, but I am mentioning companies like Google/Meta/MS/Adobe/Midjourney who have closed their sources. I like that some of them give their stuff for free to try (shame on you MJ) but still, you don't have privacy with those or are delimited to what can be generated (use of specific names/topics/detailed depictions etc. )
broski please how can i submit my film. i tried the website. first AI Action Movie
Where can we see
That's perfect, we'll take a look. Is it on TH-cam too?
Dude, you looks like young Christoph Waltz, AI version of course
Why thank you ;)
So can I use surah now I have like amazing stories that I would like to bring to life using AI but I don't know if I can use surah now or if it's like not available I don't know when it's available so can you like please give me some news about that.
Because if I could get my hands on surah I could really start making some very good movies that I think would really start getting me some attention and with some attention I can pull people into the faith I'm in which is Christianity.
Sora is only available to a small number of studios right now. We may get access to it later this year but it might not be the full power version. I would start experimenting with tools like Pika, Runway, and Haiper in the meantime. If you make something good enough, you might even get an early Sora invite.
@@user-on6uf6om7s oh that's great thanks
Sora isn't available yet - and will first be to the studios, then hopefully general population :)
not to be the negative guy but if someone finds a workaround for the safety filters you could make some really nasty stuff I hope they really lock and key the security on this thing :/
True! People certainly need to be careful with these tools!
What's so funny, is these companies who hold back from releasing updated features, now run the risk of getting behind the tech is evolving so fast..fun to watch innovation break capitalism 😂
👍🏽 👍🏽
What can this do that can't be done with Autodesk Maya which does it all better (720p😊 vs4+k) with more control, more surrealism and realism, magnitudes cheaper in terms of resources used ($80 gpu cards vs $30k per gpu card) ? Forget the power consumption.
We'll have to check that out - one of the issues is that it's far too technical for average users.
Thank you so much!!! Super great!!!
So glad you enjoyed the content!
I'd like to see a decent result. Usually the tutorials says "look, this is easy, of course you need to fine-tune this". The question is if this fine-tuning actually works.
If you ever have more indepth questions feel free to jump in our free discord community. We love to talk and walk people through things!
Claude Opus still not available in Germany if not using vpn software….
Oh, interesting! Thanks for the clarification!
@@curiousrefuge Claude Sonnett is availabel but only with knowledgebase till March 2023
please have security real loaded security at the parties.
What?
was my concern for your safety confusing? sorry @@curiousrefuge
Sora ai will replace 3rd artists?
At the beginning, it will enhance 3D artists who embrace it and use it in their own work :) A win-win!
Something tells me that it will be difficult for the general public to get their hands on these tools. Of course, you can already access it for free on the internet through a simple browser. But it's still too rudimentary to get anything exploitable. And something tells me that the development studios will be pressured to prevent individuals from using them. In this I am not optimistic at all. It is all the more annoying that I have been writing a story since last year and that I would like to do something with it with one day.There are also ecological and environmental considerations that must be taken into account since all of this goes through servers that are extremely energy-consuming.And as much to say that it will probably not contribute positively to global warming, that all the rest. I am talking of course about future AIs who will be able to do things much more intuitively than now.
True! Some of these tools may be difficult at first - however ever since the announcement of Sora , many competitors have made enormous advancements and so the first 'sora' may be accessible soon via a different app altogether.
I hope. Because I would like to do something with my story. I have the ideas, the character and its environment, I have already written tons of pages even if it remains a huge draft.And I admit that I have never used such tools at the moment. So I don't know anything about animation and animation software or sound. Hopefully in a few years there will be an all-in-one that will do everything with one tool.
The videos tools are just not ready for production. Full stop. Audio tools instead could be used in a different range of works.
True, for many projects it's not quite there yet. However, for smaller niche projects and for concepting it's an unbelievably great use!
WTF, the take part gold stones makes me very dizzy!
Very trippy! :)
Sora is starting to get old - it can do this, it can do that.. but apparently you can't have it because at first and for a long time (unless competitors catch up) it's only going to be available to Hollywood. Maybe start investing time into companies that want to invest in you.
Interesting perspective. We do love so many other tools :)
It's hard to trust people who use the term "misinformation"
Either way, we appreciate you watching!
1.sora 2. Haiper 3. Adobe
Seems a bit pointless to get a professional filmmaker's opinion on Sora. *_Obviously,_* he's going to prefer real actors, real cameras, etc because that's how he's made movies his whole life. A professional filmmaker is going to want more precision control than some Joe Blow wanting to piddle around for fun.
This is sort of like getting Paul McCartney's or Van Halen's opinion on Suno AI. Both Suno and Sora are intended to bring music and video (respectively) to the average Joe, not to be utilized by professionals who have millions of dollars at their disposal. Obviously, Suno is never going to offer the precision of an actual recording studio with real people playing real instruments, nor is it meant to.
Suno and Sora are just the "Wix" or "WordPress" of music and video production.
We see your point, but this first release of Sora is basically a beta. Consider this - in 1-2 years, Sora will have the control that you are talking about so it will be interesting to see what all types of professionals say about it. We can't wait to see what happens :)
@@curiousrefuge 1-2 years? I doubt it will be that long, as rapidly as this tech is evolving.
By the end of the year I expect Sora to be available to the public.
Another form of pixar
These tools are helping to democratize being able to make animated stories :)
Wish we had a streaming ai service! Non Politics non woke just good entertaining ai driven stories and videos!
Bit unfair only filmakers getting the tool...
Don't worry they get to test the tools that will eventually put a lot of them out of work
@@soggybiscuit6098or the studios will make it inaccessible for indies and also lay off their own people too
@@OldNerdLogandoubt it. Do the maths.
10s of Millions of users paying $50-$100 or hollywood who churn out prob less than 100 sci fi epics a year (inc prime,netflix etc).
@@OldNerdLogan OpenAI stands to make billions off of subscriptions to indies. Hollywood would have to top that money, and they might. But then the tech will eventually be cloned for public subscriptions. I''ll pay it. HW might be able to stop it for a few years, tho.
@@OldNerdLoganThen someone else will make identical toolsets you can use.
I think sora will create new "art" shows, where "artists" put on display their prompts. Just like those weird art shows where a guy stacks sand buckets on top of each other, or an artists just spills paint on a canvas and calls for a million bucks. You know these weird reddit dwellers that say "i'M goInG To MaKe A MoViE" "I'm GoInG To MAke Star wArS beTTER than DisNEy", they're the one that will jerk off to this tech. It's certainly not going to affect the movies and 3d modeling in this current form. Digital art is cooked sadly, especially concept art. It's got so good that you can just prompt some concept and adjust it, the quality is good enough for corpos to save money on an artist
We appreciate your perspective and engagement! We'll see how this all turns out once Sora is available
This looks ridiculously complicated. It should be possible to supplant such a walking person by a prompt in a natural language, like: "Replace the gray-clad woman by the orange-clad". A filmmaker has only a few decades for his work. If he constantly has to work with tools of such a backward slowness, he'll be robbed of the opportunity to try out thousands of ideas.
u yapping show me a easier method
This is still emerging technology, much of this will change in just a mere matter of months. It will be interesting to see what it looks like (and also your perspective of it) when that time comes. We appreciate you engaging!
@@curiousrefuge Yes, it's quite obvious that such downsides must still, and certainly soon will, have to be overcome. What I could imagine is that until now they aren't because the LLMs (large language models) would still not accomplish such tasks reliably enough. They could lend a person supplanted because of a natural-language prompt additional legs when they interpreted "gray" as referring also to the ground a person walks across, or the like. Current LLMs apparently are entities one cannot easily overlook the workings of, high-end outcomes of avant-garde developments released a little hastily because of how impressive and usable they are already in such a state.
But it could also contribute much to the persistence of such more primitive procedures that whenever you talk to such a machine in ordinary language, this again requires tons of computational processes - which could often slow down a step just as much as the current, more old-fashioned procedures.
No wonder studios are agreeing with every union , because none if them will exist in 3 years
It will certainly be an eventful few years! The industry is changing :)
I’m not against it but ease of access to everything is lowering the value we put on real art. It happened with the digitalization and democratization of music with streaming . Now we are starting to see it in films with apps and streaming. More mediocre content will only make the mediums less special over time like it has done with music. So many people make great music now but no one cares because anyone can do it and it’s not special anymore. Don’t let that happen to film
We appreciate the comment and your engagement. We definitely see your point. Like all technology that get democratized, it's true it will mean there will be much more content (like phones/camera). However, truly good content will always rise to the top.
YEAY Now we don't need TH-camrs anymore! And other creative people
This won't replace creative people.
Which part is testing SORA? Kinda click bait-ish, it’s just reporting news ?
Hey there! We report AI Film News every week, Sora is not available for public release but artists have just recently been able to test it which was showed their work.
MORE SHIT MOVIES = MORE GOOD MOVIES!"
AI is very, very new - it's only a matter of time before the content gets to a point where the mass audience will no longer mind if it's AI or not (or can tell). We understand right now many of these videos look like a proof of concept or demo.
I don’t mean to be the old man yelling at clouds, but am I the only one that thinks Generative AI is not worthy of any admiration? I’m not against using AI to make certain mundane tasks easier, but I just can’t get behind something that (in my humble opinion) cheapens the creative process and art in general. I’m not looking for an argument, just a discussion. What are your honest thoughts?
These are just tools by themselves - just like Photoshop. What we admire are the stories that people make using these tools.
It cheapens commercial artistic endeavours, which is a good thing! Creating art for money is a corruption of the creative process and runs counter to the idea that it is an expression of self and a creative outlet.
I predict that this sort of AI will make it very difficult to make money from creating art, ultimately liberating it from exploitation by the incumbent entertainment industries.
I'm so happy that Midjourney is no longer reliant on Discord. Gawd I FUC$ING hate Discord...I refuse to use Discord for anything...good riddance!
Many people are just as excited as you are :)
First?
Yep! Chicken dinner heading your way!!
@@curiousrefuge I'm preheating the deep fryer now!! LETS GOOOOO!
Hey Tim!
"...best achieved with cameras and actors." ... actors mostly stand in front of greenscreen now. Camera is just a camera. If AI can master lip-synching and create savable characters, it's over Johnny - IT'S OVER! lol learn to code, Hollywood.
Fair point! This will be very disruptive to the whole industry!
sora walling off the beta like that except for a bunch of hollywood guys, was a cheap move. the space is slowly migrating from indie film tools to hollywood only, just as many predicted.
If the studios buy SORA to keep it out of Indie hands, it won't stop another company from cloning it eventually. Is a studio like Disney ready to waste billions? They have before, so maybe.
It will be interesting to see how it all plays out over the next few months!
Barefoot - tetanus?????
Don't step on a nail!
I wonder how much they paid these "artists" for them to make this shitty ass movies, i bet it was a pain but the money was pretty
It's very telling how they said that nothing tops cameras and actors and VFX. Oh really? Good on them for protecting their jobs, but... the writing is on the Wall.
For T2? This was a community driven project that wasn't paid.
This video is just one big ad disguised as news with AI bullshit in the title. The reality is, AI is just a toy. Despite all the blah blah blah, there is not a single instance where the quality gets even close to being usable. Even voice cloning and TTS, which have been in development for literal decades, still can't really fool people and are full of problems. I don't think AI can classify even as a gimmick anymore given how horribly useless everything these companies produce is. It's merely a toy, with very limited professional use, and that's ignoring the blatant problems with IP theft companies like OpenAI depend on.
AI is literally in most productions released from Hollywood now and has been used in audio production and editing for years. Similar to good VFX, good AI work is indistinguishable from reality.
Voice cloning is so good that many Netflix shows currently use them but we don’t realize it.
You need to cope right now. Sora AI is far more advanced than any other aI tool and it will wipe out majority of film industry
@@VaderandRodaregoatedfr I'm a software developer you twat, not a filmmaker, in case the huge UNIX logo didn't tip you off. My assertion comes from my long time in the field. I have LLMs running locally on my machine, alongside various other types of AI systems. Sora AI will wipe out the industry as much as ChatGPT wiped out writers and book authors. That is to say, more than likely completely irrelevant as more than a gimmick.
can i just watch the freaking videos without somebody interpreting what the f im watching
Hi there! We're sorry you may not have enjoyed the video. We want to let people know why these tools are important to them which is why we give context. If you'd like to see more clips directly you can access the at their website (for example, OpenAI has the direct clips for Sora videos).
Hey your video was fantastic, we really enjoyed it! 😊We've recently started creating our own AI Films. It would mean the world to us if you took a moment to check them out.🙃
Thank you! Will do!
@@curiousrefuge Thanks so much! We're excited to share something special soon: an AI-made Dune parody. Stay tuned, hope you'll enjoy it!