I Was Wrong About AI Video...
Вставка
- Опубліковано 11 чер 2024
- In this video I talk about all of the latest advancements in the world of AI Video Generation.
Discover More From Me:
🛠️ Explore hundreds of AI Tools: futuretools.io/
📰 Weekly Newsletter: www.futuretools.io/newsletter
😊 Discord Community: futuretools.io/discord
🐤 Follow me on Twitter: / mreflow
🧵 Follow me on Instagram: / mr.eflow
🐺 My personal blog: mattwolfe.com/ (Currently Being Overhauled)
Resources From Today's Video:
AI Predictions 2024: • 2024: The Year of AI (...
MReflow Tweet: / 1749871366281810013
HBCoop Tweet: / 1762901045087203695
Pika Labs Tweet: / 1762507225455604165
EMO Project: humanaigc.github.io/emote-por...
Morph Studio AI Clips: techcrunch.com/2024/02/28/mor...
LTX Studio Twitter: / ltxstudio
T2 Remake Twitter: / t2remake
Anamorph Film Tech: techcrunch.com/2024/02/28/ana...
Sponsorship/Media Inquiries: tally.so/r/nrBVlp
Mailing Address: 3755 Avocado Blvd Unit 287, La Mesa, CA 91941
#AIVideo #AITools #ArtificialIntelligence
Time Stamps:
0:00 Intro
0:12 My Prediction is Wrong
1:54 Sora
2:12 Pika Lip Sync
3:17 EMO Generative Expressive Portraits
5:47 Morph Studio and Stability
6:46 LTX Studio
8:48 T2 Remake With AI
10:12 Anamorph Unique Films
12:23 Final Thoughts - Наука та технологія
This is the kind of thing that makes me frustrated with so many people sleeping on AI. It's not just that there are so many capabilities/applications, it's the insane pace at which those abilities are improved/polished.
Even if people do know whats coming,,what tge hell they gonna do?
Burn companies to let chinas ai take over the world?
By the time you learn, really learn, to use, say, Sora, or how many commas to use in a string, Sora 1.5 will be out. Once you release your first project 2, etc
Yeah, all the people that think AI is just another fad and won't really go anywhere are really going to be blindsided.
How can you sleep on something who's barrier to entry is to the floor? And continues to get lower.
It's good nobody knows about it, means you can capitalise on it
I'm waiting for AI to turn books into Full Length Feature Films, to recreate famous Naval Battles and historical events, it's going to be a crazy wild ride.
Yes! There are several books I've read that I think would be great as a movie or show but I knew wouldn't ever be popular enough to get properly made. Now it can be a reality. Crazy to think about.
you will live in a horror movie LOL@@SignumEternis
Or to remaster tv shows or rewrite whole bad seasons and canceled tv shows
It's coming. Maybe 2 years, not 20.
There's no limit to the fun to be had with AI. I've already fed image generation images of my old self, and made pictures of young me doing all kinds of crazy things, sometimes with other people who I never met, or who weren't the same age at the same time.
Thank you for not including the word "shocking" in the title. I have almost started to skip videos with the click-bait titles, because I find nothing shocking and most don't show any evidence that the entity mentioned in the title has actually been shocked.
Sounds like you’re soft and are too sensitive to be on the internet. Log off.
Shockingly accurate
Yeah seriously, I’m sick of the Clickbait
"You won't believe this new shocking thing that will change everything forever and is never seen before", these days I'm shocked only if see a title with some sort of informative description.
Yeah I skip those instantly
EMO paired with Sora is almost all the way there. And I say that as someone who has won awards for my work at major film festivals. There is going to be a lot of schlock made, but when real artists who understand film get hold of these tools, watch out. Note: I fully support the making of schlock! I want everyone to feel the joy of creating something, of seeing their vision - however amateurish or clumsy - come to life. This is why generative art is an almost absolute good. Also: Brian Eno is a genius and a hero.
Right. The idea that this will be catastrophic for art/artists is flawed imo. It's going to be revolutionary for sure, but we will continue to seek art, and we will continue to rate art, and have preferences for certain art. (this creates a selective pressure, which drives progress)
Maybe it just raises the bar, maybe it makes people focus on different aspects of the art object.
Maybe if it becomes really easy to create AI generated video scenes, the camera angles, certain effects, the story development, or even aspects we don't even really know about right now, become more important.
I'd like to make the example of digital/electronic music.
100 years ago nobody knew about something like "sound design" you had acoustic instruments and that was that.
Nowadays, the sound of a guitar or a string section is infinitely variable and morph-able through digital manipulation, or the absolute base tone of a sinewave is used, a sort of "fundamental element" we couldn't work with before. Clearly. This has become a major aspect of music. How well Mixed is your track? What quality does the sounds have. Listen to any electronic music and try to view it from the perspective of Mozart. It's basically impossible to comprehend.
Maybe we find aspects in film to play around with that we don't even recognize exist today.
The ability to manipulate any aspect of a moving picture is amazing. The visual effects, the genres, the story telling. It will Change everything.
Think of any limitation of real recorded film. It's all completely of the past, just like the many aspects of classical Orchestral play are of the past.
This will do to Film like what Synthesizers did to music.
It's gonna be a blast.
Instead of ruining film, this will increase its variability, the number of genres and aspects to be judged a thousand fold.
Can you imagine?
No actors, no directors, NO FILMING, no restrictions, no costs. This is incredible.
I agree, I'm in the VFX industry. For a while I did freak out like a lot of people but have come to realise that these tools will empower artists to create the content of their dreams. As a visual effects artist it was very time consuming even to make a little short movie by yourself and so as a vfx artist we have to fit in a team where everyone works on individual elements for a shot. I chose to be a matte painter but when I started I did learn a bit of everything. Now i'm gradually incorporating AI image, image to video and voice to voice generation to help speed of the whole process for creating little short films. I'm working on my first animated short at the moment it still takes a fair bit of effort even with the use of AI. I only use it if its up to a professional level standard so some shots are still done the old fashioned way. In the end though its the story that matters and how watchable the movie is. The best and most innovative creations will be down to those who can put it all together in the right way and come up with some unique.
Thank you. I was telling a friend the same thing yesterday...it's like using mid journey...lots of trash will be created but some really outstanding work will be revealed and hopefully go viral. People who have very creative minds but no access before this will shock the world...Hollywood movies have become excessively boring and predictable. I welcome this.
It takes away the pressure to appeal to the masses. Now ideas can be realised that would otherwise never have seen the light of day. It's a great time for real artists.
I imagine that is the future of every show. You will no longer watch NCIS on TV. You will simply press play and a never before seen personal version will generate for you. Everything will be created on massive data sets to be unique. Combine that with personal preferences and I don't know what this world will look like. And this is technology that is here. NOW. It's crazy.
Great job man. Seen pretty much all your YT output and love it. Don't be shy of the sponsorship. You are clearly genuine.
We’re in what Kurzweil called ‘the knee curve of exponential growth’.
Exactly!
Technically the knee curve changes depending on the scaling you put in for X and Y. The difference is that our own idea of fast progress and slow progress is kinda fixed, and that is the region we are crossing now. Mind you, it won’t stay an exponential curve forever. I just hope it keeps going for some time before it begins to slow down.
You start with image recognition. Then you reverse the process (you generate random noise until what you have gets recognized as what you were prompted to output) and you have image generation. Then you apply some kind of logic with keyframes to that and you have video generation. At no point does the generator need to be conscious, intelligent or creative or understand what it is outputting. This is a knee curve of asymptotic growth. It's leveling out. They're just applying data science and it gives you an appearance of creativity, thinking or consciousness. This hype is like considering it a miracle that a library had a book on your preferred subject.
Knees are cool.
@@GlaciusTS At the time this curve starts to plateau, a paradigm shift will have kicked in starting a new A.I. curve. That's what I put my money on. We're at escape velocity regarding intelligence, right now. It's truly impossible to say where we will be in 5, 10 or even 50 years.
I don't think people as a whole really want for their movies to be random each time they watch, to where it makes a difference. People have always bonded over the shared experience and being able to talk about it with others.
I want both. I want to tell my AI streaming service what I want a movie about, it makes it, I enjoy it, and then it can be saved/shared with others if I want. If we didn't get the latter, I'd still be happy with the former - beats scrolling through Prime and Netflix crap for 40 minutes to settle for something like I do now.
Well, what happens with AI is that you refine results. This is where the human judgment comes in. "Okay, I like this video, but the character should be slightly younger, it would be better if the villain dies in this way, please make the sun brighter in this scene, etc."
A chain of several people could make a movie incrementally better until it tops anything Hollywood can do.
This is a great point.
I mean if your thinking of humanity as a whole movies have not been around that long. Cavemen telling stories around a camp fire were probably told different everytime.
Also think of Star Treks holodecks where you can just enjoy a randomly generated story with your friends.
This is getting bonkers! 😱
Thanks a lot for keeping us up to date! 😘
Exponential growth is scary. :)
@@NScherdinYES! VERY!!!
I've found you really need to take current estimates for ai development timelines and divide them by 2. Then you need to divide it by 2 again to get the actual estimate.
exponencial grow, hell yeah
So divide by 4? 😂
@@gonzalezm244divide by 2 again. then put a usb cord in your forearm and enter the matrix
@@JankJank-om1opBro stop playing around, Elon is going to release the USB capability later this year.
@@JankJank-om1op😂
I have been following AI for a while now and am very bullish on it, but I am SHOCKED by the EMO Generative Expressive Portraits. Seriously, SHOCKED! Just those previews alone are incredible, I had no idea we were this close to that level.
I'd love to talk to a historical figure using this tech. Imagine an AI that's been trained on the life of an historical figure and we pair it with EMO. If it could run fast enough to have a conversation it could revolutionize education.
Incredible.
Were you STUNNED as well? Perhaps the entire industry was surprised? lol. But yeah, those really are incredible. It's really striking how lifelike they are. Once all this different tech is combined it's gonna insane what can be created.
I mean to be fair the demo ones are going to be cherry picked to an extent. But I agree, a big step up!
It’s amazing. There is no going back anymore, the results are too impressive. Not just the results in the MACA field either…
@@greenockscatman Watch the Joker one. That is what got to me. It is riveting to look at all the micro-expressions going on; a genuine acting performance. WOW WOW WOW
Hopefully the AI telling the "history" has trained on the correct source material?
I could see studios using that dynamicslly generated movie approach you described to A/B test different sequences or versions of a movie to eventually arrive at the "definitive version" - I think that would be really cool!
I still find it so ironic how fervently the music copyright is protected while the copyright for the art used for learning models isn't.
Visual art isn’t controlled and owned by mega corporations.
Music is often owned by companies. Most art online are just users posting their work onto public websites. Before the AI art controversy, it was the norm for the past 15 years that anyone could just scrape users' stuff online for analysis, research and other uses- as long as the websites that host the user's stuff allows it.
@@chrism1503 As much as I understand this both are "copyrighted" even though, it seems, copyright of one this isn't synonymous with other copyrighted material and particularly how it is enforced.
@@flyingstapler1241 I guess in all fairness, the art sites people uploaded to for the part decade couldn't forsee the rise of AI and how scraping would be a thing.Thanks for the insight!
I've become so much more bullish on AI over the last year, to the point that I wouldn't be surprised to see anything anymore. Advancements are happening on a moment to moment basis now. What a wild time to be alive.
Extremely wild times! It's so hard to plan your future nowadays because everything is so unpredictable..
@@pictzone it's incredibly exciting, even if some of that is genuine fear of the unknown.
All in on NVDA!
I still can't quite comprehend why it's advancing so quickly. Is compute really pacing at this clip as well?
@@ohhhhhcool the best way to really think about it is through the concept of anything that "compounds". Each movement doubles the previous, so 1 and 1 is 2, then 2 and 2 is 4, then it's 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, etc., etc. See how quickly it expands? We are at a ridiculously high point now where every movement jumps beyond the previous point to a new, almost incomprehensible, new high point. It can be hard to comprehend, but everything is now moving at an incredible pace. It's very understandable why it's hard to perceive. It is constantly catching me off guard, even though I think I get it... I don't, lol.
The moment true AGI is fully created, ASI (Artificial SUPER Intelligence) will probably happen in basically the blink of an eye, and that's the thing no human actually understands. We are talking about an Intelligence that is truly unfathomable to any human. It's this unknown that has me excited and apprehensive at the same time. That's why I say it's truly a wild time to be alive. Maybe it takes 50 years, but it could be 5 years, or maybe we're actually right on the cusp right now?
Yeah, compute power could hold it all back? That's really what Sam Altman was getting at with the $7 trillion number people are talking about. What he was really saying is that he believes it will take that level of commitment, across the board, for all sectors to come together and be truly viable for the advancements that will truly change the world. He was talking: chip tech, infrastructure, etc., everything, for it to all come together.
The thing is one knew this was gonna happen (with ai), but the speed at which this is happening is really amazing and mind blowing. One also knew that ai would grow and learn and evolve at very very fast pace, but still seeing it actually happen and happen so fast actually, is really mind blowing.
Can't wait for your video on the LTX Studio Tool and other video tools, that type of stuff will change everything. As someone in the AV and Videography business it's going to rock that world. Things like training videos, corporate presentations, product announcements and advertising will all get easier to produce. There will be a reduced need for camera operators, sound and lighting techs, video editors etc. Why have a trainer or corporate executive sit and read a teleprompter when you can input a picture and a small clip of their voice and generate it all by AI. Wow!
Always learning something here!
Great Stuff & CONGRATULATIONS MATT On yer NEW Sponsor🎉😊🎉
Eventually you’ll put on a VR headset, you’ll choose a location and the characters, who mind you will have their own large language models, and you’ll interact with them as if they’re really there. It’ll be processed in real time and it’ll look so realistic, you’ll forget they’re not really there. Of course that might be some time away, but the societal impacts will be massive. If you think people are disconnected now because of social media, wait until real time lifelike language and video generation in real time happens. Wait until people are able to completely disconnect from this reality that we’re all experiencing now. At the quality where the viewer is able to forget they’re in a simulation is only a few years away.
yea before that happens the entirety of the working class will be unemployed thanks to agi and the economy will collapse, so I wouldn't look forward to anything unless you're a literal billionaire
@@user-jq6ro5rt6s where this shit is going to end :0(
That desk lamp behind you is killer! Where did you get it?
All I need is 100% consistent characters (animated). Then I can stitch all the clips together into scenes, like a regular show would be done anyway.
Brian Eno is most known for being a musician! It said that in the article, you just started mid-sentence. The movie 'Clue'm from the 1980's, had four different endings so people would have different experiences and have to watch it multiple times to see all the endings. But when the VHS tape came out it had all 4.
yeah, pretty sloppy wrt to a legend like Brian Eno.
As impressive as the last tool sounds regards variations in the film individuals see. Does anyone think this sounds like the beginnings of thought control? Content tailored to the individual? I'm interested in AI and it's capabilities but am concerned it's potentially more nefarious possibilities might sneak up on us through the guise of entertainment.
The largest untapped potential for AR AI is having the Ai change small bits of your viewing space as you walk. Like auto translating all text /signs; removing people from beautiful landscapes (like a beach scene) or augmenting what you see like making the outershell of things translucent so you can see the insides of machines
Wow, great update - it is bold to make predictions in this environment but I appreciate that you do it 🙂
Im interested in using AI to create scenes with a product. Like a window brand. I will upload reference images but the only issue is it always generating something too abstract. You have any insight on this?
Hey Mat...what is the best tool for uploading a source image to animate in video?
There are the smooth arty things like this and then there is the bleeding edge tech demo of what can be achieved which will come out next. Not polished but a mind blowing pure performance demo.
When LCM came out 4 months ago I was perhaps the first to do a RealTime camera -> SD LCM -> video at 15 fps in my first iteration. I've doubled the speed since then. I coined the term RTSD back then.
On the non-video side I've recently hit 200 images per second with sd-turbo 1 step at 512x512 using multiple performance optimizations.
All of this is on my 4090 on a home PC.
Thank you Matt, very excited about LTX and look forward to seeing your videos about them.
OMFG!! That EMO project is absolutely NUTS! Soooo impressive. Jaw dropping! I can only imagine what is not in public domain
I'm really into the effects you use. It's cool to see how you use AI to make your videos better, and the quality keeps getting better with the tech.
@1:41 how do you have access to this? From this channel getting you famous ish?
I’m going to get a new computer but don’t know much about components and whatnot. Could some of you help point me to what I’ll need in order to run these Ai programs natively? I understand this isn’t available to the public yet, but what should I buy to make sure that when it or something similar is, I’ll be able to use it?
The anamorph thing will probably be best for RPG video games, particularly if you can carry outcomes across sequels
Imagine how much storage we're going to need if people can generate shows and movies and then want to share them with all their friends. The amount of storage we're going to need is just ungodly.
and the amount nature has to suffer for that. dystopia greeting...
Invest in data storage companies.
Nah, if you just store the prompt, the RNG "seed", the model used, and other input values that was used to generate the content, you can just re-generate the exact same show/movie/whatever on demand by using those same exact input values again. No reason to store the final output. This is how video game randomizers work.
I think that won't be a problem in the future, like AI will be able to recreate things on the spot. Like if ai makes a movie, there will also be a code/seed which you can give it, so it'll be able to create the exact same thing again. No need to store or send huge files anywhere.
I don't think a lot of storage space will be needed. In fact, I think, this will lead to incredible forms of data compression.
Ha! This is exactly why I've stopped making time based predictions about the future of AI! Always happy to speculate on things that will happen, but I'll rarely give a date! I'm always wrong, and it is always faster. From here on out, I just say "That'll happen sometime between tomorrow and a later date!"
I can’t wait to make alternative endings/sequels to my favorite movies.
😂 Matt, you keep teasing us on the behind-the-scenes look….👀 Can’t wait
Mind = Blown. Are any of these suitable for entering an entire written story? Like if I wanted to upload a short story and have it transformed into a video? (as opposed to a film script)
Absolutely love those singing ones. It takes me back to a UA-cam channel idea I had back in the day. 🙂
Not too surprised with the music videos. Lots of source material to train on(The Voice, American Idol etc....). :)
The problem is the rate of progression. I'm not learning a tool, only for a better tool to come out a week later after I've got a grasp of it.
I made an AI app back in nov 22 and its no longer needed. Was replaced within a few months.
Thats why nothing is really taking off.
I definitely think movie studios will start using generative AI video to replace more expensive action shots very soon. Way easier too
I think there will be tech to make 30min videos this year, but I doubt it will be available to general public at reasonable price. My knowledge could be outdated but my recent information is that openai didn't disclose much details about the time and processing power was necessary to generate their demo clips. To create a story that's coherent over 30 minutes or more, would require the model to go back and forth to correct errors in the story and the visials. This multiplied by additional 30 times (30 min vs 1) will make such feature very expensive for average users, unless new optimisation techniques or new inference chips appear on the markets, we will only enjoy the new demos, and maybe some paid movies.
Mr. Wolfe. Could I make a suggestion that in your videos, you put in the corner the current tech you are talking about? I find myself watching something your are demoing, I then realize it's pretty cool and would like to try it, but I didn't catch the name of it from the beginning. So for each little section, if you had in the corner "LTX Studio" or whatever, it would make your videos easier to follow.
You should see the chapters titles (which reference the tech he's talking about) in the bottom left corner when you tap the screen.
Hey Matt! Love your channel man. Thanks for doing this. I’m a musician, is there a AI platform out there, or on the way, that will let us upload multi tracks, like ProTools multi tracks, or something of ou original music, and it will generate a Fully mixed Song out of it? And let users be able to tweak the mix, maybe it could be done in different time periods, like give us a mix from the 60s, give us a mix from the 70s, or a modern mix, etc. We should be able to pick and choose. Is there something out there anywhere?for musicians, that would be the holy Grail. Thanks.
Perspective movies. Where you see the movie from the perspective of the different factions in a movie setting.
Thank you for all the info, Matt!!
Imagine a chose your own adventure movie where you seamlessly direct the movie to the narrative that you envision for it in real time.
5:47 lol "flims", like in that Simpsons episode
I predict the first (or nearly) use of the capability to generate 'infinite' video variations will be in advertising so that each time an advert runs it will be different and therefore more attention grabbing and therefore more effective.
No, what's really impressive at 3:58 is that the earrings move.
EMO and LipSync are really mind-blowing. I keep getting distracted by the small screen that keeps dancing or shows your face in the background 😅 But it’s fun.
Awesome stuff! Does anybody know how long the scenes from LTX can be?
I want to see someone make an AI version of the Circle 7 Toy Story 3 script and storyboards.
is any of these ai available?
same shirt from your prediction video? it's like you went back into the future!
I said already last year with model scope that we will have the first ai movie end of 2024 because you could also see that it took just two years for diffusion models to get to production quality.
In the mid-90's I ran some "graphics" classes at the local "Film Institute" for creative types wanting to learn about what they might be able to get out of a computer for TV or Film... for a handful of students I had to show them how to handle the "mouse" first. 30 years later...
I remember "machinema" being a thing where people were using things like the original Quake engine to try to make films. We have come a long, long way. Sora in a few years will look terribly primitive.
This stuff is moving so fast. Matt you're really helping the masses stay educated but also helping us make money and create new businesses with these tools. Very appreciative of your efforts
5:50 - DID YOU SEE IT?
Lover of crossword puzzles notices TechCrunch article headline has a typo. Misspelled 'FILMS' as 'FLIMS.
crazy crazy developments, in just one year!! What's more coming!!!
WOOOHOO , so proud to have taken part in Our T2 remake 50 artists or awesomeness
I said 10 years for a full 90 minute movie or whatever in 8K
Welcome to the Curve! (Exponential, that is)
It would be so cool if we could have something like a "holodeck". We put on the VR headset and then speak a prompt, and then the AI system creates very realistic environments and situations complete with characters. I think this will happen within 5-10 years.
LTX studio is a game changer. On the wait list already hoping to get access
LTX looks pretty damn awesome. Looking forward to that content. Cheers Matt.
There will be a point, where instead of watching an existing movie, we will simply write the kind of movie we want ("An action movie on Mars, the hero is a robotic Elephant. Humour. 90 mn long). and the movie will be generated just for us. Everybody will get movies and series taylor-made to each customer ! O_O
Or adaptation of existing books will be done in a few seconds (Generate a 8 episode season adapting book 4 of Reacher), and BAM ! O_O
Think of all the comic books or manga adaption that will become possible ! ("generate a live-version of the X_Men trade HoX/Pox").
Yeah
Oh, and I wish these videos would include the information I want the most : can I play with it right now? :) And if so, where?!?
Keen for LTX studios!
Amazing stuff!
I guess this year, we'll be able to watch a whole episode of Will Smith Eating Spaghetti. And I don't just mean that troll clip that he made where he tries to emulate an AI effect.
That last one with Anamorph makes me think about all the youtube Star Wars fanfiction I see that start with "What if...". Maybe one day they'll be using this tech to help bring their story ideas to life. Someone already used AI art last year to help illustrate their reimagined version of the Sequels.
Yup. We're living the moment of filmmaking.
7:30 takes the fun away 😮
That terminator not the first AI generated movie though. I've made an 80 minute one called The Eternal Recurrence from an original script, and there's two others made before mine.
Great to hear from you brother I wish to become like you but different language
No one cares.
Hey Matt, just FYI, I create the 1st feature AI movie last year and got it into actual movie filmfest not AI film festivals. We don't put it on UA-cam because of some of the restrictions for some film festivals and I didn't get permission from Stars Wars, my movie is called Dj Gregu it's over a hour long all AI footage, and music. Last year another animation my team did the old-school way got 19 film awards. So it's good to see folks catching up. Hit me if you want to know more
I work as a software dev at an animation studio. Yesterday, our CEO sent out a message saying he's going to host a panel to answer everyone's questions and concerns about AI video generation, what kinds of AI tools we will and won't be using at our studio, etc. Whether it's disruptive or not, there is absolutely a shift happening in the film industry right now from GenAI tools.
Yeah I can't wait to make music videos with this
anamorph sounds cool, the idea of every showing being unique is novel.
So we can take tv shows and change their styles or upscale them, remake cartoons as live action, have ai adjust the audio to match the lag to make sure it never gets out of sync. Maybe redo entire shows with different actors or voices or characters. So essentially ai will alter media so that the most popular version will be what is shared not the version released by the media creator. Fans of a piece of work will win in the end as anything not made for fans will just be altered to please fans.
I'll bring this back up for any future arguments we have
Bring what back up? For arguments on what?
You could have everyone watching and listening to a different movie using directional audio, and strobing glasses (like 3d)
i just was really shocked with your title but once i saw the video, now i am good
With AI and so many average people can be creative to create content, the challege will be how to find the gem needles in the massive haystack of content. Aggresive promotion will become vital if you have great content.
This is wild, i do see fan based IRL Manga Adaption coming soon...imagine the movie then makes more money, then the manga ever has made...
hope there is a middle way for both type of artist to make it.
So where do i go to use it? You didn't tell us that.
Damn, I can't wait for Sora AI... or at least LTX Studio, I've already registered on a waitlist
Great video, Matt.
Can't wait the day when I'll be able to input AI with source material (book, scripts, visual references, voice actors and music composer samples) and it will spit me out the movie/series I wanna watch:)
Something about EMO fells like there was some human feedback somewhere in the creation of the 'emotional' renders.
I remember watching of shows, and thinking "maybe something different will happen this time!"
Oh, how I miss my twenties.
Where we can download and use EMO?
Again, government will have to decide which stuff should be available for public use and which not.
Ps : Like, i dont understand how stock market will last.
I am just waiting for video models to generate full movies in what I enjoy watching with me being the director
Please talk about shinefy!
When I clicked on the vid I was genuinely hoping it was about the timeline being longer, not shorter.
"Stop the world I want to get off..." is going to become a very common sentiment over the next few years. For your own sanity, I'd suggest learning to love the breakneck velocity of innovation. Or you could try moving to a cabin in the woods, for however long that can protect you.
The big question I have is when these video generators move beyond the demo phase and into the "affordable for anyone to use" phase?
Pretty sure Sora will just be part of your GPTPlus subscription. Though you may only get a few generations per day or something as it is very compute-intensive apparently. Something like EMO is a research project but everyone and his dog is going to try to figure out how they did it.
Thanks for this 😊
Usually predictions keep getting pushed back but with AI it keeps getting pushed sooner
🎯 Key Takeaways for quick navigation:
00:39 *📈 AI video quality and length have significantly improved in 2024, contrary to previous predictions.*
02:15 *🗣️ P Labs introduces lipsync feature, allowing AI-generated videos to synchronize with uploaded audio files.*
03:23 *🎨 Alibaba's Emo research showcases AI-generated expressive portrait videos with audio, pushing boundaries of realism.*
06:41 *🎬 Stability partners with More Studios to create an AI filmmaking platform, offering a storyboard visual drag-and-drop builder.*
07:09 *🎥 LTX Studio by L Trix enables AI-generated movies from prompts, including storyboard creation and character consistency.*
08:59 *🤖 AI animators are remaking Terminator 2 entirely using AI-generated video clips, showcasing advancements in AI animation technology.*
10:21 *🔄 Anamorph's generative technology can reorder scenes to create unlimited versions of a film, offering unique viewing experiences for each audience.*