Meta Announces the 'World's Best' AI Video Generator. No Joke.

In this episode, I dive into Meta's groundbreaking AI video generation model, 'Movie Gen,' which has set a new standard in the industry. The model boasts impressive features like producing high-quality 1080p videos with natural movement, sound integration, and the ability to transform images into videos. I explore various examples, including a girl running on a beach and a woman in a pumpkin patch, showcasing the model's impressive capabilities. Additionally, I discuss Meta's Llama 3.2 update, highlighting its enhanced image recognition and seamless integration with various platforms. Despite its current research phase, Meta's AI developments indicate a promising future, and I emphasize the community's urge for open-source access. Join me as I explore these exciting advancements and share my thoughts on their industry impact.

▼ Link(s) From Today’s Video:

-------------------------------------------------

▼ Extra Links of Interest:

Let's work together!

Thanks for watching Matt Video Productions! I make all sorts of videos here on Youtube! Technology, Tutorials, and Reviews! Enjoy Your stay here, and subscribe!

All Suggestions, Thoughts And Comments Are Greatly Appreciated… Because I Actually Read Them.

00:00 Introduction: The AI Video Generation Revolution
00:39 Meta's Groundbreaking MovieGen Model
01:50 Exploring MovieGen's Features and Examples
04:51 Meta's Llama 3.2: A New Era in AI Models
07:57 Advanced Video Editing Capabilities
15:15 Community Reactions and Future Prospects
18:16 Conclusion and Final Thoughts

[MattVidPro]

BTW Literally just timing that the Ad Sponsor for this video was for Llama 3.2 - (my thoughts on Movie Gen are my own)

[AndyPerezCaba]

Congrats on a meta sponsorship! Thats crazy 🎉

[SkyForbes-.]

What was that OpenAi Sora thing again? I can't remember...

[ElishaBradley]

the video with the lightning it wasn't supposed to be realistic thunder sound btw. the prompt asked for orchestral sound which it gave the orchestral bass drums in time with the lightning was brilliant, more brilliant i think then you realize, very interesting that the ai did that.

[JUDY-zwdk]

The fact how far have we come in text to video, is crazy. I was just rewatching your video called "Create Videos Instantly with Meta's Text to Video AI! - Text to Video Generator". And there was one of the first video generators from THE SAME Meta Company. In that video you were saying: "What a long way we passed in ai comparing to a start of 2022" you also were saying that videos gonna be a lot cooler in future, and you were ABSOLUTELY right! (Ironically it were posted somewhere in end of September 2022) Now i can say how much have come in those 2 years! And there will be a fully generated movies, somewhere in coming years! Knowing how fast Ai developing, it can be even faster. So i have to say, thank you for sharing this amazing stuff with us.🔥🍋

[sneedtube]

EDIT: Nah they're straight up streaming an actual parallel universe

This is definitely a world simulator case closed

[TheIllusionRecords]

ok yeah, independent artists are gonna have insane music videos at the tips of their fingers

[nikfpv]

This solves a lot of issues people have with RunwayML. Meta took those notes and nailed it. Well done.

[MrJaggy]

"Make a Scene" was never released, this one might not ever be released too...

[notnotandrew]

Their largest model is 30B parameters? Those 1080p videos we saw can be created on your laptop? That's awesome! I'm not even into generated video but I really hope they don't disappoint everyone by refusing to open-source this.

[wtflolomg]

Not released? No better than SORA.... EXCEPT, they did release the research. This is far more important, because right now, there are half a dozen serious projects that are knocking on OpenAI's door, pushing video generation about as far as it can go with the current tech. They will be pouring all over those papers and implementing improvements immediately. If it involved new training, that will happen. This is a very heated race.

[SardiPax]

The biggest issue with the girl flying a kite video is the kite physics. The kite is facing ahead of the girl and the string connects to the front, it is also in front of her so presumably the wind is behind her, yet the directing the kite is facing is the wrong way around for that. Either the kite should be behind her (closer to the camera) or the front of the kite should face toward her/the camera.

[xitcix]

If this gets open sourced, things are going to get absolutely crazy. Something like this could make super accurate porn of anyone from a single image really easily, it would blow up the news. It would be really amazing if it were open source, but that's gonna cause a massive amount of panic and increased hatred for AI.

[levonkenney]

you should upload another video on suno ai there has been a whole lot of updates since your last video

[planetmuskvlog]

Another SORA we’ll never get our hands on?

[mohamedzewail]

I'm really happy to see you sponsored by Meta, Matt. Congrats, dude! You deserve all the best.

[adamfroman]

The kite is upside down. That kite would crash immediately.

[I-Dophler]

It's been absolutely mind-blowing to witness so many AI breakthroughs in my lifetime! The progress we've seen, especially in areas like video generation, is nothing short of revolutionary. From being able to generate videos purely from text inputs to watching AI models that can edit and enhance existing clips with stunning precision, it's clear that we're entering a new era of content creation. Meta's new AI video generation model is another massive leap forward, setting a new standard for immersive content and pushing the boundaries of what's possible. The quality, detail, and fluidity in these AI-generated videos are just incredible – the kind of things we could only dream about not too long ago. And it's not just about creating videos from scratch, but also editing existing footage seamlessly, which opens up countless possibilities for creators everywhere.

What's even more exciting is how fast all of this is happening. Every week it seems like there's a new model, a new tool, or a new capability that makes you rethink what's possible with AI. It's an exciting time to be following these developments, and I can't wait to see where things go next. The future is looking brighter than ever for AI and its role in shaping the creative world. Also, well done on getting sponsored! Woo hoo!!!

[tonyanimations]

Do you plan on making a video on Scenario AI ?
They recently added a way to train custom Flux Lora's and to use them in Sketch to image. I would love it if you can make a video on it.

[TheIllusionRecords]

I actually like how the sound fx kind of turn into music