Welcome to the podcast.
I'm your host, Jaden Schaefer.
Luma AI has just raised $900 million in a funding round led by the Saudi AI firm Humane.
In addition to that, Luma has released a new AI model that lets users generate a video from a start and an end frame.
I want to talk about some of the insane use cases and what we can see with AI and video on the show today.
But before we do that, this episode is sponsored by Delve.
I'll see you next time.
I'll see you next time.
All right, let's get into what's going on with Luma.
So recently they raised $900 million at the end of November.
This was led by the Saudi AI firm Humane.
And there's something interesting going on here, because the two companies beyond just getting this investment, they're also partnering to build a two gigawatt AI supercluster, which they're calling Project Halo.
And they're building that in Saudi Arabia.
So you can imagine there's a lot of kind of um, funds and work and partnerships happening inside of Saudi Arabia, which I think is smart on their part.
With all of this, Luma AI is now valued at over $4 billion, which is an incredible valuation.
So they kind of had this big funding round.
And along with this big funding round, they rolled out a brand new feature, which is really cool, which essentially is letting users generate a video from a start frame and an end frame.
And what that looks like is.
You essentially can give it one picture, which would be the end frame of one particular clip, and you give it another clip and it basically makes a video that merges those two clips together.
And so what this is really powerful for I've been watching a whole bunch of tutorials on YouTube is for transitions.
One guy in particular showed a really cool tutorial where he was making kind of like a travel vlog.
Now, I'll give like this YouTuber's kind of examples as examples.
But imagine this is essentially replacing VFX.
This is a massive industry in Hollywood and film creation.
And this model is incredible.
So I'll give you some of the incredible ones.
But I just want to give you from like a practical idea so you understand what it does.
There's one YouTuber.
He's like up in a hotel room and he like points out the window and he's like let's go like take a tour of the city down below, and then all of a sudden the camera like just zooms out the window, drops down to street level, zooms forward, and all of a sudden he walks on the frame from the side and starts like talking, and so anyways, it just felt like this like you know a 50000 camera shot where you would have to have, like either a really fancy drone maneuver or some really crazy VFX or this like incredible thing.
And all of a sudden he's like great, let's now go chat about what's happening at that building over there.
And the whole camera just zooms all the way over to that building.
And then he's standing up in it and he starts talking.
All of a sudden he points in another direction.
The camera starts spinning around him and does a 360 and then kind of zooms off down the street that he's going to give you a tour of.
Okay.
The reason why this was so incredible to me, and maybe some people listening are like okay wow, that just sounds like what you'd see on like a normal TV show or movie.
These are all VFX and special effects that are usually only reserved for big budgets and big Hollywood films.
Films and sets and cameras, or just a ton of effort.
All of this can be done in your, you know, in your computer, on your computer, in your office.
They don't, you don't need a big budget.
Anyone can do them and they look incredibly professional and they're really exciting.
So One thing that Luma was showing off, when they were kind of showing off the capabilities that they had with all of these effects, they were showing how you can have, you know, one person.
Let's say you're riding a horse or pretending to ride a horse down the down the street.
You can use their use, their tool to literally transform you in that motion, in that clip, into you, actually on a horse.
So, like you can act something out and it can turn into a real horse which like, as you can imagine, it's not just like you're AI generating, you know, like a whole film.
You could still have like actors and people acted out, and you have those same people, those same actors.
You just got all the VFX.
Like it is really cool.
The clip they showed was just a random guy, like kind of like putting to gallop a horse down the road.
And I don't know, California or some random street.
And all of a sudden he's like in the wild West on a horse.
It's the same person, different costume, different setting.
It's amazing.
Another one that was really interesting for me was in the e-commerce space.
You can have it replace products.
And so like yes, we could do this with ChaiGBT image generation, but to be able to do this with video is really impressive.
So they'll have it where like, one person is holding a random can of you know like a protein powder, and they just had to instantly switch to like five different brands of protein powder.
The reason why I think this is cool is because if you, you know, need to create an ad for some sort of product and maybe you don't have the physical product with you right there, you can use an image of what the product would look like.
Then you can kind of act out the whole ad holding something different and have it swap it in.
So I think this is really cool.
Another thing that you can do is to restyle things.
So they had a sample where it was like these two people just like play fighting with swords.
And then they had it turn it into, you know a whole movie where it's those same two people dress up as knights with armor and a castle with real swords, actually fighting
You can do a lot of other really cool things where you take a picture of a specific person and you have it change them into a completely different character.
You can, like I mentioned with the start and end frame.
So just to give you some context on how that actually works, they have one where it's like a picture of a ice hand and a picture of one where it's a fire hand.
And they just say, start with the ice hand, end with the fire hand.
And it's like this whole animation where the hand that's ice all of a sudden lights on fire and turns into like this fiery fist.
So what's really cool is it builds all the pieces between the two clips.
And so, like the example I was mentioning, where there was a guy talking in, like the apartment building, and then all of a sudden it zooms in and he's on the street,
He just recorded two videos, one of him talking in the apartment building, one of them talking on the street, and he had it fill in the clip.
And also you can explain what you want it to do, right?
So it's like zoom out this window, zoom down to street level.
Keep a smooth motion like you're still giving it text and explain what it's going to do.
But it will fill in all of those pieces and do exactly what you need.
And it looks absolutely amazing.
They had another one that was just giving me ideas, where it's like a video clip of, like a football player walking into a stadium, but it's just an empty stadium.
And then they, you know, have the AI go and it's the exact same clip, exact same person, exact same.
Everything's the same, except the stadium's now full of cheering fans, right?
And so you can imagine for commercials, you can imagine for movies, you can imagine for products.
This is incredible where like, you can get a rough draft of exactly what you want and you can have it go.
Make it so much better.
They also had another one where it was like girl pretending to cry and then they made it like so she actually had tears coming down her eyes, which is really interesting.
They also have it where you can.
You can completely change someone's persona.
So you can, I could be the actor in a movie and I could do everything, like all the acting that I want, and then we just upload a picture of a different person and all of a sudden I become that person, which is kind of cool if you think about.
You know, one person could technically go make an entire movie they could actually the roles of all the different people and just have the AI change them into all of the different roles, and I think that my prediction is that we'll see a whole bunch of really cool indie films which are just one person doing all of the doing all of the roles in the whole film.
Of course, they had a bunch of funny ones, like a girl running up and you know, pretending to punch the you know, punch the air and then they turned it into her being the Hulk and she was smashing through a pillar.
They had a car commercial concept.
That was kind of cool where it was a car you know driving at nighttime and then all of a sudden it's driving through a the snow, and it's driving through the fall and the leaves and through a sunset.
And what's interesting is it's all one clip like so one very cohesive clip, and it just switches through all the scenes.
You know, something we've all seen in like car commercials before.
And this, you know, in the past would have required them to go to all of these different places.
Line the car up perfectly on the camera, have a car next to the car, driving like.
I have a friend that shoots these kind of productions and it's pretty intense.
They have this like camera rigged up with all of this insane equipment on it.
There's people in the car that are manning the equipment.
There's people in the car that are driving the car.
There's someone in the other car that's driving the car.
There's so many pieces to this and now it can all be um, you know, done with ai and it can.
It can shoot all of this stuff, which i think is really really cool, really exciting.
Of course, you can turn a real person into a cartoon.
They have it where you know.
There's a guy walking through the desert and they change him walking through the snow.
They have it where there's a guy just like laying on a on a box with a cape, pretending to fly through the air.
It looks super cheesy.
And then all of a sudden they turn it into, he's actually Superman actually flying through the air and make it look really realistic.
A person in a grocery store, they all of a sudden turn her into Shrek.
You're flying through the air and all of a sudden there's like buildings rising around you.
So There's all of these really exciting, I think, use cases that I'm really excited for.
I think we're going to start officially.
My prediction for 2026 is we're going to start seeing single person produced movies.
We're going to be able to see incredible indie films come out, and a lot of people are going to become their own movie producer, which I'm really excited about.
Here's a quote from Emmett Jane, who's a co-founder and CEO of Luma Labs.
He said, generative video models are incredibly expressive, but also hard to control.
Today we're excited to introduce Ray 3 Modify that blends the real world with the expressivity of AI, while giving full control to creatives.
This means creative teams can capture performances with a camera and then modify images them immediately to be in any location imaginable, change costumes or even go back and reshoot the scene with AI, without creating the physical shoot.
So I think this is going to save companies a lot of money.
I'm excited to see what they're able to do.
They have a new platform called Dream Machine and they're hoping to compete with companies like Runway and Kling, who have some of these capabilities which have been out for a while.
So I'm excited to see what happens.
They've obviously received a ton of money.
They have investors like A16Z from the past, Amplify, Matrix Partners.
And of course, this latest round is led by Saudi Arabia's Public Investment Fund.
Well, I guess it's their public investment fund-owned company, which is called Humane.
So Saudi Arabia gives the money to Humane, and Humane makes the investments.
Thank you so much for tuning into the podcast today.
If you learned anything new, make sure to leave a rating and review on the show.
And, as I mentioned, make sure to go check out the sponsor of today's shows Delvecom, to help with all your compliance needs.
