Moonlake: Causal World Models should be Multimodal, Interactive, and Efficient — with Chris Manning and Fan-yun Sun

Peve Lin Jajan, and then she ended up working with Ron and Chris Manning here. And then so I got connected through to Chris initially, actually, through my cofounder. What is Moon Lake?

with, actually, NVIDIA research during my PhD years on essentially generating interactive worlds to train reinforcement learning agents or embodied AI agents. And then there's two observations, one in academia and one in industry. An industry, like folks at NVIDIA are actually paying a lot of dollars to purchase these types of interactive worlds, whether it's for the sake of evaluation or training the robots or policies or models.

And then in academia, same thing is happening. And more specifically, when I was actually working with NVIDIA on the synthetic data foundation model training project, we were actually generating a lot of synthetic data and showing that, hey, you can actually these synthetic data are actually as useful as real world data when it comes to multimodal pre training. But then, like I said, there's a lot of dollars being paid out to, like, external vendors or or, like, other folks to manually curate these types of data.

It was very clear to us that, okay, on our way to let's call it embodied general intelligence, models need to learn the consequences behind their actions, which means that they need interactive data. And the demand for those types of data are growing exponentially, but everybody's sort of thinking about it from a pure, say, video generation perspective or something else. But we feel like the the true, actually, opportunity is actually building reasoning models that can do these things, like how humans do these things today.

So that's a little bit on the genesis of Moon Lake. And I think the reason I got into world models was partly a philosophical take of the on the world where I, like, you know, blame the simulation theory and stuff like that. But on the other on the other hand, it's really just like, oh, like, there's an opportunity there that I feel like nobody's doing it that way I think should be done.

I can say a little bit about that. Yeah. So the overall goal is the pursuit of artificial intelligence.

Yeah. Actually, I wanted to add on a little bit Yeah. On our previous point, which is just like Chase, stop it super quick.

I I do feel like sometimes people confuse, like, oh, like, we're taking an up a method with with abstraction. That means they don't believe in bitter lesson. Like like, that's just false.

Right? Like, we are believed as a bitter lesson, but then I feel like the question that we always discuss is, like, what is the right abstraction level today? The analogy I like to make is, like, let's just say we can encode and decode represent all of images, videos, audio in bytes.

Then the most bitter lesson approach is to train a next byte prediction model as opposed to the next token prediction model, where it's like, it's natively multi modal because you just but it's like, well, yeah, like, to to Chris' point, it's like the scale and computing need to achieve that. So that's why we always come back to, like, okay, what is the most efficient way to do it? And and reasoning models, to to the point of this blog post, is a showcase of, like, hey, we're actually just, like, reasoning about the world and reasoning about the aspects of the world that that matter for me to learn what I want to learn from this role model.

Yeah.

Yeah. So this this really just walks us through the reasoning traces of, like, okay. So let's just say if we wanna build a world in this context, it's really just a game demo that that shows the variety of interactions that this world model can build.

And, yeah, it's really just a reasoning traces of, like, okay, you're prompted to create a bowling game. Like, how did it achieve what you saw, that level of causality, interaction, and consistency? Right?

That's the benefit of having a reasoning model. Right? Like, because you can you can say, oh, like, maybe in this particular context, I want to learn how to bowl.

And then you can say, okay, then what is it important when it comes to learning how to bowl? Okay. Maybe it's like, I need to understand the the basic of, like, physics, and I wanna throw it over them.

I wanna know that when I when it resets, it's it's a new game. So I know that yeah. Basically, you know you know you know to pick up the ball.

You know the ball's gonna cause the pins to fall down. You know that what's important to this particular bowling game is to score, and you know that the score corresponds to the number of pins that fell down. So it's just like, if it's a model that sort of knows what it looks like, knows what a bowling game looks like, but doesn't actually allows you to practice over and over again and to understand that, oh, like, what it takes to actually get a high score, then it sort of doesn't actually allow you to learn what you set out to learn within the world model.

Right?

So there is an underlying physics engine. Yeah. In that sense, there's some overlapping things to Unity.

But the way we think about it is, like, physics engine or tools or code are cognitive tools, like, borrowing Chris' term. Right? Like, tools that the model can employ as means to an end.

So today, maybe you say, okay, in this particular context, we care about physics. We care about the long term causality consequences. Then, yes, we deploy employ physics engine.

And then maybe tomorrow, we say, okay, we're we're training, let's just say, drones, where we only care about really fluid dynamics and the visual aspect of the world, then then, yeah, maybe we don't actually The model actually doesn't have to use a physics engine, or maybe it employs other types of representation or physics engine to achieve the task.

thing. But in fact, we can already do multiplayers. Oh, yeah?

Okay. I haven't seen any demo sessions. Actually just, like, prompt our our models to say, hey.

a data constraint, like, the more data, the the better this reasoning model can almost basically act as humans to, like, operate a variety of tools and softwares to build whatever is necessary. And then there's a sort of fidelity constraint, which we're actually solving with another model, Reverie, which we can talk about later. But it's like, well, it's not as easy to get to photorealism with the approach that we're taking.

But we think there are better solutions to that, which is we can dive into later later. The one one thing you note here is it's a diffusion model. Right?

So there's there's a few approaches, diffusion, caution, splatting.

Yeah. Totally. So within our world modeling framework, we think there are two models that we train.

Right? Like, there's the multimodal reasoning model that we just talked about that essentially handles mainly the the causality, the persistency, and logic determinism determinism of the world. And then Reverie is our bet on saying, okay.

Like, while all those model can take care of all these things that we just talked about, its limitations compared to existing, say, video models is that it doesn't have as high of a pixel fidelity right off the gate. Right? And reverie is to say, hey, we can actually take whatever persistent representation that we generate with our multimodal reasoning model and learn to restyle it into photorealistic styles or arbitrary styles you want.

So this model is almost to say, hey, I'm going to respect the persistency and interactivity of the world that you created, but my only job is to make sure that its pixel distribution is close to what we want.

We actually believe that this is gonna be the next paradigm of rendering. So it's gonna replace how raster rasterizers. It's gonna replace DLSS today because it not only has these pixel prior that's learned from the world such that you can literally play any game in photorealistic styles, which is a lot of people's desire when they do GTA.

Right? Like And all the mods, all the people adding perfect lighting and all this. So skins for worlds, let's call it.

Skins. Let's call it skins for worlds. You can call it skin.

You can call it customization. You can play it how you want. Right?

Yeah. Exactly. And I think another thing that we really pointed out specifically in this blog is the programmability of it.

Right? So what this means is that this renderer well, historically, renderer is always a derivative of the game state. Right?

You're saying, here's the game state I'm rendering out of frame. But here, I'm saying, actually, this renderer can be part of the gameplay loop. I can say something about the lines of if upon getting 10 apples, I'm gonna my weapon of choice, my bullet's gonna turn to apples.

And that's that's possible because we can say we can basically dynamically have certain game state trigger the the preconditions to the renderer such that the rendering is now part of the game loop too. One thing is to just say, okay, it's it's it's the appearance. But the second thing is also to say, there's these novel interactions that are possible because this renderer now has actually priors of the world.

And it's up to the artist to figure out what to do with it. It is up to the creators. Yes.

Yeah. And I also think that's actually another big argument that we're making and the reason that we're picking back taking the bet we're baking is that a lot of the times, whether it's for Embody AI or gaming, like, you want a layer where human can inject their intentions. Right?

So, for example, let's just say, in the context of gaming, it's obviously, like, my creative intent. But maybe in the context of embodied AI, it's like, oh, like, I take this foundational policy, and I want to actually fine tune it to deploy in my house. So you want to almost say, inject have a layer where human can say, oh, here's the distribution of things I want to create to achieve my goal.

And I think three d graphics as it as it is today is basically the layer for people to say, hey, what do I care about in this world?

Now the stance that we take on this front is, like, we're not gonna be more creative than our users. Ship it out. Yeah.

But we wanna make sure that we're building things in a way that really allows them to express their intent.

Yeah. I think this is, like, a great question and probably one of the hardest questions in world models, because, like, I think it always comes back to what are you building this world model for? And depending on your end goal and purpose, the evaluation should differ.

So in the context of games, then the most direct way of measuring is how much time are people actually spending in this worlds that you create. And if your goal is to say, example, in the context that we just talked about, like, hey, deploying deploying action in body agent, then your your end metric is then, okay, after training in these worlds that you generate, how robust it is to when you actually deploy to the target environment. But then, you know, it's it's hard to measure these end metrics.

So today, people have, like, these proxy metrics that I call that basically try to measure what we really care about, which is the end metrics. But then, frankly, it's different for every use case.

But it's a lot easier to change with code as opposed to a model that is learned primarily on data of real world and virtual worlds that are I guess, like, example, Genie, where, like, there's actually training a lot of real world data and a lot of virtual gaming data. And it's hard to say well, maybe it's easy to say, okay, wanna change the visuals in, like, the time period of of the world.

I mean, are there other examples you have in mind where existing models you know, like, it would be easier to do that's not using code?

and that seems like your approach. Right? Yeah.

The point I'm trying to make is that there are very many, many different types of world simulators. And, like, having a world simulator that can produce pixel coherency is very, very useful for games and, you know, marketing and all these things, but it's not as useful as people think when it comes to causal reasoning, when it comes to embodied AI. And, yeah, like, it this this title is true.

Like, we're not saying that it's it's, you know, not a great world simulator, but actually, in the blog that we we we we wrote, the bet is more so that there are gonna be disproportionately large share of value of real world task and virtual tasks where high resolution pixel fidelity is not needed. And, yes, video models have their values. Yeah.

Yeah. No. I mean, I I think the the interesting or more technical question here actually is where do you draw the boundary between what's handled with, let's say, diffusion prior and where when what's handled with symbolic priors?

Yes. And Okay. Okay.

Because I go there. This boundary can actually be fluid. Like, I think, like, maybe what you're trying to get at is like, okay, people are saying, pixel prior everything.

But what we're saying is, okay, there's a boundary that we draw where this is where we think provides the most economical value for the domains and things that we care about today. And I actually do think and it's just something that we do internally all the time, which is like, okay, given new equations that we learn or new elements of the world and that we we learn, or maybe some other knowledge that we acquire in the process of developing the models. Should we still be maintaining this line exactly as it is today, or should we move it a little bit left or a little bit right?

Right?

the reverie model Yes. Actually, we have a few iterations of them. They're actually at slightly different I know.

Values. You should do that. That's a cool dimension to show.

Yeah. Is quantum mechanics the diffusion prior of our world?

I don't know I don't know if, Chris, you wanna say it, but I think I think, generally, I feel like physics is better with symbolic priors.

That that will

and we're very hopeful about that. Yeah. Yeah.

Yeah. I mean, we are we are very focused actually on commercialization in the sense that, like, we do we do really believe in the data flywheel app approach Yeah. Where we put this in the hands of the creators and the users, and then they will teach us when what capability our model should improve, and that's why we are we are actually, you know, like, products in beta.

Yeah. Focusing on gaming. What what's, like, the adjacent thing to gaming?

Embody adjacent, basically. So maybe we can we can I'll I'll maybe start with where we see the platform in three years, which is like, okay. The users would tell us what they want to achieve.

The end goal could be, hey. I just I wanna make something to teach my kids the value of humility. Or it could be, hey.

I wanna fine tune my drones to be really good at rescue situations. I could be vacuum robots. I wanna, like, train my manipulation or, like, vacuum robot to be very robust to my office.

Right? But it's like whatever it is The robust to my office. Like, very robustly with in my office.

But then it's like whatever end goal that you want, our world model will say, okay, given what you want to achieve, let me generate a distribution of environments such that I can train and evaluate whatever it is you you want. Yeah. Right?

Maybe for the purpose of games, it's just the end simulation and that's the end product. For certain policies, it's like, I can train it within these environments and then help you see where your policy is failing or not.

both. Right? Sure.

Same same thing. Yeah. I think it's just this role model that allows people to train any policy that can act in any multimodal environments.

No. Honestly, there there are so many actually applications in adjacent markets that our role model can go into. Yeah.

But, yeah, it's it's sort of fun to riff riff on, although on the execution side, we sort of we we need to stay focused with, like, okay, what are the capabilities we wanna unlock over time, and there's a roadmap for that. But, yeah, we're we're just ripping on sort of, like, the possibilities. I feel like whether it's endless.

Yeah. It's like Classic. And then The embedding for a possibility and endless in my mind is very close.

Yeah. I think this point, like, the it's sort of a more more pointing to the benefits of using an game engine as a tool that's available to the model. Right?

Because, like, part of the spatial audio is from the code that is underlying the simulation. And while we do give our model access to other types of audio models as tools, none of them would be spatial, I think. Right.

But that's exactly sort of more point to, we're giving our model an abstraction or a suite of tools such that it's able to achieve that. And you can argue that sort of spatial is like a like a emergence out of the the tools that we and abstraction that we provide to the agents. And I think that's the beauty of this this this approach is, like, there's a lot of things kinda like how humanity's built technology, and they're like Lego blocks that built on top of each other.

And it's the same thing here.

We do want to basically, like, we our moral model, like, the one we're training is basically towards the goal of having a combined latent representation across all these different modalities. Right, such that it can, like, reason across these different modalities. So for example, if I close my eyes and I you play a video you play a sound of, like, cars skidding away from me, I almost can, like, visually extrapolate that trajectory in my mind.

I think that that type of capability, we want our model to be able to reason. Right? And that's the reason that we're sort of taking this multimodal reasoning approach.

It's like we want this combined latent space that can Yeah. Oh, you said latent space, and we like that here. We have to play the the bell every time that someone says latent space.

No. You gotta train daredevil one where you you you it's only audio, but you have to work out where everything is.

Yeah. On the model side, we we are actually striving towards basically a self improving system, but what that means is that we need people to set up the self improving system. So more more specifically, people who have the intersection of knowledge within cogeneration and computer vision and graphics.

Right? Yeah. That's that's sort of the core research background that we look for within our team, and and the majority of the team today do have, like, both backgrounds.

Got it. Yeah. Alright.

So I think If you've written a game engine, you're come talk to us. Right? Oh, yeah.

Definitely. But I do think that the line is blurred, like, increasingly blurred these days, where it's like, if you have a general understanding of vision and graphics.

then then maybe we we care a little bit more about having graphics knowledge.

in your world. Ah, I see. Yeah.

In that case, if you yeah. Definitely, if you've written a game engine before, if you've RL ed a variety of coding models on different objectives, like Easy.

it's done. I think, yeah, there's there's a there's a versions of that that are done. But I I think we are aligning audio, text, language, and video.

Yeah. Yeah. Yeah.

And, basically, we have these role models that are able to act as agents to, like, act in these worlds and extract long horizon videos and encoding that back to the models to sort of self improve. So it's an insanely exciting but also technically challenged problem. So people who wanna do their lives best work, you know, one makes a place.

How big are you guys? Where are guys based? We're currently based in San Mateo, although we're moving up to SF.

Very cool graphics and design, by the way. Actually, at the at the time when the when the when we started the company, we were thinking a lot about how do we make a company name that gives people the vibe of, like, OpenAI, but for, like, almost like industrial light and magic vibes. Wow.

Because it's like we care about creativity and using that as a funnel to solve AGI. So then we were we brainstormed a lot around, like, DreamWorks. Right?

Like, industrial light and magic. And so there's a few few basically space of things that we feel like are very, very semantically close to the company's identity. Yeah.

And then it ended up being Moon Lake partly because of the DreamWorks vibe, you know, the DreamWorks Moon Lake. Exactly. Yep.

So that was a little bit of that inspiration. And then the moon was sort of like a it basically was, like, about the reflection. The reflection part also implies the self improvement loop.

Wow. We sort of, like That's what really believed in, and that's the path towards multimodal general intelligence.

Yeah. And Walt Disney is also, like, one of my favorite founders. He's like his his story like, at the time, you're like, okay.

I'm gonna create this, like, immersive park. Like, people can't can't don't even have that technology to create it virtually, but, like, you know what? Let's just build it such that people can So he's the first world modeler?

No. I I yeah. I'll tell people that.

Like, theme parks are world models too. Yeah.