Dyna-2: General Purpose Physical Agents with Hour-Long Autonomy — Jason Ma, Dyna Robotics

0:00 All right, we're here on site with Jason Ma, the, uh, co-founder, chief scientist of DynaRobotics. Thanks for having us in your office. Of course. Thank you for coming. Yeah. It's, it's been a blast. We got to see... You'll see a lot of B-roll, but we got to see, uh, drinks being pulled, a lot of evals, robots with knives for the first time. [laughs] Uh, so under the hood of all this, you have a world action model. What is that?

0:22 [laughs] Yeah. So last month, we released our world action model called Dyna2. So it's the first world action model trained on more than one million hours of data. So going back to your question first, world action model is a model that's able to predict the future video, the future state of the world, so hence the name world model.

0:40 But in addition, you can also use that future generation to also produce actions robots can execute in the physical world to control the robot. So why do we do that? It's a very powerful architecture because if you think about how humans act in the physical world, in our brain, we're implicitly thinking about how the world will evolve based on our actions.

1:00 So now we have a model architecture that's able to do that to guide robot actions. So what we find is that by doing so, the model is much more dexterous, much more robust, and much more capable of generalizing to environments, to scenarios it's never seen before. Yeah, I th- Yeah. Uh, I wanna prompt this question a bit. I think a lot of people hear world model, and they think you're outputting video or world or 3D state. Um, you're also doing action.

1:27 I think most people don't understand what models are under the hood of robotics. So what is input, output? What's the training data? What is the architecture? Yeah. Give us all the details. [laughs] Yeah, yeah. So, uh, for inputs, the most important input is video, right? You know, you can think about videos that the robot is capturing with its cameras, right?

1:48 So video or images of the, you know, past many seconds, that's the input. But for robots, what's really interesting is that robot also has a lot of actuator, a lot of, uh, joint encoders on itself. So the robot also knows, you know, the angles of its joints. So those can also be inputs to this world model. That's something that's very different from just a pure video model that's only looking at videos, right?

2:10 So that's the input. And additionally, for robots, you want to be able to talk to it, to instruct it to do different tasks, right? So you can also feed it like language instruction. So those are what I consider the basic set of inputs that go into any robotics models, including world action models. And once we have that, the output, at the, uh, basic level, it needs to output actions the robots can execute on the hardware.

2:34 But in the case of world action model, what's really interesting is that in addition to the actions, you can also output future video, right? So basically, you can almost output the future video or some representation of it first, and then use that to guide robot actions. So the way I think about it is like, this is kinda analogous to, you know, chain of thought. You know? Basically, the robots, the models are reasoning in the visual, the physics space, and then outputting actions.

3:01 So that's the key difference between world action models outputs versus other kind of models, robot models outputs. Um, you put out a good blog post around your model. I think a lot of people understand language models or transformers, next token predictors. In this case, you shared quite a bit. There's like diffusion, flow matching. You're predicting next world state plus action.

3:23 Um, one thing you say is a million hours. How do you determine the architecture? Like, what are the choices there? What is the architecture? I feel like in language models, there's a lot of take something off the shelf- Yeah... take trillions of tokens, do prediction, SFT, RL, and you have- Mm-hmm... a model. What about this?

3:40 How do you go from nothing to your model? Is there vanilla architecture off the shelf? Do you determine all this? And what's, what is the current state? Yeah. I think, yeah. So the mental framework to think about how do we build models for robotics is that there are essentially a couple things to look at.

3:57 One is data, and the second is, like, how do we make a model learn as much as possible from that data? So let's first talk about the model actually, right? So the blog post we released is our Dyna2 world action model. So under the hood, it's a diffusion transformer. So transformer is essentially like the language model, you know, uh, uh, sequence-to-sequence match- Decoder only... right, yeah.

4:20 But what's really interesting is that in visual domains, right, if you think about generating images, generating videos, and in this case, robots looking at images to generate actions, people have found the diffusion kind of architecture to work really well. So what is a diffusion model? A diffusion model is a model that takes some random noise and just through iterative computation can go from noise to something that's really high dimensional, that's something really highly, like, defined and have very fine-grained feature.

4:51 If you think about generating images, right, you want sharp edges, sharp corners, color contrast. And for robotics, even though you're not generating image, you're generating actions that needs to be very precise, very, like, smooth, right? So it has kind of the same kind of, uh, consideration, the desideratas, generating image.

5:09 So that's why from first principle, you know, the whole field, not just us, have been using this kind of diffusion transformer architecture to train robotics foundation models. You need the transformer aspect because transformer is the thing that can just ingest the most amount of data, can learn. Then diffusion helps you to refine the outcome to get to the smooth, precise, fine-grained motion you need to actually execute on the robot. So that's the architecture that goes behind a lot of our recent breakthroughs.

5:37 And we, and we've covered flow matching before on the podcast. It is mostly an efficiency speed-up. Like, let's say you normally would take 40 to 100 diffusion steps. You can reduce it to four. Yeah, that's right. So with a, uh, diffusion model, one thing that's very hard in practice is that you need to tune the scheduler. A lot like by how much you're tuning or converting the initial noise input into the final output.

6:01 So flow matching is kind of like a special case of diffusion that makes the transformation much simpler. In practice, you don't have to tune as much, and it works just as well. That is why I think about two to three years ago, most generative modeling for image and video has shifted from A diffusion model to a full matching model, but the concept is the same, and full matching is just a special, simpler to tune, uh, variant of diffusion.

6:25 Yeah. It actually makes me think that, so full matching obviously has trade-offs in quality. Yeah. But actually is very suited for action diffusion and, and, and robotics because you get to reevaluate in real time. Yeah. Whereas in, in images, you're, it's basically one shot. That's right. Which is pretty fun. We got to mess with the laundry folding robot, like, you know- Yeah... throw a random towel in. Yeah.

6:47 Um, the models adapt, right? Yeah. I think one key thing about your models is that they're tr- they're made to be generalists, so one foundation model and you adapt them a lot. Yeah. Uh, how does that work? Basically, in 45 minutes of mapped data, you can generalize towards... We saw an example of, um, pulling drinks out of a fridge. Yeah.

7:07 What's the, what's the mindset there of one general model, and then how do you adapt to different tasks? Yeah. So the way we think about it is that, uh, for robotics, we need to build a very general, strong base model, right? And then because if you think about how a human interacts and behaves in the physical world, right?

7:28 We have very strong base knowledge. We can just readily do a lot of tasks. But if you want to become an expert at any, any given task, there is some amount of training that needs to happen. And I think to train general purpose robots like the one we're building here, the same principle has to apply. So our approach is to first train a general model that's training on as much data as we can find.

7:51 So that is when our Dyna 2 report, our model is trained over a million hours of human data as a starting point. And then given that strong foundation, what we found is we can quickly make the robot or make the model do many, many different tasks with a very minimal amount of fine-tuning to just tell the model, "Here's the task I want it to do, and here's very few amount of demonstrations," and then the model can readily do bunch of the tasks that you have seen in the office.

8:18 Yeah. Um, we actually have quite a few people who deal in data, who sell data. Yeah. Um, what is good data? What is bad data? What do you wanna buy? Uh- [laughs] Yeah... uh, I think these are all o- open questions. Like, obviously, I think the, the top, um, important thing is egocentric data of humans doing tasks. Uh, what else would you modify on that basic statement? Yeah, yeah.

8:42 So I think data is the most important thing in robotics today because unlike language model, there is no internet, right? So every company that's building model, building hardware has to think about data. So there's, uh, roughly couple of different ways of operating here. One is that you can collect a lot of data yourself, which is what we are doing as well. You have employees here doing that.

9:04 [laughs] I mean, the robots everywhere in the office, right? Uh, but, uh, if you think about how much data you need for robotic foundation model, I think the amount of data any individual company can collect by themselves is just nowhere near sufficient, right? It's kind of like, you know, you're trying to build an internet by just having Reddit or just having Facebook. Th- they're a great website, a lot of users, but it's not enough, right? So when we think about data, there has to be some sort of data acquisition happening, right? You buy data, you look at data that's already out there. And then there's many different tiers, right? Um, you mentioned data quality.

9:41 I think data quality in robotics is a non-stationary problem in the sense that what is high-quality data for your model changes over time. And when your model doesn't know anything, [laughs] like when it's zero, any data is useful, right? You can use a lot of video data. Uh, you know, in the case of egocentric data, I think it's actually a very good way of pre-training robots because human data is so diverse. Like it covers every single physical activity you can possibly imagine.

10:11 And that gives robot the model such a strong prior on how things, how the world can evolve around a embodied actor human. But the issue with human data is that before your robot looks exactly like a human, there's gonna be some gap between human actions and the robot actions, right? So just by training on human data, uh, today it's not very likely you can just execute robot actions directly. So you still need some amount of a- adaptation using robot's own data, right?

10:43 But the robot's own data collected on a robot is gonna be much more expensive and much harder to collect. So the kind of recipe we're thinking about is like, how do we take the largest source of data we have today, like egocentric data, to give the model a strong prior over everything physical manipulation can cover, and then combine that with a small amount of robotics data to make the robot, uh, you know, master the whatever task we're interested in.

11:10 So data quality and getting data in robotics is really, I think it's a fascinating multifaceted problem. But the reality is that there's not a single winning strategy like, "Oh, here's one kind of data. I'm only gonna collect on that," which will give you a strong foundation model. You really need to think about, uh, this human data covers this part of robot cognition I need.

11:34 Robot data covers low-level reflex dexterity the robot also needs, and put them into one model, and I think that's what gives a really adaptable, really strong model at the end. Yeah. I- I'll- I'll- I'll prompt you, uh, just for, again, for the people listening who want to work with you- Yeah... um, what, uh, what data vendors do you want... Uh, this is your call for data. What- Yeah.

11:56 Yeah, no, that's a great question. And, uh, yeah, so in our Dyna 2 blog at the end, we actually have a email address- Uh-huh... called datapartnerships@dyna.co. Okay. So if you're a data vendor that has a lot of high-quality, let it be egocentric data, uh, robot data, video data, or anything you think could help robots, you know, send it to that email.

12:14 Our team will take a look, and we'll get back to you. Yeah. But like a- a- at your stage, are you- You're, you're probably not buying any more egocentric data off the shelf, right? That you probably want to be like, "Here's what I need. You go-" Is there a niche that's lacking? It's like you need a scale AI, you need a McCore for, like, doing this. I see. The- It's like custom data...

12:32 people that don't work in robotics are probably thinking like, okay, some niche factory, like, you know- Uh-huh... you showed putting tubes in. There's only so much data- [laughs]... you'll find about that, but, like- Yeah... would a lot of that help you? [laughs] Yeah. No, I mean, if you really want to nail a task, having data from that task will certainly help you a lot, right?

12:50 So for us it's like we still think, you know, just a million hours of data is, is not enough. It's just the beginning. Like, we showed the first model on a million hour scale, but we need more than that, right? So we're still looking at a lot of just raw, diverse data, uh, that we can get, but we're also thinking about how do we get more custom data either collected in-house or by our partner for, like, really, you know, specific niche tasks because I think you still need that to get to, you know, excellence.

13:19 That, you know, any model needs to actually be useful in the physical world, right? It's kinda like humans. You know, uh, you know, if, uh, you know, Sean or me, you know, we go work at Rosewood Hotel tomorrow, we're not gonna be able to make the bed the way they want, right? We need a lot of training.

13:34 So that kind of data, like special data, that's kinda like what McCore is doing today for language model, I think will be very, very valuable for robotics as well. Yeah. So hotels, you know, strap a GoPro on your employees' heads and [laughs] And then [laughs] yeah, there's probably more va- There's, uh, you know... people put themselves in there, in there. Wait. Yeah.

13:52 Uh, so, so one thing that we discussed is, like, the, the closing this interview, uh, and, uh, what's very interesting is, like, now you can basically fabricate environments on the fly. Yes. Yeah. Uh, uh, have you tried that with the, the frontier models? Um, what's your experience? Are, are they good enough? Are you using any world model robotic- like world model companies to generate these and train on them?

14:13 Yeah. Yeah. So, uh, I think generating synthetic environments, sy- synthetic data, simulation data for robotics is going to be one of the next frontiers, right? 'Cause today people have looked at egocentric data, robot data. You know, those are scaling, but I think synthetic data will be a new dimension of scaling.

14:32 So going back to your question, like, yeah, we have been exploring, like, using frontier models to generate a synthetic, uh, environments for us on our data, on our environments that we care about. And surprisingly, I think now it's at a point where this technology is actually very much worth, like, going really, really deep into.

14:51 So for some context, so, uh, in my PhD I primarily focused on simulation-based training for robotics. Back then simulation wasn't ready, but we had this paper called Eureka, which was the first paper taking like a large language model, at the time it was GPT-4, which is how to, like, self-generate, uh, environment, generate reward functions or reinforcement learning and just, like, do things in simulation.

15:15 And at the end, I think, uh, we had this really nice demo of the robot spinning pencils, right? That was really cool at the time. But that kind of thing was still very, very limited at the time because back then the model didn't really have good vision capability, didn't have a innate 3D understanding, right?

15:32 So it was basically just generating code and, like, using very limited feedback from text to try to come up with new simulation environments. But today, like, the vision capability of these frontier models are crazy. They can do, like, real to sim, through CADing. There's 3D reconstruction, all of that. [laughs] So like now it's at a tipping point where this technology could really accelerate robotics.

15:55 And for us, right, we think about that, like, as like a way to accelerate our internal progress, right? Because we're building models, we're building hardware. So if a frontier model can give us more data that way, like generate really tailored data for ourselves, that's extremely useful for us. And another thing we've been thinking about is instead of using these models to generate, like, 3D environments, we have also used our own world model, right? The world action model, uh, I was telling you about.

16:23 Because the model not only generates action, can also generate videos. We- the model can also generate its own training data. So that becomes a, you know, self-fulfilling, like, flywheel that I think will also be powerful. So overall I'm just really excited about this direction because I think even though we have scaled to a million hours of data, we are deploying a bunch of robots.

16:42 The robots are act- you know, are already starting to do very cool things. We're just at the beginning, just the tip of iceberg for, like, what is there to come. Yeah. For those on the research side, we recently had Runway on the podcast. Yeah. One of their things is robotic world models. They're, um, shifting a little towards there, so I'll, you know- Yeah...

17:01 nudge people to look at it. Um, one of the questions I have about you making your own models is what is the constraint you operate in? So if you draw a parallel to language models, you see like GPT OSS20B is- Yeah... very fit for your average MacBook. Yeah. Then there's a larger one that's fit for a single GPU. Um, do you work in the constraints of robot, uh, model on- Mm-hmm...

17:25 on device, in cloud, a hybrid, and then how does that affect your scaling law of data? Yeah. Um, you know, a million hours is great- Yeah... but, like, do we just scale up parameters, data? Yeah, no, that's a great question. So I think what's really different about robots versus language model and other applications is that for, uh, coding agents, language model, it's okay if it's, it has latency, right?

17:50 Yeah. Like, like, you know, today, whenever I ask ChatGPT, it takes, like, couple of seconds before it spits out anything. But for robots, like, the inference latency has to be, like, very short, right? The robot needs to act. Like you can't just hold a knife and not do anything for three second. That's gonna freak people out, right? [laughs] Yeah, we saw someone put their real hand under a knife.

18:10 Our model's, uh, pretty smart, so that's okay. But, uh, in general, like if you think about very dynamic task, like for example, you want your robot to like, you know, hold a, I don't know, a tennis racket and, you know, juggle the ball, right? It's this very dynamic, very reactive. You can't just like... The phys- physics, gravity doesn't wait for you, right? So for that reason, robotics model needs to be able to inference very fast.

18:32 And the fastest way to do that is by serving on-prem, like on device, right? Because if you call a model- Network latency... but the, there's no guarantee- Self-driving car... there's no guarantee, right? Like, so for that reason, the model that gets deployed eventually has to be of reasonable size that can be served locally. But that doesn't mean that the initial model you train also has to be small, right?

18:56 Like in language model- Can distill You have a large- People do distillation, right? Yeah. Like the best way to make a small model very smart is by doing distillation. Right? You train the biggest model, then you distill. Right? So I think in robotics we're getting there, but I think that's what people are going to converge on. Yeah. How about like, uh, inference?

19:16 You have your own inference stack or... A lot of this is architecture. You've pulled diffusion and whatnot. I'm guessing you can pull off of something. Anything you wanna share? Yeah, so we build our own inference stack for whatever model we deploy, because I, like I was saying, like latency is so important, so owning the whole stack is extremely important for robotics, right?

19:36 You want to optimize, like the inference speed for the kind of model you're running, and there's a lot of work in there, so that's what we do. Yeah. You're really full stack, like- [laughs] Yeah... everything from like, uh, you know, the, the actuators all the way to the industrial design, ID. Yeah. Which I didn't really know as a acronym. Yeah. Well, I, I guess- Obviously to the models and data as well...

19:57 uh, I'm curious, like are these your own actuators or are they... How much of this is your own? Like of course the design, you know, it's not like humanoid, but- Yeah... actuators, like how, how far down the robotic stack? Yeah. So the way we think about, uh, what we need to own versus what we can buy is like if the, uh, input output and, uh, of a component is very, very clear, and like there's already extremely good mature solution, then there's no point of building house. Like for example, on this robot, uh, you see the lidars, right?

20:29 So those are black circle, those are lidars to help the robot detect obstacle, um, navigate surroundings. So that's a very mature technology that people have been using for now decades. It's on self-driving car. So we don't need to build lidar from scratch, right? Same goes with some cameras. But for anything that we think is extremely crucial for the whole system's performance at the end, our vision is to be able to own that and build that ourselves.

20:57 So one thing is, uh, you mentioned actuator, so that's something we're working on and we're going to own ourselves. The reason is that, like the quality and how you design the actuator really matters for how well you can model the robot, right? Actuator is what drives the robot. To be able to model your robot very well in simulation or to do any kind of modeling, you need to select actuator that's very good. That, you know, can actually be easily modelable.

21:23 So that's what we care about eventually, so we're gonna own that. Same thing with, you know, the robot's proportion. You know, this robot's proportion is like we measure against real people. Why do we do that? 'Cause we're pre-training on human data. Oh yeah. [laughs] Best transfer, right? So there's a lot of these kind of, uh, decision that goes into hardware that's very much tied to like...

21:44 Because eventually, you know, it's hardware that carries intelligence. Intelligence is the extremely important thing. We want to make sure the hardware serves for that. Yeah, yeah. Uh, I would just highlight, I, I don't know if this is relevant to you, but, uh, one of the most interesting robotics discussions I've had is someone who was very stressed out about actuators and specifically magnets.

22:03 'Cause apparently we don't make magnets in the US. No, we don't. And there's only two countries that make magnets. Uh, rare earth material. Yeah, right. The, I don't know what the, what, what's the chemical, what's the material, I don't know what the thing is, but- Yeah, yeah... uh, so yeah, if, if anyone's interested in that, that's, uh, apparently a, a bottleneck that we have. No, it's bottleneck. Not making it. Yes. So what do we do? [laughs] Yeah. I mean, I mean it's a good question, right?

22:27 I think, uh, there's a lot of, uh, exciting discussions on bringing manufacturing back to the US but, you know, for robots especially, that's actually much harder than, um, than what it seems. And I would say even now there isn't a company in the US that's truly building everything from scratch in the US, right? Like even when they say they assemble robot, they build whatever they're actually doing.

22:53 In the US, they're still sourcing the raw materials from other places in the world. And I think that's something to really be aware of as a robot builder, right? Like I think, uh, like as a full stack robotics company, we realize just how long the supply chain is and how long the whole problem is. It's actually extremely hard.

23:12 And I think these days a lot of the discussion centers around a robot foundation model, model data. But I think hardware and just how hard it is to build good hardware is very under emphasized. Yeah. But I think that's gonna be extremely important. This is a, a platform, right? Yeah. Actually- A platform. Yeah. There, there are some companies that choose to be, let's call it platform-less.

23:34 They only do the model, and they'll say we'll deploy on any platform, we'll deploy on Optimus. Off-the-shelf robotics. Uh- Among- And so, uh, it's a choice, right? Like obviously you, you, you own everything so you can do that. Uh, do you see yourself building more in the US or do you see building your own global supply chain? What, what, what's the direction that is likely? Yeah.

23:53 So we are currently exploring many different options, right? Because with global supply chain then there is some, you know, geopolitical things, right? Shipping. Shipping. Also shipping delays and all that stuff. Literal ships for robots. Yeah. And, uh, in the US, you know, it's just that going from zero to one will take some time, right? So we're trying to think about a solution that can both allow us to build more in the US while leveraging global supply chain, right?

24:20 But I do think, uh, going back to your previous question on like some companies choose to just build model, we build everything, I think, I think it's very important to get this right. I feel that for robotics, the hard thing is- Uh, like building the whole system, or the necessary things to build a whole system.

24:38 Like previously, I was at DeepMind, I was at Nvidia primarily just building models, and I saw the upper bound of the system when you don't control the hardware, when you don't get to optimize that, right? Because if you think about the value creation for robots, eventually robots are here to deliver a service to solve some physical tasks.

24:57 So if you wanna solve physical tasks, you wanna be able to optimize everything, and that's important. I mean, we have the robot. You gotta tell us about the design. Yeah. Um, your own custom platform. What do you call this? [laughs] What do we call it? Yeah, so it's not humanoid. Um, it's very smooth, but what's the, what's the rotating arm? Yeah. So we think of this as generally like a semi-humanoid.

25:20 So basically the upper body is, you know, uh, like a humanoid. It's got, you know, of course shoulder, got arms. It also got, like four degrees of freedom on the waist. So you can basically do all the upper body motion that a human can do. But it's a semi-humanoid because the lower body is wheels. All right, so why do we choose wheels? Because I think there is a difference.

25:40 There is a, you know, optimal, like practicality versus performance trade-off. I think with legs, the energy con- the energy consumption will be a lot higher. It'll be a lot more unstable. You know, safety will be extremely, extremely hard to guarantee. So right now, for legged robots, there's actually no guarantee that the robot will stay upright, balanced the entire time.

26:03 I'm sure you guys have seen a lot of those, uh, robot dancing videos, like, like the, like the robot just fail. Seeing dancing, yeah. It's interesting. They, they fall over very easily. No, but you also see the demos that you can tell are not production, where you can box the robot, and you can push it over, and it's very good, but that robot is just good at balance, right? Yeah. Like there's a... Yeah.

26:21 I, I think, I think the more interesting thing that you guys have innovated, so like we have, uh, one of the prototypes, uh, which is the Mu2. Yeah. [laughs] You can call it. Back here. This is, this is just like an up-down- Yeah... uh, axis, and you've chosen to then move towards something like this, which, um, has the, the ability to lean. Mm.

26:38 And we- we're gonna show, uh, the laundry demo where you have to lean into machines. Yeah. Right? That's another degree of freedom that I don't even have the words for it. It's not pitch. It's not yaw. Yeah, yeah. It's something else. Yeah. So yeah, h- how did we think about that, right? Like, uh, so actually it's very used for exercise.

26:53 Like, you know, I think sometimes, a couple years ago when I was exercising, I accidentally like, you know, like pulled a muscle. Mm. Right? I'm sure you guys have that experience too, right? Then the next few days, like you just couldn't move, or like you pulled it where you slept in the wrong position, so your neck is very stiff. We have all experienced that.

27:12 And then whenever you experience that, you kinda realize, wow, so many muscles, so many degrees of freedom that we use in our daily life that we take for granted, but when it's disabled, like it's very awkward for you to do some task that you feel are very natural, right? So that's also how we think about, like when we design robots. When you have like a, this kind of design, which is this purely elevator, could only go up and down, we realize a lot of the, like human torso movement.

27:38 Like we use our like, you know, waist back a lot for like day-to-day physical activity, right? Even like, if I wanna get this bottle, like I'm not just moving my arms. It's actually very strange for me if I, I have to intentionally think about fixing my upper body. I would always like do this, right? Mm.

27:54 So there's just a lot of like subconscious way that humans move their body that I think it would be equally useful if robot has a degree of freedom. And if you further think about, okay, we want to learn from human data. Like humans are moving their torso all the time. You wanna be able to, like ingest that into the model to control the robot.

28:14 But if you have a form factor that doesn't allow for that, then the transfer becomes less and the robot becomes less dexterous, and it's just more awkward. What's the data labeling side of this? Like I, you know, you've seen people, um... Is this the last like post-training, fine-tuning side of actually the hour of using the robot teleoperated or remote controlled mapping motion? Or like does some of it generalize over? Uh, when you say labeling, you mean? Like taking the model you have that predicts state- Uh, yeah... to mapping it to your specific hardware. Oh, yeah, yeah. It generalizes very well.

28:52 So for all the demos that I have shown you on the new robot, 'cause the new robot we, you know, it got ready not so long ago. So we quite literally do not have a lot of data compared to the stationary arms we showed you, which had much more data, right? But a new robot, the model has not seen this in pre-training, in some of our pre-trained model. It's able to learn the task that you saw, like very quickly.

29:17 And I credit a lot of that to like, you know, like even before the hardware's ready, we're already committed to pre-training on humans. Because we knew at some point our robots will be like human. You can get ahead on the training side while hardware is getting ready, and I think now we're seeing that, like the transfer is actually there and is actually very helpful. Talk about the eval side of this. Yeah.

29:37 So you can have predictions, you can have it do stuff and, you know, every now and then we see it miss stuff. You test on very niche things as well, so like filling a pill bottle. Yeah. Um, how long does that process take? How do you think about the aspect of actually deploying on a new task, evaling it, iterating? What does that look like? Yeah.

29:57 So if we wanna learn a new task, uh, you know, we basically do a little bit of robot data collection. Then the model, these- What is a little bit, like? Yeah, like a couple hours. Couple hours. Yeah. We're, in some of the tasks, like just tens of minutes, right? Yeah. And then- The chopping one was 45 minutes. Yeah.

30:13 The open the fridge was on- And that's an interview, that's like an onboarding task for one of your engineers. Yeah, yeah, yeah. [laughs] So, uh- Fun, fun thing if you wanna work here, you know. Yeah. [laughs] Find the task, have the robot do it now. Yeah, yeah. Pitch me a new task, and I'll come to the office, uh, you know, collect some data, train a model, and see what happens. Tie your shoes. [laughs] Yeah.

30:31 Surely it's tying, yeah. That'll be cool. But, uh, yeah, so you know- A little, a small amount of data collected, train a model, then we just deploy the model in the real world. And then we look at the performance, then- To, to fill in, collected means, like, you are, you have a person actually operating- Yeah, operating... remote control- Yes... doing the task for a bit. Yeah.

30:49 And then you train? Yeah. And it does okay, you improve. Yeah. How do you measure? Yeah, yeah, so when you're training the model, there's, uh, you know, certain metrics you can look at, right? Just like train your model. It converges nicely. Now, after that, for robotics today, we just run the model in the real world, right, on the, you know, same setup that the data is collected to see how the model does.

31:14 Sometimes the model just works great out of the box, right? You don't really need to do much more. But when the model's, you know, like maybe like 70%, 80% there, then you see how the model behaves. You figure out what it's not good at, then you can, like, collect a bit more data, then iterate, right? I think of this as, like, the robotics version of, like, reinforcement learning from human feedback, right?

31:36 The robot rolls out. You can score it, good, bad, s- uh, somewhat good. And then that data, the robot's only experience, can be used to train the r- model, and you can also just collect more data yourself to train a model, right? Eventually, you know, our goal is, like, you know, this robot is gonna be our platform. Anyone should be able to, uh, train tasks on it. I wanna get a bit of a prediction and, like, timeline to where that goes.

32:03 Right now, the tasks are, um, somewhat simple, like- Yeah... take laundry out. It's a towel, you can't crush it. Um, chop some food. Yeah. There's also the side of, like, okay, set up this filming studio, right? Yeah. Move stuff around, um, bring a table, set up lights. I think manage my data center. Yeah, yeah, plug in stuff. Um- And, and also, uh, re- like, uh, recharge another robot. Yeah, yeah. When, when do we see the factory?

32:31 Or, like, w- what do we need to scale up? Um, is it model? Is it everything? It seems like your hardware is set for that. But where do we see it go to a point where I can go from instead of just cut one thing to, you know, okay, here's a recipe- Yeah... like, full send or- Yeah. Yeah. Yeah. I would say to get to truly general purpose robots, uh, like, like basically a robot or a model that does anything. Like- Not, not, not, not that.

32:59 Like, okay, so to keep it somewhat grounded on a technical level, like- Yeah... when do we start to see us pushing that feedback, right? So- I see... instead of very simple chopping- Yeah... uh, maybe a five to 10-minute sequence, right? Yeah. Where now the eval step, it can break at many more points. Instead of just fill one pill bottle- Yeah... uh, 10 different pills all in a sequence, right? Yeah.

33:18 What's limiting us? What are we pushing on? Yeah. I think, yeah, I think couple of thing. One is that, like, uh, the laundry demo that, uh, you have seen, yeah, it's gonna be, like, one hour to finish a whole loop, right? So the longest duration that a robot can do, like, a full job is already in the hour range. Nice. Right?

33:36 But, you know, to be able to do that for any task, you know, I think the bottleneck is coming back to data and, uh, hardware, right? So for some of the tasks, you know, for example, setting up the space, you know, you might need more than, like, a, you know, the parallel jaw gripper, right? You might need something more dexterous, like a hand, like a human hand, right? But then if you wanna train a model on the hand, then you're bottlenecked by the hardware stability.

34:00 Like, hand today, the best hands, dexterous hands, are still not, uh, durable enough. They overheat pretty quickly. Right. So that becomes a bottleneck, right? Like, you, we, you know, you quite literally have to put a, a, you know, ice bag [laughs] to cool. We're not- Human... we're not just auto researching, make better model. We have, we have hardware limitations.

34:20 Yeah, so hardware limitation is a big thing. But then if hardware is limited, then that will slow down everything, right? Because to know that you're making progress on any task, you eventually have to eval on the hardware, right? So then without the eval being very, like, spontaneous, very fast, that also limits your research progress, right? Because then how do we collect the right data, right?

34:41 I think, uh, how do we collect the right data to do extremely complex, like setting up this, uh, studio, right? Because you can imagine, like today, most of the data is, like, vision-based, right? It's like video streams and, like, you know, some, uh, uh, encoding of your hand motion where the robot's, uh, joints. But humans have a lot more sensors. You know, we hear things. We feel things. Smell.

35:06 Yeah, you smell things, right? So we subconsciously use all of those to help us be so good at manipulating the physical world. But today, like, if we're not capturing that at all, then even when we scale to millions of hours, you might just be missing critical dimensions of the task, so you'll never be able to solve the task, right?

35:25 So I think those are, like, what I consider, like, you know, what's dragging out the timeline. But for tasks or for workflows that you can collect data today to high fidelity, then I don't ne- I don't necessarily think there is a fundamental bottleneck, right? Just like what we are demonstrating, uh, to the public this week, you know. We have a physical agent that can do the entire laundry workflow, right, as a human would, right?

35:50 Because for that task, we don't really have, like, oh, you actually need more than a parallel jaw grip, but you actually need some sensing modality that's just not available today, right? So I do think the biggest bottleneck to have a much more general thing is still hardware and data. Yeah. I wanna take this chance to talk- uh, let you brag about your customers and partners. Yeah.

36:10 Uh, you have some announced, uh, relationships. Um, we can start with Best Western and then maybe some of, uh, some of the others that you're, that you're proud of. Yeah. Cool stories, anything like that. Yeah. So we have a deployment blog post, uh, on our website that goes into much more detail. I really recommend watching the video, 'cause I think the video is- Which one? The Best Western one?

36:30 Uh, no, we have, like, a one video for all the customers we're working with. Okay. It's a really, really nice video. Shows what we're doing, and, uh, yeah, I think it's great. But yeah, we have customers, uh, you know, Best Western- That's in the hotel space. We fold a lot of, uh, laundry, like, uh, you know, towels for them.

36:47 Uh, every hotel needs it, every gym needs it, every fitness center needs it. But- What's the deployment so far? Uh, how... Like, where can I go to? Which best restaurant do I go to to see it? Yeah. So if you wanna go somewhere, I actually recommend going to our other deployment, Din Tai Fung. Have you guys been to- Yes. Yes. Yeah, you know the best- Very thin wrappers. Yeah, very- We should have filmed there.

37:06 [laughs] Yeah, we know, we should have. [laughs] Should have just filmed there. B roll. Yeah, but, uh, in Din Tai Fung, like, uh, if you go to Din Tai Fung or other like, uh, you know, somewhat upscale restaurant, there's napkins on your table, right? So our robots, our robots today fold thousands of napkins for, uh, many, many different locations, many different Din Tai Fung locations in the United States. So that's actually our biggest scaled customer, right?

37:32 So I wanna make a distinction here. Uh, today there's a lot of robotics companies that's talking about deployments, talking about pilots, but what's really hard is not putting a robot in one location or doing a pilot. What's really hard is that the customers believe or validate that your robot creates enough value that they wanna deploy your robots across many of their sites.

37:54 So that's what's happened for us with Din Tai Fung, which to the best of my knowledge, hasn't really happened for other companies doing similar like, uh, general purpose robots. So at Din Tai Fung, our robots fold thousands of napkins per day. It's... The quality standard is extremely high, you know, 'cause they want like, okay, uh, the second fold, where the seam of the napkin is only like, a couple centimeters below the, uh, I don't even know how you call it, the, the, the pocket so you can stick the chopsticks in.

38:22 So it's actually a very fine grain, high precision task. I can show you. At some level that's what robotics are good for, robots are good for too, right? Um, humans are not the most consistent- Yes... especially in small details, whereas a robot can somewhat do it perfectly. That- It's interesting that's the example you bring for general purpose robotics versus, uh, uh, at some level I think people would assume- Yeah...

38:45 you can have a specialized robot that will fold napkins perfectly to the exact, you know, millimeter as opposed to something like this, but... Yeah, I'm sure, uh, maybe you can, right? But the hard thing about m- manipulating deformable object, right, by napkin, towels, everything's deformable. There's just infin- infinitely many possible configuration.

39:05 So that's really hard for a specialized robot, right? And what we find is that by developing a general purpose model in a robot and adapting it to these use cases, you can actually get to a production quality faster. And what's really nice is that now that Din Tai Fung has scaled, has signed a much bigger contract with us, they not only want our robots to fold napkins, now they want our robots, you know, like cut, you know.

39:28 That's why we had the- I mean-... example, to cut food to the other- Is that actually the second use case? That's one of the use case they're thinking of. Oh, wow. I mean- Din Tai Fung needs to cut so many ve- Yeah... vegetables per day. They need to do the dumplings. Exactly. Do the fold. That's the final boss. That's, that's the final... [laughs] Uh- I think you can. I think you have the dexterity- Yeah...

39:44 and the precision to do this. Yeah. You have a little problem. But like, no, don't break the dumplings. Yeah, yeah. [laughs] I mean, that's, that's the obvious benefit is you can adapt them anymore. Exactly. I think the other thing you really touch on is the speed, right? Yeah. To make a custom napkin folding with the exact ply, whatever- Yeah... you have to make the machine. You can't just Amazon this. Yeah.

40:05 But this can do it pretty fast, right? Yeah, this can do it ex- very fast, and now it's actually al- it's at human parity. Yeah. You know? And this robot doesn't get tired, doesn't need to, uh, you know... You, you can... And I think I really like the fact that you mentioned consistency, right? Like, one issue with Din Tai Fung is like, you know, their LA location with New York location, Seattle location, they, they fold napkins slightly differently, but from their pers- perspective, they want consistency.

40:34 But you know, that's exactly what these models are good at, 'cause models are, you know, just, uh, it's a set of weights. You can just like put the weights in the different robots across different locations. You can OTA update them as well. Yeah, you can just update all of them at once. You know, it just... Once it works, I think it's truly like- Yeah... incredible.

40:52 Final boss is Cheesecake Factory, if you know the story there. Yeah. They're from scratch. Everything is very consistent down to the measured gram. Every sauce is made. [laughs] And like anywhere you go in the world, you'll have the exact same thing. Yeah. If you don't know, you don't know. Yeah. But you know, that's the, that's the final- I mean, the, the term that comes to mind is like, kind of like bitter lesson for robotics, right?

41:10 You, uh, instead of the specialized robots, which is a lot of, uh, industrial engineering, you build a generalized platform that can- Yeah... uh, do a last mile fine-tuning and, and- That's right... do all those things. Like give the expert add-on outside of your what is, what is more bitter, bitter lesson robotics? [laughs] Yeah, I think there's several bitter lessons for robotics. One is that robotics is going to scale just like any other domains that machine learning has solved, right? So what does that mean? That means have as much data as possible. But in robotics, the one thing that's a lot harder is like, what is the right data is much harder to figure out than other domains, right? I mean, language is text. Uh, in video games, just video game data.

41:51 But for robotics, like what sensors you choose determine what kind of data you have, right? But if you have the right s- set of sensors and just scale that data, I think you'll get predictable return on model performance. So in that way, robotics isn't so different. So that's the bitter lesson. As a corollary of that is that, like, there's a lot of research in robotics that's coming out with like very specific priors. Like for example, we need to make the robot see objects as they are. Draw some boxes.

42:22 That will give robot better perception, right? But I think the bitter lesson here is that just keep the algorithm simpler and scale the right kind of data, then generalization, better behavior will emerge. So that's not so different. And the second, I think that's the main bitter lesson, which is like, at the end of the day, scaling will work, but in robotics to scale and how to scale is just much harder, 'cause there is no internet.

42:46 You have to come out with your hardware, come out with the sensors, get all of those right, and then scale. Then I think, uh- I mean, there's the argument that, you know, the world is very made for humans- Mm-hmm... it should be a human, right? As hard of a hardware problem that is, think 10, 20, 100, 200 years down the line, can we make a really efficient human? But you have to be practical with what we have now, right? Yeah.

43:13 Yeah, so the way I think about it is like, yes, eventually the robot should... If you wanna build a general purpose robot, right, like a machine that can be adapted to as many use cases as possible, I think human form factor will be the best one, just because that's what we have been doing, uh, for ages, right?

43:30 But I think the best way to actually build a humanoid robot is to take a iterative approach. Because robotic is so hard, the, you know, the whole problem, the whole stack is so long, it's really hard to know a priori what's the right decision to make, right? And because hardware decision is much more expensive and much more costly, just 'cause hardware iteration is a lot slower, you actually want to take a incremental, iterative approach, not going all the way to the final form factor.

43:57 Because the chances that the thing you build actually doesn't last, it's not good enough, so you just end up scratching it anyway. You know, like, it's public news, like, you know, Optimus, you know, already gone through many, many iterations, and they had to scratch their designs, right? And I think by iterating with a semi-humanoid form factor and actually putting them out to real customers, or, and getting models really deployed, you get that feedback loop that you wouldn't get if you're just trying to build, like, you know, warehouse humanoid, and you spend like five, 10 years building it. I think the iterative approach will actually get us closer to the final robot faster.

44:33 And I will say, you know, for those listening and not watching, it's probably the smoothest robot we've seen. Like- Thank you... every humanoid you've seen is jerky- [laughs]... and even humans don't walk as smooth as this thing probably. Because, because, yeah, the model is- Uh-... can, can move every single m- motor at once. Yeah. You have more than two cohort, yeah. Yeah.

44:51 Yeah, and that's actually a very hard problem, because if you th- just think about, like, uh, the model, you know, it's outputting action, versus action, you know, dimension, right? The bigger the dimension is, the harder to fit, uh, the model to the data. But here the action space is quite large, but, uh, uh, we did very good job to make sure the model fits well and gives you a smooth- Can you bound the order of, of magnitude of the action space, or is that...

45:14 Yeah, so it's- Right... in the orders of, like, uh, you know, about 100 action dimensions. Yeah, and these are- But actually get that-... all continu- continuous, right? Continuous, right. So the model is just basically predicting at very high frequency, you know, like 10, like at least 50 dimensional vector, like 10, 20 times a second to get that continuous motion.

45:34 And pr- potentially predicting one to two seconds out, so that is another 100 frames or whatever. Exactly. There's a bit of a history, um- Yeah... similar to context of previous state, I guess you have. So context is a very important thing in robotics, right? 'Cause otherwise, if you're just making decision based on the current observation, sometimes it's really hard to tell, like, what to do next, right?

45:54 Be- for example, like if you think about opening a microwave, like in the middle part, it's unclear whether or not you're- Closing... closing or trying to open it, right? So, like, context is also really important. Yeah. And I, I have a s- screenshot of the, the fridge demo where it is, there's agentic planning and tool hauling and all those things. Yeah, yeah. So we can show that- I mean-... for listeners...

46:13 I think that's an interesting side. After the model- Yeah... how do you wrap it around into agentic harness? Like, what is- Clock code... final deployment? [laughs] Clock code, a UI. You have a lot of cool UIs. Um- Yeah... but you know, what, what goes into wrapping model output towards final thing? Is this custom per use case? Is it you have your own SDK? Yeah, yeah.

46:38 So I mean, eventually we will have our own SDK, but, uh, the way, the way we think about it is we want a general, like, harness layer for robotics foundation model. Uh, today that is our own model, right? The way I think about is, like, for any, like, real task, long task in the physical world, there is a lot of, like, a high-level decision-making you have to do, right?

47:00 It's not just, like, robot, you know, only reasoning about the physical world in the action, continuous trajectories, right? 'Cause that's a very, like, low level, the thing you're actually executing in the physical world, but that's not the intent that, you know, like if, when we are doing decision-making, we think about steps of a recipe.

47:17 We think about, oh, like, I finished this, I need to do that. While I'm waiting for this part to finish, I need to go somewhere else to do something else. There's a lot of high-level cognition, decision-making that's happening, and I think for robots to handle even, you know, the laundry task we are showing you, it has to have that capability.

47:33 Because, you know, it needs to reason that, okay, the washer is running, so I should go do something else while the washer is finishing. But now the washer is done, I should take, I should take the towels from the washer and put it in dryer immediately, right? Because if not, then that becomes a bottleneck in the whole workflow, right?

47:50 So there is a lot of reasoning that needs to happen for robotics, and I think in theory, you could just, you know, try to learn that with a, you know, low-level model that's outputting actions, but it'd be extremely inefficient. So that is why having a hierarchical system, like putting a harness, like agent- agentic system around our world action model is what allow the model to compose different capabilities, different skills to solve a much longer horizon task.

48:16 It's very much like in language model, right? Like two years ago, three years ago, the model was only good at autocomplete. Like, humans have to, like, tap, tap, tap to get the model to keep spitting out code. But today, just, like agentic system just carries the end-to-end task, and that's... And, and I think that's important, because I think today in robotics, people are still thinking about autonomy at the very task level, like, like one minute, 30-second task.

48:42 But really, the value is about tasks or, like, workflows that go on, like, unsupervised for hours, and that's what we're building towards. And that's why I think thinking about- The robotics problem, not just as a modeling problem, but rather as a system problem. You need good agent, good harness engineer, good model, good hardware, everything.

49:02 It's like how we think general purpose robots ought to be solved. Do you guys see benefits from language model coding tools getting better? So new Opus 5.5, new Astra. Um, do you have language models in your agent harness? Yeah. How does AGI being developed in language model space affect you guys? [laughs] No, no. Yeah, no, absolutely. Like it affects us very deeply in many ways.

49:29 One is that, like we are extremely AI native company, right? Like coding agents, all of that is a very fundamental part of our work, and we're very welcoming of that. Like whenever there's a new model, we're the first one to test it, right? Like we being-- like when GPT-6 came out, we had like a company-wide hackathon the same week, and the thing that our engineers built was incredible.

49:52 And many of them are-- many of the features we built is already made in production, right? So we really like using coding agents to accelerate our work. Uh, just for the people interested, uh, anything surprising in GPT-6 that you found? Yeah, I think, uh, it's about the ability for it to like construct virtual environments, right? Turns environment to simulation.

50:12 We talked that, uh, [laughs] twenty minutes ago. I think that's really exciting. It's interesting. There were so many like CAD Zillow to 3D-ish- He lost the hackathon to the a CAD, uh- I won the-- I did the CAD bullshit [laughs] other way, other way. No. Oh. That was... I won a GPT hackathon just for fun.

50:31 I was like, "I know nothing CAD." Yeah. Um, might as well play around with it 'cause every demo is doing it. I'm like, "It's gonna be garbage." And I'm like, "Well, it's kinda grounded, it's kinda good," but- Yeah... it's just hackathon stuff. [laughs] Yeah. Um. Cool. I, I, I, I was gonna round out, uh, some of the use cases and then go into a little bit of company culture and your personal background. Yeah. Yeah.

50:52 I have two, uh, two technical questions. Yeah. Uh, one is, are there any like goals you wanna set for yourself, like six months, one year, two year timeline that you see robots should be able to do, fill in the blank, that you're kinda working towards? Yeah. I mean, in the six months to one year frame, our goal is definitely like scaling up our robots and our model and truly getting them to deploy- I mean like a task. Like, uh, right now it's cutting or opening a fridge.

51:21 Is there... Like, do you set yourself to, by this timeline, we should, to stay on track, be able to do whatever? Yeah. I think, uh, in terms of deployments, we're exploring many different industries. There isn't a particular task we think we have to nail, right? Because our goal is to build a general model or general platform, and we think of each different task as like validation of our platform as a way, as a wedge for us to figure out how to build everything better.

51:51 So today we are doing a lot of that in laundry, in restaurants. Uh, we have already deployed tens of robots, but our goal next year is like overall just deploying hundreds of robots. Uh, my part two on the last technical question is, how do you see the rest of the industry? So fill in for people, what other approaches are people taking to world models? Why are they wrong?

52:14 Um, what is different but good about their approaches as well? Just what is the landscape of robotic world models in this space- Yeah... and why-- how have you come to your current? Yeah. No, that's a great question. So I think robotics is the most exciting space today, and there are just so many, uh, uh, you know, amazing companies that I truly respect.

52:37 You know, they've been doing great frontier work like us, and, uh, it's very exciting. And I think what we are seeing today is that there's still many different approaches to robotics. Like we believe that hardware needs to be built in-house. We need to own the whole, uh, chain. Uh, we believe that scaling human data is extremely important.

52:55 But there are companies that's betting on just scaling robots data or scaling some sort of intermediate like data collection devices, right? All of them are in play, and, uh, I think it's different approaches. We'll just see what happens. I think right now, uh, it's still a bit early to tell, but we feel pretty good about the way we're building things just because, like we as a company, culture-wise, mission-wise, we're not here to just build a model.

53:21 We're here to truly get robots deployed. So we deploy as much as possible, work on the hard problems, so we identify problems much earlier on. So then I think that really helps us identify what is the true bottleneck and then solve it. And I think for, uh, folks, for people who are only working on one part of the problem, it's kinda like fog of war.

53:41 You don't really see the whole picture, and you might be building something that's extremely cool technologically, but may not be what people need or want at the end, right? I think there's a whole graveyard of robotics company, and I think all of them had very shiny, very cool robots. But, uh, you know, just figuring out what people actually want is actually...

54:00 It's very easy to say, you know, like build something people want. You know, everyone says that, but it's actually really hard to execute a robotics company. On the technical side, you have your world action model. Is that standard across industry? Is there other directions people are taking? Yeah. I would say that's also not standard, so we are building world action model on top of world model.

54:20 So there's many companies building world model. You know, you mentioned Runway, they're also getting to adapt, adapt video world models. Right. Uh, there are also companies building, uh, vision language action model. VLA. So that's like taking vision language models and adapting them to robotics. Uh, so I think there's many different ways to go about it. But I think at the end of the day, like what we call the models is less important than what we, like believe the model needs to be able to do at a first principle level. So I wanna explain that a bit.

54:51 I think for robotics, the most important thing is that you need to be able to predict the future. Right. Like robotics, any physical system is fundamentally about based on the information you have at the present moment and before, what should you do next? So it's really about predicting what comes next.

55:11 So predicting robot action is like predicting what the robot should do next. But if you think about it, robot is just a part of the world, right? Robot influences the world, but the world also influences how the robot should behave in that environment. So if you think about it this way, then the maximal, the most first principle thing you can do is just predict as much of the world that you can possibly predict. Right.

55:35 So that's why we believe predicting the future is extremely important, and today the instantiation is world action model. And VLA only predicts robot actions where some more abstract representation of the world. But whenever you're predicting abstract representation of the world, you're encoding your own bias into what's important to predict. And the bitter lesson tells you, let's remove human biases.

56:01 Let's just take the most raw representation, the camera inputs, and just predict all of that. Right. So that's how I think about, uh, modeling for robotics today. So just to press you a little bit on this, you're, you're basically you're saying your approach is better than the VLA, right? Because you're, you're, you're, you're emphasizing prediction. I fully buy that.

56:20 I, like you've sold me on it. Yeah. If you were wrong, what is the most likely number two [laughs] So- That someone else is doing? Interesting question. Yeah. So yeah, there is a caveat in what I've said before, and I think that's where I could be wrong, right? I said that you need to predict the world as much as possible, right?

56:40 But what that doesn't say is that does your sensor have everything about the world that you need to predict? I think I touched on this a little bit, right? Like if you're trying a world model that's purely based on videos, I think it'll be a very powerful world model, but it may not be a sufficiently rich enough world model because you're missing critical aspects of the world like sound, like touch, right?

57:06 So I think what, where we could be wrong is like, okay, I have the best world model on vision, vision-based world model, but tomorrow someone else comes out with a vision plus tactile world model. They're just like 100X more data efficient just because, you know, vision is a very high information or like a highly redundant, you know, input because, you know, every other pixel that they don't change that much.

57:33 You can see, you can understand the world without perfect vision. But tactile, you know, other modality might be much denser in information, right? So maybe having that, predicting that would just allow you to use much less vision, so you can get to the same performance- Oh, oh... with much less compute, with much less data, right.

57:51 But- I think you'll be able to adapt very easily. No, exactly. [laughs] The reason why- There's also a level of how much data- Right. So the reason why I bring this up is because I think if vision tactile world model or vision tac- audio, whatever world model, is the final thing that solves robotics, then I think robotics companies will be able to do that better than video generation companies is trying to pivot to robotics. The reason is like to understand sensing is actually very complicated.

58:20 Like we've been building that since the company started, so we know the bottleneck, we know how to combine them. But for if you're just trying to adapt a video model to robotics action then like, you know, if we were wrong, if, you know, video model is not enough, then like you don't really have the knowhow to get tactile to work very well. But we have that. This is a side question, but I, I think it's a quick one.

58:42 Uh, heat seems very important. Is infrared just another color to you or is that a- Yeah. You could be another- Just another color... could be another color, another input- Yeah... to the model. It's, this- Could be used- And-... in, in cooking scenario, actually. Like I know companies that's using infrared to like know how cooked [laughs] the food is. Yeah.

58:59 'Cause that's really hard to tell from RGB camera, right? My, uh, very naive pushback on your last point- Yeah... is, uh, maybe we don't need to understand hardware cloud code or Codex can just, you know, pipe code it in. Yeah. Uh, there's a lot you can do without understanding, but of course- It's just not in their distribution.

59:16 Yeah, yeah. [laughs] Uh, so, so I wanted to bring it a little bit to your personal story or your background. Um, we both went to Penn. [laughs] Uh, I did not- I just found out today. Uh, yes. Crazy. Yes. Uh, I, I, you know, my, my, my big regret is I didn't join the self-driving, the DARPA challenge. Okay. Which, uh, then I could have been like early- Yeah... Waymo or something. Um, you were at Grasp. Yeah.

59:36 And then you, you've also worked at, uh, Google and Nvidia. What are the different focuses of, let's say, academia versus Google versus Nvidia? You've seen all these places, right? Yeah. And then like that obviously caused you to say like, "Screw all that, I'm gonna do my own thing," right? You could have joined OpenAI's robotics lab. Yeah.

59:54 Uh, so maybe just take us through chronologically, but also like, uh, one thing you learned from each. Yeah. So chronologically, yeah, maybe it's useful to talk about how I got into AI in the beginning, right? So, uh, as a kid, so I spent time in Japan and China. I grew up playing Go, you know, the Go in AlphaGo.

1:00:11 So when AlphaGo came out, I was a senior in high school, and I was mind blown, 'cause I've been playing this game for many, many years. It's so complicated. Like I think before AlphaGo came out, people were like, "Okay, Go is one of the game that AI will never solve." The search space is like larger than the total number of atoms in the world.

1:00:29 How could you possibly search for solutions in that? Then like, you know, RL, reinforcement learning, just like cracked it, right? And I think that was how I got into AI initially. Like I was just really interested in like models or systems, agents that can self-improve, right? So when I got to Penn as a PhD student, I just really wanted to work on reinforcement learning. So I'm co-advised.

1:00:52 Both my advisors were interested in, uh, RL at the time. One of my advisor was more in robotics. He did a postdoc at Berkeley robotics doc. So that's how I got into robots. Uh, at Penn, you know, uh, we have GRASP Lab, I think one of the best places to do robotics. And, uh, so that's how I got into doing RL for robotics in my PhD.

1:01:13 And then, yeah, I did a bunch of research thinking about how to connect reinforcement learning, uh, with human data, how to do reinforcement learning, uh, theory in a very rigorous way. And then I went to Nvidia. That's when GPT-4 came out. At the time, we were very interested in just solving a pen-spinning task. I don't know why. Like, uh- It's the dexter- it's the dexterity limit, right?

1:01:37 I mean, I think- That you can solve Rubik's cubes and then you- Yeah. Time I was working with, uh- Jim Fan... Jim Fan and Yuko, which you guys probably also know. Like, yeah, we were just like, "Okay, let's try to do dexterous pen spinning." The initial idea was, like, learning it from human videos, right? We were trying to watch humans do it.

1:01:53 But then I realized, okay, today we don't have a dexterous hand that can spin pencil at all, so how would I even be able to evaluate the algorithm even if learning from human worked at all? So I said, "Okay, we need to first try this in simulation." Then as soon as I tried to set up a simulated environment for a robot hand with a pen, I immediately saw what the issue is.

1:02:15 The issue is, like, like, if I had to come up with a simulation environment, the reward function for reinforcement learning to solve pen spinning, that's impossible. Like, I wouldn't be able to come up with all the terms that you need to get RL to learn the right behavior. And then I, I was just thinking, "What if a model that can write code can do that?"

1:02:34 That's what became the Eureka project, right? We basically... Like, in my opinion, we had the first, like, recursive self-improving robotic agent in simulation, 'cause we really, like, used that to solve a new task that no one has demonstrated before. And that, that, that allowed me to see the potential of using language model, using coding models to really accelerate robotics.

1:02:57 So then I went to Google DeepMind, 'cause at the time, you know, I mean, even today they have one of the frontier models in Gemini, right? So I was looking at how to use Gemini for robotics. But eventually I realized, like, I had a lot of fun doing academic research, but there was always a question in the back of my mind like, "Does any of this matter?" You know?

1:03:17 Like, when you're working on this kind of thing in the lab, like, you know, you ask me, "How do we come up with so many interesting, fun tasks?" You're basically doing your own eval. You're grading your own homework. Like, I wanted something more real. I wanted something that could actually be used by, uh, millions of people in the world. And I felt 2024 was when we started Dyno, like end of 2024.

1:03:37 That was the right time, that we can take some of these technology to actually putting them out in the real world. So that's why, uh, I started Dyno at Lynn in New York. And, uh, fast-forward two years, we built our platform, our model. Our robots are in Jinjaifang. Go check it out and, uh- Go eat. [laughs] Yeah, go eat. It's been a journey and, uh...

1:03:58 But what still gets me really excited is, you know, like some of, a lot of the ideas that I temporarily abandoned when we started Dyno because, you know, to build a company you have to be very practical. Like, I think simulation wasn't ready in 2024. But now with, you know, like, the new frontier models and what we're seeing internally, like, this might change, and, uh, that's very exciting. Yeah.

1:04:20 And, uh, so, N- then also let's talk about company culture. Yeah. Um, what principles you're, you're building with the company with. Uh, obviously, like, the, the, you're, you're responsible for the science- Yeah... and the, the research, uh, agenda. Yeah. Uh, and it, it just seems like when we meet the engineers, they're, they're very engaged and then they have a lot of ownership. Yeah. They have, they have a lot of motivations.

1:04:38 I'll say it's, it's interesting to see how many side projects people just work on. [laughs] So some of the demos we've seen, it's just the engineers that work on it were like- Yeah... "Yeah, this seems fun. Might as well do it." Yeah. Um, to your point on fun evals, right here, like two feet behind the camera- Can't talk about that. No, I mean, the thing is just the e- Can we talk about it?

1:04:58 The eval is just- [laughs] There's highlighters, four highlighters being rubber banded together. Yeah. It's like who comes up with, "Ah, yes, we need to test putting rubber bands on highlighters." It's just- Stationary... fun stuff. People are very, uh- Yeah... you know, creative here. It's a beautiful space. Yeah. I think culture... Yeah, so let's first talk about, you know, how we run the company, right?

1:05:17 Like, we think of ourself as like a, like a, like a, eventually like a research plus product company. Mm. It's kinda like OpenAI where you do frontier research, but eventually all of them make their way into Codex, through ChatGPT. That's how we think about it, too. We're not here to just do open-ended, like, whatever thing that interests us at the time. We're always, like, product-driven. Like, we wanna deploy our robots, our platform to a new customer.

1:05:43 What are the technical problems we have to solve, right? We have to solve generalization. We have to solve whole-body control. We have to solve how to adapt quickly. We have to solve how to make the robots follow language. We have to solve dexterity. So once all those problems are very clear, then, like... I mean, our engineers are very capable.

1:06:01 They're very smart, so they just create instances of the task. They create tasks that will stress test those capabilities, right? Like the j- the grabbing a drink task is just how we test language following, right? You can just specify what you want. If the model follows intent, follows instruction, it should be able to do it.

1:06:19 So that's how we think about, like, at the top level, we identify what are the things we have to solve, then we have a lot of bottom sub-project to, like, actually drive execution to solve them. And, uh, you know, you hire the right people, and everyone here is super passionate about robots. You know, they've been doing this for a long time. They love it.

1:06:38 And we just create a culture where it just naturally happens. And, uh, it's, it's a really fun place to work. What's the split between, like, uh, technical research software versus hardware- Yeah... side of the company? Yeah. You know, we recently grew so much that I don't have the perfect number, but I would say on the hardware side, we have, I would say probably like 30%, 40% are hardware, just because hardware you need a lot more people.

1:07:03 And then, like, you know, we have, like, around 10 people working on models, and then we have, like, software engineers, uh, robotics engineer. We have a lot of, uh, very talented people on the business side. You know, the most fun thing about building a robotics company is, like, you see everything. You know, everything under the sun that could possibly be helpful will be helpful, right?

1:07:24 And, uh, so you just get so much exposure working with so many different people from many different backgrounds. We're all very passionate about robots, and I think it just, it's just been really eye-opening for me as a founder. 'Cause when I was a researcher at a lab or, like, a PhD student, you know, I was talking to researchers every day, and that was great. That was really good.

1:07:45 I still get to do that every day, but now I also get to learn, like, just from so many different people from different walks of life. Like, some of our, uh, people who collect data, they had no robotics experience. They were work- working in retail, working in some other jobs, and they came here. You know, they also became very well trained in robotics, and I think that's just pretty amazing th- amazing to see.

1:08:07 And I think of myself as, like, a... I'm still learning, you know. I think of myself as just always learning in the robotics space, and, uh, I haven't learned so much as I am now. You know, even though, you know, you think about me now, I'm not a student anymore. I'm, I'm here, like, running a company, managing a lot of things, but I still feel like the day-to-day learning is, like, just so much.

1:08:29 Um, and then just to close off, uh, where is Dyna going? Uh, two years, five years, what's, uh, the vision? Yeah, I think our vision is to truly... The high level mission is truly bring, like, a f- you know, robots and, uh, intelligence, right, into the physical world. Like, act- that means, like, actually seeing our robots deployed, like, thousands of, uh, customers, even more, millions of robots. That's where we wanna get to. What's the, what's the name from Dyna? Dynamic?

1:09:01 Is there a- [laughs] Is there a story? Yeah, there's a, there's many stories. So- [laughs] I, I na- I named Dyna. There was a lot of, uh- Take the credit [laughs]... [laughs] internal, uh, different options. But, uh, how the name came to be was there are two things. One is, uh, so, uh, Turing Award winner, Richard Sutton, who was, like, the pioneer of reinforcement learning, wrote a paper called Dyna many, many, many years ago. It's about basically a agent that self-improves in a world model.

1:09:30 And I felt, even when we started the company, like, we weren't doing world model yet, but I felt that's how robotics will get solved. And now our, you know, technology, technology direction is, like, converging to that, so I felt that was a really neat inspiration also because I worked in reinforcement learning, have a lot of respect for Richard.

1:09:49 The second thing is that, um, [laughs] it's a bit more personal. Like, I grew up in Japan and, uh, you know, the year I was born there was just, like, a cartoon show. I don't know if you guys have seen, like, Ultraman. Yeah, of course. Yeah, yeah. The, the, yeah, the Ultraman that year was called Dyna. Oh. Oh. Yeah. So yeah, it's a very subtle thing not people know, but I guess now everyone would. [laughs] Yeah.

1:10:10 The team is finding out too. Yeah. Yeah. Thank you. But, like, I, I, I think Richard would appreciate it. I don't know if you ever talked to him, but- No, I've seen him at RL conferences. You should just be like, "Hey, this, this company- Yeah. Yeah, I'll invite him, come... inspired by you." Everyone knows Richard for better or less anyway. Yeah. Yeah. Yeah, also better or less. [laughs] Yeah. Okay.

1:10:27 Uh, any parting thoughts or calls to action? Calls to action. Yeah, I think, um, I do feel like as an industry, robotics is still very demo oriented. Like, I have done a lot of demos in the past in my lab, you know. Even here, like, we had to show some demos, but I think we need more people that's truly, uh, mission focused on doing whatever it takes to get robots deployed, get, uh, robotic foundation models truly integrated with hardware, and seeing them at real customer sites.

1:10:57 I think we're a pioneer on this regard, and I wanna see more in this field. And for the audiences watching, if this mission inspires you, if, you know, solving the hardest problems today is what makes you wake up every day and you wanna contribute, you know, we're always hiring, and, uh, yeah, looking forward to that.

1:11:14 But I think just- Bay Area? Anywhere else? Uh, yes, just Bay Area today. Just Bay Area. But, you know, as company grows, never say never to other locations. Yeah, but, uh, that's what I wanna say, and, uh, I wanna just see more robots in the world. Yeah. Cool. Congrats on everything. Uh, I mean, this is, this is incredible progress for two years. Thank- Uh, yeah. Amazing.

1:11:34 Yeah, I really appreciate it. Thank, thank you for, uh, visiting us. I hope, you know, it's, it's been fun, and, uh- So fun Yeah, I would say, like, yeah, I, I, I don't regret doing this on site. Like- Okay... yeah, we do a lot of in our studio. Yeah. Like, this was worth it. Yeah, yeah. I think for robotics, gotta do it on site. It just- For, for people listening in, you know, watch the, watch the video version.

1:11:55 We have fun B-roll