One Brain, Any Robot: Skild AI's Skild Brain Explained - Ep. 295

0:00Abhinav Gupta:Robotics is a data problem. Unlike language or vision, there is not much data in robotics. There is no internet of robot data. So if that's a scenario, we cannot pick and choose which data we use. So we go, in a most general fashion, every single instance of our brain which we deploy for any kind of task or any form factor that contributes in making the brain better for the future scenarios.

0:27Deepak Pathak:Welcome to the NVIDIA AI Podcast. I'm Noah Kravitz. I'm here today with Deepak Pathak and Avanab Guptal from Skilled. Skilled is a robotics company that's building the OmniBrain, a universal brain that can power robots across any form factor to tackle any task. It's amazing stuff. Very excited to find out about it from the source. And so let's get into it. Deepak, Avanab, welcome. Thank you so much for joining the AI Podcast. Thank you so much for having us. So, Deepak, maybe you can start and tell us a little bit about the company, about Skilled,

1:00Abhinav Gupta:and then you can both talk a little bit about your roles. Yeah. So at Skilled, as you mentioned, we are building a general purpose brain. So we call this omnibodied intelligence. Any robot, any task, one brain. So think of what ChatGPT is for language. We are building a general brain for any physical device or any kind of robot. So this is absurdly general. You can have a humanoid or a dog-like robot or a robotic arm on a conveyor belt, all being controlled by the same shared brain, shared intelligence behind the scene. So why do we go so general? And the reason is robotics is a data problem. Unlike language or vision, there is not much data in robotics.

1:43Abhinav Gupta:There is no internet of robot data. So if that's the scenario, we cannot pick and choose which data we use. So we go in a most general fashion, every single instance of our brain, which we deploy for any kind of task or any form factor that contributes in making the brain better for the future scenarios. So this is the main goal behind this. And personally, in my role, like I've been, we both have been professors before this. So we are extremely technical. We have been involved in bringing up these technologies in the robot learning area for the last decade and more. So our role is both on the technical side to make sure that these things get built and they are super general, transferable.

2:31Abhinav Gupta:But our focus is also a lot on deployments. We do not believe deployment to be a, it's not hindsight scenario. Like for instance, in case of chat GPT or language models, folks did research for several years. But once it was ready, you have a million users in seven days. Maybe one day, I don't remember. Maybe 100 million users in one month. Right, fastest growing product. Physical AI is not like that. Things take time to deploy. So for us, deployment is our first priority from day one.

3:02Deepak Pathak:Yeah, makes sense. And you mentioned being a professor.

3:05Abhinav Gupta:You're at Carnegie Mellon? Yeah. And the company is based in Pittsburgh? So company has HQ in Pittsburgh, but we have offices in Pittsburgh. Now we are also in Bay Area, the San Mateo area. And one office in India, Bangalore. Fantastic.

3:18Deepak Pathak:And Avana? Yeah, I think one thing which I want to start from is like, the reason we are actually so excited about this is because we are almost rethinking the way robotics is done traditionally. Like traditionally, robotics has been a very classic, like a vertically oriented field, right? I mean, so what that means is if you think before this era, you first decide what vertical you want to place the robot in. So like, let's say I want to build a welding robot. Now you go and start making your hardware, which is very specific to welding. You start making your software, which is very specific to welding.

3:54Deepak Pathak:Now, the problem with these kind of deployments has been is it's very easy to guess the first 80 % or 90 % of the performance. But then you hit this wall, which is called the corner cases in the physical world. Right. There are so many corner cases in the physical world, like someone might lead a package in front of you and now it becomes a corner case and so on. And so that is why if there is a corner case now, because you are at 90 percent performance, you will still not be able to get it completely automated. Human still needs to be around to make sure the corner cases are handled and so on.

4:24Deepak Pathak:And that is why it has not been traditionally robotics has not really gone big mainstream, essentially. Now, however, things have changed when AI came in. Like if you think language also before this whole came in was very verticalized. There were different companies building chatbots. There were different companies building search engines. But once LLM came in, they became the horizontal platform. And now everyone is building on top of that horizontal LLM platform. That is exactly how we are now thinking about robotics. We are building this horizontal general purpose brain. that will and this general purpose bin is can then be fine-tuned for different verticals essentially and our thesis is that if there's a corner case of one vertical becomes the central case of the other vertical so now the data is from everywhere and so now it will be able to handle these corner cases through the data play with the different verticals in in terms of what Deepak was talking about I mean we are definitely like very similar in that profile because both of us are professors So we do not divide our work like, oh, I do business and you do this kind of stuff.

5:30Deepak Pathak:We are more think of it as extension of each other's brain and thinking about it, strategizing about it and the whole and really, really focusing on deployment.

5:38Abhinav Gupta:Humans are limited in the sense we cannot enter each other's brain. We are fusing the omnibody intelligence in the human way.

5:46Deepak Pathak:I have a feeling from talking to you guys for five minutes that you might be closer to fusing brains together than you realize. I don't know. You seem to be on the same wavelength. So what was the inspiration? I mean, you discussed, you know, in some ways, the inspiration for Omnibrain building that horizontal platform. But were there deficiencies or gaps that you saw in existing robotics foundational models? Or what was really the impetus to say, hey, we need to go do this a different way?

6:16Abhinav Gupta:I think if you look at the current systems, I think I've already alluded to it. in a way when the robots are currently deployed they are they behave more like machines right so everything is measured everything like in factory setups everything is so for instance if you look at a classical automation line you will have a robot but around the robot you'll have a big cage everything will be measured very precisely the whole setup may cost several times more than the robot itself. Then if anything were to change, you have to redesign the whole setup. And then people talk about consumer applications where things change.

7:00Abhinav Gupta:Let's say your home, right? You don't, you can, no matter how many sensors you put, you cannot measure everything, single thing to 0.1 millimeter accuracy. Sure. Right? So this, this whole paradigm of robotics has, the main shift in robotics has happened going from this programming in the behaviors to learning the behaviors which means you learn that from data so now the engineering part has gone from okay how should my robot move what failure may occur to thinking where the data will come from or how can i make it high quality how can i get it at scale and that's where the shift has come so we saw the shift uh in academia like uh we could be began seeing results one after another like we could get a result today and demo live demo in a conference the next week.

7:48Abhinav Gupta:So for us, it was like either we bring it to the masses or we are the ones who just get eventually replaced by it in some way. So it was just a no-brainer for us that this is the future of robotics. And I think this realization is also happening at the same time in the general field. You can see the excitement around physical AI in GTC. we are working with several major players in the space to bring this. So this is not really, oh, this happened, hence this should happen. This is the way to scale. If you do not do this, it is almost impossible to scale the way how things have been in the robotic space.

8:29Deepak Pathak:I noticed on your blog, on the website, I was reading an article about training on video data. Can you talk a little bit about the benefits and why you're training on video data? And is that the primary way, the only way you're training your robots? Or are you bringing data sources from other places as well? So, yeah. So, I mean, when it comes to robotics, we have multiple choices when it comes to data. Sure. So there are three main sources of data. The first source of data is videos. Or maybe let's start with the robot data itself. So now where you will do it is you have to collect robot doing a task.

9:10Deepak Pathak:And that data itself can be used to train the robot. However, this is very hard to scale because you're collecting data with robots. So for every data point, you need a robot. You need humans to control the robot because the current robot and we call this teleoperation. So you have to collect data with teleoperation. The good thing about this data is it's the richest form of data because robot itself is doing the task. So you can read all the sensor values. You can read all the motor commands that are going in the robot and so on. The problem with this form of data is very hard to very, very hard to scale.

9:40Deepak Pathak:and so when it becomes hard to scale it's very hard to learn large-scale AI models on type of the second form of data is like something like videos now in this case you are that there's huge diversity of the data because we are collecting videos in US people are collecting videos in India China everywhere so you can you have huge diversity of the actions everywhere and so on so this is a scalable form of data highly diverse but the problem with this form of data is that it's not rich enough. You do not know what exact actions, what exact forces people are applying to do it. And then there's a third form of data, which is the simulation form of data.

10:18Deepak Pathak:Now, in this case, it's highly scalable. Simulation is as scalable as it gets. You can collect trillions of examples in a day, for example, and so on. It is also, you can measure all the forces in a simulator and so on. But the problem with simulator is there's always what people call sim-to-real gap. Like simulator cannot be exact replica of the real world there's always some difference and so now you have to bridge this sim to real gap either through algorithms or some other data and so on and so for at scale we use actually all three different forms of data we believe every form of data is critical because every form of data is complementary to others like i mean if you think videos are scalable and diverse simulation is scalable but not diverse and uh and then the third one is the robot data which is the richest form of data so every form data is useful but some data has different metrics videos is not as good quality for robot training as like for example the real world data so what we do is we use the video data to pre-train our models this is the data that is available in billions already so we can pre-train our models to build a build a model however the problem with videos is if we can learn everything from videos.

11:31Deepak Pathak:Deepak gives a great example that if we can learn from videos, all of us would be Federers. Because we will watch Federal and we'll start playing like Federal and so on. So that's never going to be sufficient. Just watching videos is not going to be sufficient. If it was sufficient, I could dunk a basketball, but I can't. Exactly, we cannot. And so that is where for us simulation comes into play. We get the idea of what the task is, what the action is from video, but then we practice it in simulation. We robustify it in simulation. But again, simulation is, there's still a gap. Remember, sim2real gap still exists.

12:04Deepak Pathak:And now we take this model which has been pre-trained on videos and simulation, but before deployment, we post-trained it on the real world data, on the small amount of real world data that we can collect in factories or whatever task we are trying to solve. And that makes it precise and help it solve. So you get the robustness from this pre-training data, like the corner cases remember i was talking about these corner cases yes those videos and that simulation helps you to robustify okay and to make it precise is where the post training data comes in right right um and so this you can also find analogies with language i think ai has been

12:40Abhinav Gupta:mainly successful right at a massive scale for language data right but the same recipe is there like you have this when you are building this general model like to go general first and then you go specialized model the general model is training on all of internet data right like from different sources, different articles. But then let's say you are open AI, you build chat GPT, and then Amazon comes and say, oh, I will deploy a robot. So your model in my amazon.com website, then you will take that model and you will fine tune it. Right. Right. And then you deploy it. So then data from just amazon.com will be very high quality for Amazon, but very low in amount.

13:18Abhinav Gupta:Sure. So it's used for post training. Internet data, maybe it's low quality because people are different things and maybe there is junk text many many many places so it's low quality but at massive scale in pre-training time so this separation of pre-training and post-training is how the current ai revolution is is governed right you're even at nvidia right you have chips for inference you have chips for pre-training and this is the same separation we are building to robotics and which is why we are seeing this immediate access to variety of applications which you would not have otherwise.

13:56Deepak Pathak:You've talked about this a little bit, but maybe kind of to put a narrative around it for the viewers and listeners. Can you talk about kind of what it takes, the process of building, testing, and deploying, bringing to market something like the Omnibrain?

14:14Abhinav Gupta:So it's a very complex question because it really depends on the scenario, right? Like in language, it's very easy because you can ask a question. it's just prompt does everything oh sure so the general recipe which we are going towards is that the behind the scene brain is shared okay so any single action you will take will improve the brain now how do we orchestrate the deployment of this brain so the idea is let's say if you have some task if we have seen that task before let's say if it's a task of moving around or walking or jumping over things. We can do that already very well. So in that case, you can just take the brain, put on the robot, and we'll just work off the shelf.

14:55Abhinav Gupta:Then you can build applications on top, like, okay, I want to use the robot for taking a selfie or security inspection. That's the second part, right? But let's say now you go to a different task where the robot is, I don't know, like assembling a GPU on a conveyor belt. Now it's a super different task compared to what people generally do even humans need training so in that scenario what we do is on the on that robot we may collect data for a few days okay either do that or if or if you already have the assets then we'll collect it in simulation either way then we use the data and we post train the model and then that model takes over and it turns on the robot directly okay so in this case now what you have done you bridge the gap between what you saw before to a very different task by adding data from the actual task so it's called domain specific data right now as you deploy more and more of these robots imagine you are getting a fleet of specialists which all came from a generalist right so it's very much like you know when you're in high school you know many subjects right right I did PhD.

16:06Abhinav Gupta:I barely know any chemistry physics at this point. Right. Right. But, but I needed that to get to, to get a knowledge right now. So then when you have this specialist, then the data can pull back from all of them and come to the same brain behind the scene, which is not how, what happens in humans, but we can do it in a computer. And now this happens. Now when you have a next task to go to, you may need, you will need less data for the next task. Now this act as a, this is what we call in other words a data flywheel like you may have heard this term for self-driving like humans drive cars so this data flywheel now we orchestrate this across vertical so you start with factories they act as data flywheel for semi-structured scenarios like hospitals grocery stores i don't know like hotels you did a flywheel from there helps you get to the ultimate challenge which is like homes consumer robots so this is basically how we are orchestrating the so self-sustaining data flywheel loop from every development and this is why you probably understand now why do we have omnibodied brain because you want to take benefit of every single data point and use it for the next complex task right and does the same concept apply to different form factors yeah i mean on factory it's a robotic arm in home probably some humanoid or some other form factor for security and inspection with dog-like robot in delivery a different form

17:32Deepak Pathak:factor so across all factors so I want to ask you guys a little bit about how you're using in video technology and specifically around synthetic data and simulation as you mentioned but really just kind of open-ended how are you

17:46Abhinav Gupta:wouldn't video stuff you're using and how does it fit in I mean so our company is two and a half year old but we I have been working personally with Nvidia I think since 2018 like not at Nvidia working with them like so there is this whole the suite of simulation like isaac sim back in the day there was physics and isaac jim so we use that the physics component of that to really create these gazillion scenarios on which we can try and practice like what abhinah was describing practicing and learning so that's that we are basically the og user and and we are now working with nvidia on like newton uh as well and in fact we are co-developing better physics solvers oh great yeah probably will open source them uh together that's one one uh collaboration on simulation side second side is the video models like uh the cosmos and other models uh so we use them to data augmentation like every data point you can get that and you can create multiple variations with these generative ai models so we we leverage we partner on that front and i think the biggest of all is this the whole compute platform sure yeah those robots are the next gen next generation device right and the solution that worked for llms of big gpus enough in like servers it will look very different for a robot because robot doesn't have time to connect to a server if it's falling right right react immediately so on device edge compute this is where we are partnering as well excellent so when

19:23Deepak Pathak:you're when you're testing omni brain when you're maybe when you're using it with a new partner or developing a new feature. Do you have kind of a go-to test case, a go-to scenario that you put it through? Or, you know, walk us through what that's like kind of testing something before you're ready to deploy it. Yeah, I think that's a great question. I mean, although this is also very hard because that's a problem is something general purpose, right? Yeah. And that's what Deepak was talking about, a general purpose brain. Now, if you are fine tuning it for something specialized, like bringing a special brain, should it forget the general part of it?

19:55Deepak Pathak:Does it matter, general part of it or not it probably does not matter but then it matters if there was a corner case that was coming in and so on so those are the kind of things that matter so this is why we have been trying to develop a very specific strategy of like testing these out so the first thing of course we have to test out is on the task itself let's say we are putting let's take the example of gpu that we have been working with nvidia as well as a partner as well like the gpu like putting a bus bar or on a GPU rack on a server. Now there are two requirements. First, it has to be put properly.

20:28Deepak Pathak:So that's the accuracy part of it. And then how many, how much time does it take you to put? If it takes you one day to put one buzz bar, that's not good enough for any deployment and so on. So our testing has these KPIs that we first test on. These are the task driven KPIs that we are trying to match and so on. But just doing KPIs is not sufficient because that is where the whole idea that 90 % is done through KPIs or 95 % is done through KPIs, but the rest of the 5 % is also what matters. And that's where we go and test for generalization. We say, okay, what if someone left a box here? Or what if somehow the lights were completely off?

21:03Deepak Pathak:Or like, we change these conditions. And we have these set of conditions that we want to test in, like, even if these things happen, the robot will either continue to work, but still be safe. Safety is the third aspect of it as well like in all these conditions we have to ensure that the robot is safe of course and it's not doing any unexpected behavior and so on so we basically have this whole pipeline where we first start from task metrics then generalization metrics like if things go wrong i mean this is something which you're not expecting but you still want your robot to be robust to those kind of things and you and we have like a whole list that we develop before we deploy that okay these are things that we want to test on when it comes to generalization and last is the safety that in no no scenarios that you should uh you should break the safety violations and so on so we bought something called safety guardrails also before the deployments that ensures that let's say somehow somehow someone broke the wire or some and cut the camera wire okay because now the robot is blind it doesn't see anything so that's a safety metric that we need to make sure that now the guardrails come in and say, okay, if I'm not seeing a camera, either I should stop or at least I should not cross the boundaries that I have been given by those things.

22:16Deepak Pathak:So these are all the things that you have to test for. Again, the problem with the physical world is that it's not like an overnight sensation that you can become. You put it on a web page and now everyone can access it and so on. We have to go through very rigorous tests before we can put anything online for deployment. Absolutely. So this is one of my favorite questions to always ask as we start to wrap up. What do you think the future of robotics looks like? And, you know, we try to put a time frame when next year, next two years, things are moving so quickly these days. And particularly, as you're talking about with physical AI, you know, the embodiment of AI is really, you know, this year in particular, I think we're seeing so much more of it.

22:59Deepak Pathak:But how do you see robotics developing in, you know, the next few years, five years, whatever the right time frame is?

23:08Abhinav Gupta:I think in the longer timeline, we will be able to automate every single action that humans can take in the physical world, right? Because we are following the approach, which is very similar to how this actually thing, things happen in nature. Now the timeline, and in some sense, the longer you go the more you realize that this is the way to achieve general intelligence like currently what we have so far all the results in language models vision models it is all what people call digital intelligence but digital world if you think about this it's not more than 50 years old it's a good point yeah we're humans not intelligent before that uh right so this is this is the longer term vision right now how does this orchestrate well in our uh opinion like we you will start to see already things getting automated with these kind of models in a very short horizon but high complex repeatable maybe uh less variable scenarios first so it's like what we call unstructured semi-structured uh like industrial task warehouses they act as a stepping stone i was saying earlier to get to more unstructured or semi-structured scenarios.

24:28Abhinav Gupta:This is a spectrum. Structured is like everything is mapped, like a microwave. Inside microwave, you don't really care. You don't put your hand minutes running. It's a completely separate system, right? Other part is home, which is completely unstructured. It's a spectrum. So in this year itself, we'll start to see deployments in like factory, warehouse around people that bootstraps the next one, like hospitals, hotels, service industry, that bootstraps the ultimate consumer robots. It's very hard to predict the timeline for the ultimate home robots, but you will start to see robots for sure. And you're already seeing that happening in this year or in the next couple of years.

25:07Deepak Pathak:I think in the longer run, we all agree that robots are going to be everywhere, going doing every task. And I think everyone agrees. And so shorter term also, we are like, at least in the company, we are all in agreement that this year we are going to have like the structured places like factories and warehouses being more and more automated like the penetration will start to happen by the end of this year more and more penetration and it's a middle which is unclear and that's where we always have a betting pool inside a company also like gelato bets and all these kind of bets that we keep going on then when will these things come into play everyone has a different view like some people believe that home robots might still come in two three years but then some people are arguing that two three years is still very hard i mean we have to be honest and we have to say okay like the kind of uncertainty that can happen in the real world is very very high uh and while you're seeing so much hardware in humanoid space all also are these hardware reliable to be even put in homes today like no one has put them because safety again is a big issue like in uh when you are putting them in home what if it falls and there's a child around and something like that right So we have all these kind of within the company, all these pools going on and so on.

26:21Deepak Pathak:And I think both of us are kind of like agree on the short term and the long term, but it's middle where no one knows. And we are just figuring it out. Okay. We are playing it as long. The interesting part is it's very surprising how it's playing out. I mean, because I mean, from the AI perspective, right? When I was doing my PhD in 2008, would have never guessed where we are in AI. and it's actually continues to surprise even more and more like if you ask me three years ago where would be today that also is very unsurprising yeah and so the progress of compute and the hardware coming costs coming down has just made this all so surprising that i would say even the experts like like us who have been working in this for 20 years are scared to say anything