Autonomy Across Air, Land, and Sea
48:18 Share

There are now drones here in the barrier that are flown from somebody that's twenty, thirty, fifty miles away.

Mining has had some form of trucks without any drivers in them since two thousand and seven, two thousand eight.

China, our numbers are ship building capacity about two hundred to one.

immediately. Every autie drive was just a rich the basic .

assumptions about how software is built for the kind of traditional world of the twenty ten just doesn't work in the autonomy space. You can actually .

create a really, really incredible reconstruction of the world just using these videos, generate models. And this is not hype gene.

The cars we buy in U S. In your to exit la, they are not delighted consumer products, like when you brought your first iphone.

This has been a big gear for autonomy. For example, the fully autonomous mo driver has done over twenty million miles, the equivalent of driving to the moon and back forty times, and is now doing more than one hundred thousand rides per week. But it's not just put on me on land.

For example, the F. A granted several Operators the ability to fly commercial drones without visual observers earlier this year, and this is only just the beginning. Now in this live recording from S F tech week, we Brown ten experts from breeder and air land inc. To discuss the tony ma systems, and we touched on the real world deployments latest chips and their impact on the economics building full stack quanto find risk and regulations role in vancsik frontier moderating this panel was a sixteen a partner airing Price right along with three panels. First up, we have material named chief marketing officer of sky deo studio .

is a ten year old company based here in samoa. O we provide the option drones, more specifically camera drones. We shipped about fifty thousand. He started in consumer, especially for outdoor enthusiast, will be transition exclusively to selling into private enterprises, state, local government, the federal government.

Next, we had B, J, T night head of a product and applied intuition, but also previously spent five years at we, most recently as head of a product of their self driving activision.

A blind infusion is focused on providing developer took and software to companies that are building vacs. We provide simulation and data tools that are necessary for building a domain systems.

And finally, Peter bowman, Davis and engineering fellow, now that is extensive but previously worked on machine learning at throning.

Sonic is a full stack maritime autonomy company. We build boats that are autonomists, and we saw directly to the government. So that comprises everything from the whole manufacturing to the simulation side .

on the autonomy. We also had some amazing questions from our live audience. So stay in for that at the end. And of course, if you'd like to attend events just like this in the future, make sure is subscribed. And while you're at IT leave review, all right, let's get started.

As a reminder, the content here is for informational purposes only, should not be taking as legal, business, tax or investment advice, or be used to evaluate any investment or security, and is not directed at any investors or potential investors in any A C C E fund. Please note that A C C E ennyhow ilios may also maintain investments in the companies discussed in this the cast. For more details, including a link to our investments, please see a sixteen c outcome slack disclosure.

So i'm really excited because we have someone from the business, from the product and from the tech side here as well as air, land and sea. So maybe I get started. I would love to hear how far along in the journey to autonomy you feel like we are in your industry and maybe framing the levels of autonomy and where we really all the way there? And where do we have still some way to go?

So schedule was sort of started on the foundation of autonomy, and we did not invent the quad. Are those existed prior to the founding of the business? I think we ve all experience are seeing them and even just walking our best bias or a toy form where you have some sort radio controlled mechanism to be able to tell the drone where to fly.

The chAllenge with the use of drones is that IT always required a person to be there. And if the more you go up in the use cases into more enterprise or government use cases, the more valuable the flying becomes. And so if you're going to go fly around a substation that belongs to an energy utility and you go crash that drone, you could take out power into an entire neighborhood.

That's not just something that's cute that your kids did on Christmas afternoon. It's actually something that would cause you to lose your job because real problems for the community. So the necessity for the person is not just to be there, but they have to be an expert pilot. They have to be great at being able to fly in these really chAllenging environments. And that was the premise of sky duos.

Can we build the skills of an expert pilot into the draw itself, so that any one of us here in this room can actually pick up the drone to be reasonable, proficient? And that become sort of democratizing in terms of their access? So we started day one from billion eton's y.

And the first determination of that was the ability to follow somebody. Since then, we sort of up our game quite a bit, not only beyond obstacle avoidance, but to be able to not have person be present at all. So going back to that substation example, there are now drones here in the barry that are flown from somebody that twenty, thirty, fifty miles away.

And those drones completely automated inspection missions. So we're getting there. The regulatory environment doesn't allow us to just completely eliminate the person. But we're certain tainan making quite .

a bit of polls in the automotive space, which is passenger guards at the lower levels of automation where you would think of as for the user hands on system, but you are still paying attention to the vehicle your hands are understating with. But the vehicle can do things like automatic emergency breaking or what card cruce. Those systems, I think, are available on sort of any car you go.

And by today in europe, some of those a Mandatory ory from a regulation standpoint to be there. I think U. S. Regulations is likely behind the Mandating some other systems, but generally those are available on the where the next focus idea for a lot of the o ms is on what's god side of level two plus systems, which are these hands of system.

So as a user, you to keep your hands on the stating, will you still need to pay time, you can keep your eyes on the road and one level behind that, but its eyes off as well. So you can read a book, you can watch a movie. Those systems aren't yet deployed.

That's in R N. D days right now, just very early deployments right now. And then the level force, I think in the city, in the o we all see of the sort of main deployment on the guard. And I think there are some deployment in china that we see on the level four side as well, taking a different industry. I think maybe construction and mining side, what's interesting is mining has had some form of trucks without any drivers in them since two thousand seven, two thousand eight.

It's not the same level of autonomy as way more in the sense as those structure are following up three defind parts, they get a lot of support from the infrastructure in the mine in order to do that. But as a clear business case and an R I for that product, enhance the mining industries being investing in that. The technology itself is far behind what's available on today in automotive. So so that is a big focus to updating that technology, answering ing more of the use cases.

Maritime, I would say that takes a lot of inspiration from the cellar in car community. And the reason for that is effectively when maritime, you have this kind of two dimensional plane that you're moving around, very similar to salt cars, right? You just have a lateral planning.

It's maybe little bit more complex. You have the lunch to know access as well, and you also have much less things in your environment to ground you. So when you're thinking about like an autonomy model, the world doesn't usually just like nicely translate around you.

Think about driving uh, a car in a nice kind of like city skyscrapers block. As you take a left, the world nicely transforms around you. But imagine you're in the middle, the ocean and you take a left, the ocean looks the exact same.

And so you actually get a lot less information per frame when you're making these sort of like vision models or you're making a sort of autonomy stack. And then the other things I wanted to draw is the delineation between a perception stack and autonomy stack. For a lot of these use cases, for us at least, perception has mostly been solved, that is to say, object detection, object avoidance.

These are like very, very simple tasks, and they can mostly be done with CNN, which are twenty, thirty years old. Autonomy is a bit tRicky because you actually have to take kind of elements from the perception stack, have to make them actionable. And that is to say you have to do a lot of long term planning.

You have to actually take in that data and make decisions based on IT. And that's something I think that only been enabled in last five. Maybe you can make the argument even two years. And so i'm super cited about a lot of the work that's going to reinforcement learning and long term planning because this is super, super important in the maritime demand because in maritime, you're not often near the shore and in the defense application, you're not often in communications with back home and see you need to make a lot of independent decisions. And so these are the things that we're thinking about in the maritime autonomy space.

Double clicking on some of those sort of technical breakers. Where do you think the latest developments in A I have impacted autonomy? To what extent of the impacted autonomy? And how much change are we seeing in the industry today versus, lets say, five or ten years ago just based on the current stated, they are in things like video language models or other new transfer mer based architectures that might not existed. For example, when we got started or one sky deo got started.

yeah I think pretty significant. The bags actually in terms of the latest developments, I think foundation models do have an impact on the architecture of these autonomic systems, on these downway systems. Where do you think of contest models as replacing a number of more tasks, specific models that existed? So we see more and more companies making the shift words using or researching with how foundation models can be used.

Uh, garden court frontier of research right now is end to end. Even when I joined them more back in like one sixteen, there was a lot of hyperon end to end driving. But eventually, martial zed, that technology wasn't ready.

So now it's almost a second version of that happening now where the research has progressed a little bit for the so I do think the architectures are going to evolve significantly even though they're not yet ready for production. What that also means is that the tools and infrastructure needed to support that is evolving pretty quickly. And so we are developing new kinds of simulators that are needed in order to support those advances.

Architectures and those simulators themselves require many of these generated techniques or neural rendering techniques that have shown good promise and research. And thought, I would say, is just using some of the general I to simplify work flows as an example. If you are not on the engineer, spend a lot of time working with simulation a, but we could generate them for you programmatically in a much more confident banner today than more was possible two or three years ago. So just had different layers, whether it's the autonomy stack, whether it's the tools being used to develop the autonomy stack, I think we see pretty fundamental in bax. And that's a very interesting prety aggressively in your glazing.

These technologies just stop looking on the idea of simulation for robotics or autonomy. I think this is like a very non obvious point to people who haven't worked in this industry for the last years because I think that we've started to use video generation models as drop in replacements for simulators like unity, aras or unreal engine.

And the reason you need to simulator, to be clear, as because, first of all, of robots are really, really expensive, and you don't want to take a match of the field to have them break, if you must be hr, than basically. And you also need many environments where you need to train on policy, meaning you basically need to run IT in real time in order to make sure that thing works. And so in recent years, they've started to use video generation models conditioned on actions or sensor data.

And you can actually create a really, really incredible reconstruct of the world just using these video generation models. And this is not hype geni stuff. This is used by tesla, is used by way of common.

It's pretty exciting stuff. And I think that some people are being very quiet about IT. But there isn't .

really awesome public releases. I will say the rise of the video has been pretty instrumental for our kind of business to be able to put that kind of computer power in a foreign half pound foreign tor that can be solved for basically eleven in two thousand thousand dollars. We would still make some margin, and that's a very hard thing to do.

I'm curious, how do you think about the cost tasteful when using the high performance ships that in radio comes out with? How does the sort of autoliv for the business change when you're putting the latest G, P, S are computer onto the drones?

It's so foundational to our business. We have to have high computers on the drones themselves. What we did in our latest released that we announced about a year ago was we actually added excess capacity.

We are based trying to future prove the hardware so we can do more over time. We haven't fully used IT. We expect to be able to run additional models, including. While that our customers build on the drone itself because they can do much deeper analysis on the things that matter to them.

what's an example of A A model that a customer might want to build?

So object detection is a really simple one, as you've mentioned. But in our world, we see everything. It's not just about hate people, right? It's like, no, I want to be able to identify that particular transformer.

By the way, P, G, S. Transformers are different. The california Edison formers with transformer, they know their transformer, actually identify the transformer.

We need to be able to determine whether the thermal signature is actually saying as a problem or not. That is so specific, weren't not the best ones to be able to build that. So that would be an example of how customers can build their own stuff and and basically run IT. So you can take a media action on the road itself. That's why we put in that access capacity ability to do those kinds of things that have that kind of extensively in place.

对过， actually that brings me to another question, which is for ironic and studio. Both companies are essentially building the full vertical stack where not just the algorithms around autonomy and perception, but actually the full kind of product that goes into the world or as applied, you're much more of a software provider working with companies who are building the hardware. I'd be interested to hear some of the kind of trade officer, what are you gain by owning the full enter and system and with harder as a result of having to manufacturer deliver the final product.

We think that the ultimate value of vertically agreed as stock, as reliability, we're very committed to IT. We started with IT, and in fact, our primary competition in the world is D. J.

I, the largest manufacturer. Jones out there. And they are the complete opposite. They're happy to build the hardware and there is a whole plaza of other companies to sort to build their ecosystem around D, J, I. hardware.

And we see that those seems actually started to fracture when you get more, more complicated scenarios. And so you add an autonomy there, what happens in the corner cases, what happens in the failure most, how the things react, that's where we could completely control that. And we can test for IT and we build reliability around IT.

So so I think the downside is that we might go slightly slower, right? Because we do have finite resources. We have to choose where our investments are. Was another company that may be all of what they do is to be able to build some kind of capability on top of another person's hardware for us. We are okay to move a little slower to be able to deliver a higher quality.

kind of like a weird middle road. And I would say because we own the full stack, we also though have three types of boats s like small, medium and large, basically. And these both are very different in terms of the amount of cameras they have, the actuation techniques.

Is that being said, we still own all the hardware, especially all the computer. And so the compute is mostly homogeneous between them. But you're actual kind of a factor systems or your perception that is going to be differentiated.

And it's also a differentiated in between boats. Sometimes because we want to test things out, we want to add another camera or something. And it's actually a powerful thing to be a little bit agnostic on the hardware and the software because IT basically pushes you to create really, really solid abstractions for your perception stack or for your autonomy stack.

And I kind of pushes you to be a Better engineer. This is at least what i've experienced in my engineer team. We really likes to think about, okay, how is this going to work on boat and plus one. And so that's been kind of a fun chAllenge to work on. But I do agree that IT does slow things down sometimes.

Yeah, I completely view at that point of abstraction. So applied started with being a tools provider. So you provide and and data products and eventually our customers are liked.

Tools are great, but we don't have the intern capability to build that autonomists data using the tools. Can you help us with that? So that's how applied as he started.

We now provide alerta omy stack. We also have a trucking stack. And to the point that beta was making now more so than ever in autonomy, I think these attractors are possible where we have a stack that it's not beautiful. A single customer is meant to be reusable across customers and they can build the eventual application on top of IT. So if you only want to use the tools, that's great. But if you want to actually authority stack, either as your primary stack or to get kick start your efforts internally that our providers like applied all other ones that you can use with that abstraction to significantly the accelerate your program.

How involved is that sort of translation process to different foreign tors vehicles customer types? I'm curious, ous to hear how hard IT is to get autonomy that works for one hip of vehicle to work in a totally different set up.

I mean, I would love to tell you that it's like seamless and just happens all by itself, but there's the engineering aspect of IT and there's the organization aspect of IT. I think engineering, to a certain extent, you can engineer the best A P S. And abstractions that are still customization that you would need to do, especially if it's for the first time being deployed to different platforms.

But the organizational chAllenges is even more interesting because for some of the companies that we are working with, they're going through this big shift of going from hardware companies because wakers were primarily hardware drive to becoming software companies. And I said that reinventing themselves, hiding soft engine, hiding engineering leaders from silicon vandy eta, and you actually need a very close partnership to make this work, especially if there's radius to end real phone factors. So it's not just an engineering problem. It's also how do you build up internal capabilities for customers who you often will work with them on, like training their internal team, except in order to be able to work together on this transition of the stack. You many different .

platforms makes sense. I spent modest mico before investing at plant here, so i'm very comfortable .

with that type of model. What was your experience that in terms of how to help companies of make that shift to whether it's tony y well, it's software data .

IT wasn't necessarily strictly autonomy. But I think that there are a lot of similarities, ties that the biggest thing was how you embed with the customer to drive cultural change, which is the four delayed engineer model really invented a entier where I would be flying out to as a bijan, or somewhere in oman, or train dad, or really far flung places to help their engineers really on the ground, in the field, figure out how to use software.

And so maybe switching gears a little bit, i'd love to talk a bit about the economics of autonomous systems, also how that relates to regulation. So we've heard and read in the news a little bit lately about way mo and the economics of a way to ride. And it's really exciting to see IT start to actually make sense so that we are not just pouring money into a system for the sake of our N D.

But every single way more ride is actually, I think, profitable if you don't count the overall cost to the vehicle, which is a big cost to discount. So i'd love to hear maybe starting with sky deo, like how do you think about the economics of the drone industry where you've seen real success, where it's been slower? And maybe also how regulation has play a role in that.

We serve multiple different industries. And there R O Y calculations are going to be slightly different. I've used a couple of examples where we talk about energy utilities, and in that instance, sits about looking at the people in labor IT takes to go inspect to the equipment, whether that physical substation or transactions, es and distribution lines.

At any time, you can save the cost of actually setting a truck on site where they can put a person into bucket, get up in the air, which can cost two thousand dollars for deployment. And you can instead, but to drawn up and get the exact same information in sixty seconds, just sort of a pretty obvious ri there. We also serve public safety like first responders, police departments.

And I think you can't really put a value on human life, but you can put a value on the insurance payout. Ts that come with officers and both shooting IT can easily go to one to two million dollars every time there's some kind of use of force. And so being able to avoid that, being able to cut that down in half is really, really key. Ultimately, we're in the business of providing information for people to make Better decisions. And if you're in the really high stakes scenario, Better information about where the thread is can really make a big difference in terms of what that you end in a peaceful way, our tragic way, certainly some elements of economics, but a lot of IT is just keeping people safe to keeping the officers safe, keep in the community safe across the board.

And how about with the regulatory piece of sort, whether an Operator has a line of site on a drown, how maybe take a particular use case like police officers or something like how they think about kind of fleet expansion as those regulations change?

One of the key areas of investments right now, a constant john's first responder, where you would preposition docking stations on the rooftops of cities, and this is happening in new york, is being tested in other cities where when the nine one one call comes in, the drown automatically deployed and goes, is effectively the first responder, the first person on site, and is providing eyes back to a remote pilot. That does two things.

One is the pilot is sitting in some nice condition office, and they can immediately relay information back to any responding officers so they can respond more appropriately. They can control many, many drones at the same time. So IT becomes a one to many versus or one to one argument.

And a lot of calls can be cleared. These nine one, one calls for service can be cleared by basic of the drone saying there's no real issue here. You don't have to go outside.

That's tremendous economic savings. In order for this drone is first responder to happen, you have to have a regulatory environment that allows for a lack of visual observer on the roof top. And right now, just the twenty second primary on F A regulations, if you're putting anything in the air, IT is regulated.

The F, A, A cares about their space. They most care about just not hitting anything, don't hit an airplane in helicopter. And so they require you to stay below four hundred feet.

And they require a person to be visually looking at that drone at all times as like the both both basic rules. So now you have a scenario where you might go further up. You don't have available staff to put a personal they've talked to be able to look out. And the second is if you're in some place like new york city, the top one hundred buildings are over six hundred feet. So how are you going to be keeping everything below four hundred feet to being able to sort of work with the F.

A, A, to be able to create these kinds of wavers to this traditional relator environment that allows for on visual line site, beyond visual line of site, without a visual observer, and to be able to actually stay within fifty feet of structures so you can actually go up and over a building as long as are within fifty feet. And that was a way that just came out, actually two, three weeks or in york city. And it's a fantastic way of having the F A sort of work with these agencies to do this right for the community whistle .

keeping the exciting congres.

Do you want to Better understand complex problems and ideas like how to use A I to predict and health, or the mysterious of chat? P. T, well, then you should definitely check out what's your problem?

The popular podcast that explains the problems really smart people are trying to solve right now, posted by farmer planet money host jack up gold steam. What's your problem? Features entrepreneurs and engineers.

You talk about the future that they're trying to build and the problems they face on the way to go home. follow. What's your problem? Whatever you get your podcast.

I'd love to hear the applied intuition, the economic case, where you see where is this still sort of a slightly money losing R N D exercise versus where are you actually turning to see the economics make sense and where, if at all, regulation plays a role in that? Yeah.

I mean, on passenger guards, they've had driver assistance systems for a long time. As these systems become more capable, o EMS have an ability to charge more for them. That's the F.

S. D. System from best lament, over ten thousand dollars .

as an at test level or driving user.

You see you see yeah you there is a certain willingness to pay if the performance of the system is up there now just as discounted that. So I think the industries is in a Price finding my stage right now where obviously the best thing would be higher performance is while keeping costs the same.

But as the cost of computer goes up of more sensors under the vehicle, more soft deploy on there, because the bomb costs for the O M goes upsides, there be some Price increases. O E, M already Operate on somewhat ten margin compared to the margin we we are used to in the silicon valley. So they're not in a position to keep losing money on each of these systems.

I think the market will eventually find the Price that make IT profitable for the time to ship these systems on the, I think areas like truck and construction mining have a different, different economics of theme works. So if you think about trucks, for example, enertia regulation into this as well, we can think of unit economics as reducing the current costs. But a big part of uneconomic also.

Can you read the reviews? So the first argument is here is a shortage. So we are losing like supply in flexibility.

So like we do some work in japan and providing Donna astruc technology japan, the government is actually pushing the commercial wakes sector to rest modern autonomy because they're seeing, they call IT the twenty twenty four problem, where the drivers are engaging out and they are overworked and has issues. So the shortage problem is real. But even if you look beyond the sort of, okay, how do we reduce the labor cost and the insurance cost?

If you have a network, autonomous trucks, you can actually optimize gissing's network such that you're generating more revenue. But truck, which makes you an economics Better. Where regulation plays a part in that is too, for one is in a place like the U.

S. Each state can actually regulate these rival. Less trucks are deployed. And you can have this weird mix where the states that you are trucks is travel through have slightly different regulations and federal government has different regulations. So you do need some consistency. So if you, anna, have a drug grow from l to atlantic, which is actually a pretty heavy free through and california has to have consistent laws with that is on an texas that allows that.

And I think it's not fully that IT, but getting that the other side of IT that are regulations like hours of service that determine how long a truck and a truck driver can drive, that's for their safety from fatigue at trine for road safety. So the government has be willing to be flexible on those such that those trucks can then Operate twenty four hours such that you can then reconfigure the logistics networks and then make the supply and more efficient and make the revenue and the unit economics work. So I think those other factors in trucking and mining, I think a different sort of set of factors in the sense that the ottawa is somewhat or clear the industry why they been investing in for long time and the drugs we're talking about mining on which these autonomic systems are being deployed, just like a five million dollar machine, right? So it's not a big deal.

You put a hundred thousand dollar light on or of course, that has hundred thousand dollars that's meaningful. But in relative to a five million machine, that's not the main point. The main point is if the system stops and IT stops the mining Operations for every minute you stopped the mine, you're losing them and and hundreds of thousands of dollars. So the unit economics game that is at a mine level, not necessarily at a drug level.

I also imagine replacement cycles are different, is not as easy to buy a new five million dollar bulldogs as IT is to buy a new hundred thousand dollar car.

That's right. But the whole auto I girul tion becomes a slightly different girl lation, which is how many trucks do you deployed a mine? How do they have pack the productivity, which increases the mind revenue and how do they impact the downtime, except so we see slightly different applications. But I think generally in mining, as I said, o EMS and the minds we talk to see the business case that is, is improving the technologies that, that can reliably work bring doing down to the mines.

So I think related to some of the questions around regulation, i'd love to talk a little bit about geopolitics and some of the kind of regional or global differences in our approach to autonomy. I don't know if this is a true story, but I did hear the story about dji drones being used early on in ukraine, and china had given russians a back door to be able to determine the positions of ukrainians on the ground using D J. I.

Drones, which was a big imp test for a lot of the drone activity that we're seeing in ukraine. E, coming from the U. S. So there's a lot going on in this kind of geopolitical landscape. And I kind of love to hear how for all three of you actually given sonic as a defense contractor, i'd love to hear how kind of the role of geopolitics and how you think about autonomy and how it's come up both as maybe a chAllenge and also a motivating factor.

So sironi was basically founded on this idea that china outnumbers are should molding capacity about two hundred. And one is, the question is, what can we use as a sort of unfair vantage to leap frog that, rather than just compete on the basis of cost per ship, how can we make autonomous distributed systems to literally sort of an overmatch result over china's kind of chip link active? And so this is obviously played a major role in motivating sonics, the company.

But sort of even beyond that, I think we've seen a lot of unmanned of vehicles, which is about onest votes be used in middle east as well as with ukrainy and russia. And so I think the sort of heaving up of global proxy wars, not just great power conflict, has also LED to a lot of expansion of use of autonation systems as we see in ukraine, as we see off the coast of yemen and other countries like that. But yeah, that's the T.

O D, R. And then given this sort of D J I angle here, would love to hear how guy o thinks about .

geopolitics is a very important topic in our world. Imagine ukraine. We've been in ukraine thirty times out there.

This past ring. They use D J S in a very disposable fashion because it's the cheapest they can get. They hack IT. This is effectively a hot store vender to them.

And so they have to hack IT quite extensively so that the drones can actually fly there and not be detected by the russians under certain techniques. They have to do that. One of the chAllenges that the us.

Looks at is the threat of being wholly dependent on chinese based technology for industries that they would considered to be critical to national security. What ukraine has shown to the world is how critical drones can be in any combat. And of all of our drones and all of our drone parts are being held by chinese manufacturers, then we are wholly dependent on them for such a critical technology.

And that is exactly what the U. S. Is trying to avoid. And so there's laws being discuss within congress now to start winning the united states off of chinese manufacturer drones. So IT won't be an immediate ban and will will happen overnight, but they can sort sort of lesson interdependence and start encouraging us and us elevators to step into the full, including ourselves.

We're not asking for subjects, but we do benefit from restrictions for sure that for some of these organizations, like the federal government and state local governments, to choose western products instead of chinese manufacturer products. And one small example of the kind of power that somebody manufactures can have out, which is a sort of a smaller version of D, J. I.

Based on the exact same city in china, they decided one day that they would update their geo fence, basically where you're allowed to fly or not allowed to fly. And they turned off all access in taiwan. So immediately, every auto le dro was just correct in taiwan.

Why taiwan? So if you had an example where all of our critical industries are using a chinese manufacturer drone in one day, china says, i'm just going to turn them all off. They can immediately break him. And that's really the extent of the problem. It's quite, sir.

yeah. I think china is a conversation with every C D R C O of any major O M across industry that visit our office are we talk to. And these are like weekly conversations just to taking example from automotive and maybe government.

The applied also does a to work with the government in automotive. China is basically redefining the industry. There's no outset of light the way to put IT.

There's a few things that are happening and the beijing order to show was earlier to see a IT happened to be that in person for the all of the attention was on the local chinese o ms, like you could literally go to the boots of the international companies. Consumers actually have much interested in them, except for a couple of brands. And that's because the product innovation in the china ecosystem is second to none right now.

The cars you can buy that like buy not just driving a prototype, just vastly superior d consumer experience compared towards you can get in the U. S. Or europe.

I was talking into staff earlier about some of experience of walk up, bark yourself and bar opens the door for you. You sit in the god god system and process commands from four people docking Sidney ously. It's a delighted experience to sit on those these these production guards.

And it's not just about the product today. It's actually scared how fast the base of innovation is every six months, the innovating on that product and the cost at which they are able to do IT is very hard for the global industry to match. So at one point, you I think, well, there's enough geopolitics that maybe the global economies like somewhat united from that, and we're see that in the us.

Percent vehicles. At the same time, we look at the what's happening in the market in china, these o ms are facing a sort of flattening domestic demand. They have oversupply.

So even though china's the largest automotive market already and the largest exporter, they all have ambitions of sort of going global, right? And some of these ims have costs points that are solo like a byd that despite database in europe and despite that of in U. S, they might actually be somewhat competitive.

And that's why of the automotive executive is worry about china and teen teen about what's our strategy, not just in china, but what's our global strategic given what we are seeing in china, I think on the government side is similar. Honestly, the sense that the D O D used to be the driver of innovation, right, is a lot of dies between the silicon valley history and darpa, and how autonomy came about and how a number of other technology is, like even internet came about. And now where at the point where I think the pen taga and recognizes that it's been slow in moving towards software and moving towards autonomy and definitely slower than china, both in terms of sort of processes of like how do you do procured is actually a big value today.

So we do this conference every year in D. C, focus on national security. And the entire conversation is about how do we move faster and bring these technologies that are being deployed in the commercial world to national security much faster, and how do we have evolve about procuring processes to be able to do that across industries. We see china being a major part of conversion, a major driver of strategy for the companies.

I was gonna a end with the last question of what keeps you motivated, but I feel like that very motivating. So with that, I think we have a few minutes left. So I wanted to just open IT up if anyone in the room had any questions for our wonderful guess.

Alright, since this was a live recording and some of the questions were pretty long, we've condense them for your ears and i'm going to be punching in with the voice of her. Are IT first question, what are your experiences dealing with security guidelines and regulatory bodies?

The one thing that out says it's a little bit uninsured sometimes that designing autonomic systems, as long as they're not like true connected factories, sometimes it's actually lower risk. The example that I like to give a lot is imagine designing a drone versus a helicopter, right? helicopter.

You got to be pretty sure that that thing will not follow to the sky, right? I drown like falls out the guy, okay, that's fine. Iterate will keep going. And so sometimes IT actually helps me sleep a little bit Better at night thinking, okay, at the end of the day, these systems are directing themselves if the boat sinks, not the end of the world.

And so that's what all add to IT in terms of the actual reality, the fact that we're designing for the dod, though, yes, these systems have to be absolutely like secure, mission critical, especially terms of reliability. And to this extent, I think sonic is an engineering world is built up around this, like the way we do our systems engineering, the way we do our testing evaluation pipelines, I think is a lot more rigorous than a lot of other companies. But at the end of the day, you can sleep a little bit nicer just knowing that their unman systems .

from product design standpoint, in terms of safety, a lot is into basically failure mode. So what happens if the battery gets too low, no matter what inputs the Operator gives? The drones is gone to come back, and it's onna follow the path that I had so IT can certain navigate back from obstacles.

The matter where is out to get back to its original location, IT can also base good land directly down. There's a programing of safe landing zone. So these are all things to try, keep people sort of safe.

And terms of more security, especially with the department of defense, we sell a variant of our products that are effectively offline. And offline is a bit of a mno where they are online, but they're completely within the private network of the dog. So we have no access to IT. It's not cloud based. And those instances where if the customer really, really lenise that level security they have IT and IT basically means its engineering work for us because we have to Carry to various forward, but is necessary in our final government a generally.

instead of the automotive real man sort of veo cars eeta. What I was says that what's been done in the industry in the past, like all of the systems engineering practices that have been used to build the aircraft, like commercial airline or necessary and are used but not sufficient in the same way, some of the standards that have been put out I S or two, six to six two, that has been used in the industry for a long time. These are all inputs, but none of these, individually or even cumulatively, are sufficient.

And so if you look at some of the publications from what we customers can, customers, we actually need to do way more than what's actually any regulatory bodies asking for, or what I O space engineers has done in order to prove the safety of the system that actually validating that these systems are safety, be deployed, whether that's a passengers, there is a truck, whether the mining, agriculture, defense application is actually one of the problems, our customer struggle the most, because there is no standard blue. And for that, it's a combination of an engineering problem, plus a data science problem, plus idealities problem, lus winning consumer dress problem, because eventually, to convince you yourself and the stakeholders that this is k enough. So it's not a science problem.

But in many ways it's still the blueprint is yet to be clearly written for this. And I just don't think that regulations and standards are enough and the same applies for siber security. I think they are very short of what actually needs to be put into these systems.

What best practices do you use to reduce stress in your technical debt?

If you ask any sonics software engineer what the three magic words are, they'll tell you this, the kind of design principles that we follow is simple, correct, fast in that order. The idea is that, first of all, you want to build a very, very simple system that you understand from the electronic l. We do not integrate any technology that we don't understand from the electron level all the way up to the soft level.

And so the til D, R, that we build this ton of stuff that doesn't work at first, but that we understand super, super well. And so the systems design, like kind of physical y, pervades basically everything in our software respect. So yes, the simple, correct fast is like the way we do at at ironic. This is totally worked from. I been in.

never heard of technical dead outside. So I think technical that is inevitable in fast changing industries like the ones we play in. So I didn't really, really talking about how the underlying technology in the eventual systems is changing, which means we have to rethink the products we are providing to the market quite a bit, right? So you keep the current system stable and then you deploy a new system magically over dying.

But we've all been doing engineering for a long time. We know that it's never as clear cut as that, but I think we focus a lot on working with our customers and go we are going to actually upgrade the architectural of this product because now we can actually make use of general or any other technology or to improve these products. And surprisingly, more often than nod, if you have a reasonable road map, the customers are you pretty reasonable.

IT doesn't require you to have a trusted relationship that is not like money. Where are just purchasing based on certain features and whose the most cheapest to procure, it's hard to have that kind of relationship. But for the industries where and where software we are providing them, either that the developer tools or the actual software that's going into the wicked, this is cutting edge software and they know this is getting at software. So it's very like collaborated relationship with the customer. We are often building new products with them, and that allows us to have sort of these heartful road maps .

i've been intact twenty five years. There's always technical death that larger you're around the mory piles up. I don't know that I could add anything more. The one thing that's a little bit unique about schedule is that we have these hardware platforms and that allow us to basically sort of reset like the latest generation to be announce last year was three years and eighty million dollars in the making, and that allowed us to basically resettle lot. But we also have a pretty large cloud software component and there's a how of a lot of technical death has been built up over the last seven, eight years. I don't have a magical answer .

on now on some .

people think to get to solve driving level four or five or robot taxi statuses, we don't need lighter. What's you're take there?

alright? I can take that super easy question. So I think generally one is somehow the industry has gotten into this yes later, no later binary gams. But at the end of the day, uh, sensor is just one other piece of technology that you incorporate to achieve the ultimate goal of, in this case, level for roxi.

So if we all had a path to deploying the vehicles that you see on the road without a lighter, I get into them would have taken that. But because there's a lot of engineering and science inventions that building some of the lids that we have has, they really are like getting as riders. And so we must not building them for the sake of just one to a work on these technical chAllenges.

There's a portion of the safety case that goes back to the question of how do you certify that these are safe enough that relies on certain properties that only let us have that today. The other combination of other sensors cannot and even in cars and like china, for example, you see lights and lower levels. Tony y, so a that allows them to sort bring certain capability market faster. But also from a liability perspective, you have to really think from a manufacturer perspective that the vehicle does get into an accident and we will see accidents happen in the industry as sort of the technology's method.

Would you rather have gone the olive and that whatever was possible from a safety and liability perspective to prevent that or or not as a question that o have to face when they think of our, should I put to light up in a passenger card or not? So ultimately, will we get to a point where systems without riders are good enough to function in their respective design domain, that respective design drain could be mining the respective design of in could be cars, that could be some defense applications except up? I think in the long term, we will get there.

But do I think that at the early stages of deployment of technology, taking the safest path is the right approach? and. Order to build up trust in the technology, in order to build up trust with residents of a city, in order to build up trust with regulators. I think that's a pretty reasonable approach.

I'm super curious to hear your observations around the ecosystem to are seeing emerging. I'm sure that's also really interesting from venture perspective, right? How do you see the next couple of years coming together there?

I happy to take a first cat at this. So I think it's a really good question at something that we think about a lot historically. I'd say the industry has mostly focused around, with a few exceptions applied being a great example of an exception schlei I being another one that both businesses really took off as a result of selling software into the sort of self driving early on, at least into the self driving market.

Most autonomy companies for the last decade have been very tight. Coupling of hardware and software where all of the software was written almost to the firm are level for the specific hardware form factor. And what we're starting to see, I think part of IT is like the last three years of developments in A I and part of IT is the broad explosion of interest and autonomy across a lot of different use cases.

Is more of a software tool chain emerging IT. But there are a lot of unsolved chAllenges that we see pretty consistently across companies Operating in autonomous spaces. So even like basic assumptions about how software is built for the kind of traditional business world of the twenty tensor doesn't work in the tononi space.

You often don't have life like you can expect that you can just ship releases like every five minutes to customers and they'll be able to update how do you stream data from one place to another? Like how do you decide what decision making happens on device, on chip, on the edge versus like in a cloud environment because he has to consider a swarm or a fleet of devices. So these are all like really interesting technical chAllenges and problems that I think developers will start solving more at scale in the next several years as more and more autonomy use cases come up. So I imagine that the kind of developer tool kit around building an autonomous system is going to be more decoupled from the hard work itself over the next several years. There's lots of opportunities to build new tech in those scenes coming up.

From our vantage point, autonomy is what democratizing access to the drones, but it's not what people buy. They don't care about our autonomy. They Frankly don't really care about our drones. They care about is the information that the drones give them. And the easier way we can get that information, the faster way we get that information.

Our next level of autonomy and the broader road map is gonna less about how do you fly and more about what mission are we trying to accomplish and how do you actually, according IT, across multiple drones, to be able to accomplish that mission the safe way in a fast way. And that because much more vertically specific and use case specific, and that's really to Oscar, like the gold is, and we will create tremendous rentier. Our hope is that there is actually a very thriving drone industry in the united states to be able to combat the chinese manufacturers also extraordinary ly capital intensive, and that's hard, does a big barrier. He raised a lot of money and we're the large manufacturing and says, but we saw a long, long, long way to go.

I hope there's a sort of proliferation of a number of players coming in and contributing because I think you will be overall best for consumers is in building applications for the vehicles. So today, for the most part, the cars we buy in u in europe example, they're not delightful consumer products like when you brought your first iphone and iphone six seems like the fifty and teen.

So there is really not been a great consumers product in that sense for a while. But when you go to china and you experience those, you get some of that wow, and that moment of the light back and then you peel the layers and say, okay, why isn't this exist in the cars here? The reality is you peel back the layers behind your car.

Today, there's one hundred and fifty different suppliers that each provided a small ecu as mini computer. And the, oh, i'm integrated that in two sort of a function experience, but you almost have to redesign the car from the ground up. And that's what tesla because they could start fresh, they could start with a software engineer dog. And that's a journey setting every single vehicle that moves is on today.

And that's why applied to sort of providing the Operating system because they began the Operating system and give you a nice as d applications on p you can only see you think of the god space and what experience where do you want to that way? Is that something as as accept? And so that's what I hope the industry goes and making. Overall, we just up.

All right, that is off today. If you did make IT as far, first of all, thank you. We put a lot of thought into each of these episodes, whether it's guess the calendar touches the cycles with amazing editor Tommy, until music is just right. So if you'd like what we put put together, consider dropping as a line at rate this podcast com splash a exciting cy and let us know what your favor episode is. I don't make my day and i'm sure tomes too will catch you on the flip side.