Episode 172 - Making Real-Time Mapping a Reality

Long thought to be impossible, real-time mapping is now a reality. With the ability to update street information every five minutes, this technology is revolutionizing how road data is captured and delivered to the mobility ecosystem.

Leading the way is Nexar, a computer vision company with a portfolio of AI-backed dashcams and vision-based applications for public and private sector partners. Targeted to OEMs, AVs, and DoTs, Nexar uses crowdsourced data and edge AI to identify work zones, speed limit changes, open parking spaces, and provide more efficient routes for the public.

To learn more, we sat down with Eran Shir, CEO and Co-founder, to discuss how Nexar is making roadways safer by using footage collected from the company's billions of AI dash cams and analyzing street information every five minutes.

Meet Our Guest

Co-Founder and CEO, Nexar

Previous to founding Nexar, Eran was an Entrepreneur in Residence at Aleph Ventures in Tel Aviv. Prior to that position, Eran co-founded and served as CEO of Dapper, until it was acquired by Yahoo! in 2010. After the acquisition, Shir was the Senior Director & Head of the Creative Innovation Center, a global team within Yahoo!, based in part on Dapper's engineering team. Before co-founding Dapper, Eran founded the DIMES Internet research project, and Cogniview, a machine learning startup.

Apart from his entrepreneurial career, Eran has also served as a Board Member of the Israel Center for Excellence through Education for the last 25 years.

Eran has a B.A. and M.A. in Theoretical Physics from the Technion Institute of Technology and doesn’t have a Ph.D. from Tel Aviv University, since he bailed out to become an entrepreneur.


Grayson Brulte:

Hello, I'm your host, Grayson Brulte. Welcome to another episode of SAE Tomorrow Today, a show about emerging technology and trends and mobility with leaders and innovators who make it all happen. On today's episode, we're absolutely honored to be joined by Eran Shir, co-founder and CEO, Nexar. On today's episode, he'll discuss real-time mapping and how AI and edge technology can enable safer AV driving. We hope you enjoy this episode. 

Eran, welcome to podcast.

Eran Shir:

Happy to be here.

Grayson Brulte:

I'm excited to have you here because dash cams will play an important role in fleet safety. It's been documented with data, it's been documented by insurance companies, so point blanket will play an important role. But I'd love to know, in your opinion, why are dash cams important and what problems are they solving for drivers?

Yes, there's the simple problems if you're in a crash or an incident that can help you get out of that but overall, what are the problems they solve for drivers? 

Eran Shir:

Let me tell you a personal story from last week. Unfortunately, I was abroad traveling with my wife and all of a sudden I get a text message for my camera, for my next camera saying my car was involved in the collision, right?

So it was my daughter driving with my car. Middle of the day I'm abroad. I get a text message, I get the location of the collision. I get a link to the livestream. I tap it, I start getting a livestream from my camera of the situation as it is. I can start I'm starting to talk to my daughter and help her manage the situation.

20 year old, it's her first collision. She, she's a relatively safe driver. And I'm helping her with that. At the same time, our insurance agents get the same data, right? Like it gets the video of the collision and all the analysis. Within a minute he was on the phone with her, helping her with towing and with the, with all the red tape, what documents to collect, how to capture the imagery of the other vehicle, et cetera.

It created a lot of peace of mind, right? Think of. In a different world where I don't have my next up camera, I'm driving, like I'm getting a call from my daughter. She is, crying and stuff. Like she's been in a collision. I don't know what's what. So first and foremost I think. High-end cameras like next to the camera, provide you with peace of mind.

Peace of mind when you drive, when your loved one, drive, when your employees, if you're a fleet drive, or even when the car parked because we have also. A sent mode, like a parking mode. When the car basically is parked the camera detects break-ins and collisions and sends me a notification and livestream and all this kind of stuff.

Technology today allows for basically creating a bubble of protection and safety and peace of mind. It is imperative from an evidence perspective. It's great from an insurance perspective. It reduces. The operational cost, it reduces the headaches, and it reduces the cost for you as a driver or a fleet manager.

And so these are just a few of the good things that involved with installing a camera, buying camera, especially a connected camera, right? Like the high end connected cameras the end of the day. Ring, simply safe. All of these kinda camera devices that protect your house the car is just another room.

In your house that you wanna protect from, but it's a really complicated room with many more hazards going its way than just, burglary. So you, I think it's really essential today and you don't wanna be the last full standing, not having camera when everyone else have their point of view represented.

Grayson Brulte:

What you described, I'll use the term safety blanket. Here you are, you're a father, you're traveling abroad, you're getting a realtime notification. You're able to share it with the insurance to calm down a stressful situation that your daughter's going through because you, the way that you developed and designed the app and develop the technology.

Was the app always part of the strategy of Nexar? Was it always part of integrating that, a situation that you described? 

Eran Shir:

Yeah, actually the app was before the camera. Oh, yeah. Because when we started Nexar connected cameras were not available yet. The, there were no one had connected cameras, so we said, okay, we know how to produce a connected camera.

We'll just use a phone. And we built an app that was like a virtual dash cam, and we played with that for a couple years. And while we were walking on convincing Zen, some people that actually build stuff to build us a camera that will meet our standards, meet our requirements and so it was always a combination for us, especially on the consumer side of camera and app.

And now, already 2023, we are able to build and sell standalone cameras that are connected with their own cellular modem, with their own AI chip, all this kind of stuff where, we have the next one lineup of cameras that are very capable. They're like an iPhone 12 or iPhone 11 pro we from perspective of AI processing and edge processing.

Now we can have it as a standalone connected also to a web dashboard as well as the app. But for us the app was important to create a co holistic, cohesive, kind of modern experience that, connects the camera, the user, and the cloud to give you a full experience. 

Grayson Brulte:

So today there's hundreds of thousands of your dash cams in use around the world. You're gathering real world data. You're gathering data from incidents, unfortunately, the one you described, which allowed you to create a product called City Stream Live. That's enabling real time mapping. Yeah, that's really interesting. Can you talk about this product and why it's enabling you to create real time mapping, please?

Eran Shir:

Yeah, of course. So basically like I mentioned, each of our cameras is basically an AI device. That on the lookout, it gets the video stream, the row video stream, and the sensor of data. GPS, accelerometer all this data every second, every actually fraction of a second, right? 50 times a second, and it analyzes it, right?
It runs edge AI models on those images on that stream of frames. Detect things. It detects every sign in America, every construction element, construction zone, element, every pothole, every parking space. We detect many types of stuff. Something I think like 70, 80 different classes of objects on the road, and we capture images.

So if you probably know Google Street View, right? Yes. And if you go to Google Street View for your neighborhood for the, the street nearby, you'll see images there from a year ago, six months ago, right? That's typically the, or two years ago. In some cases, when you go to the same product at Nexar our virtual camera that is part of our real-time mapping platform, you'll get images of your street from an hour ago.

Or a day ago, right? That's the scale that we're talking about. We see all of America every day, and that's a game changer because it allows you to understand change. It allows you to understand well that stop sign has been transformed by a traffic light, or there's construction now on the 101.

So maybe I should go to the two, through the two 80, or where is the nearest parking spot to my destination, or my favorite is how long is the line to Katz Deli in New York? 

Grayson Brulte:

That's a very important one.

Eran Shir: 

That's an important one for me because I'm a foodie. That's I can get into the VI camera and literally minute by minute, every three minute or so, see the updated line for Katz Deli.

So the use cases are just immense. You can do this for rerouting around black eyes. You don't wanna go through that road because right now there's black ice on it or there's snow, and I can just go on and on. So this is a revolution in what we expect from maps. What use cases. It's unlocked.

And an important element here is that it totally changes the economics of mapping and dealing with the physical world because if you are a Google or an Apple, you are spending like a billion-dollar year in updating the maps, right? And it only up updated once a year. Now we're talking about something that is.

A thousand time cheaper that you can update every hour or every day. So imagine all the new use cases that we can unlock, that we can open up both for operations, both for cities and states, for OEMs, for AV companies. And the list just goes on. 

Grayson Brulte:

I have to say about Katz, don't forget to tip when you're at the deli line. Cause then you, and you get the taste of pastrami before you get there. So if anybody's going, that's a tip. Always tip when they have the tip jar because always you get a taste. 

Eran Shir:

Yes. By the way, I was there last summer. We, I took my daughter, the one that crashed my car last week. I took her to try it out and see if she can finish up a sandwich. That was my dare for her.

Grayson Brulte:

It's a herculean task. It is. It's a big sandwich. 

Eran Shir:

It is. It's part of adulting. 

Grayson Brulte:

And it's fantastic. After all, Harry met Sally there, so there you go. There you go. When you're developing this technology that you've built, and hopefully you're having a great pastrami sandwich on rye when you develop it, what role does Edge AI play in allowing you to get this real time app? So you're getting the Katz Deli data in real time. 

Eran Shir:

It's an indispensable role actually. This is the foundation insight that. Made us start Nexar to begin with, right? Because me and my co-founder Bruno, we built a machine learning technology, 10 years, 15 years ago at Yahoo. Yahoo bought my previous company. And through that I met Bruno. Who was Yao's chief architect. And at Yahoo! we built machine learning and AI systems and in the cloud in data centers. And then our insight back in 2015 was AI is coming to the edge and that's gonna change the game. And if AI is coming to the edge, you can do intelligence in real time.

Okay. And mobility. Was the most important human use case, right? Like when you think on the human condition, where does real time play a role, right? So if I say something wrong to my wife and I don't, fix it quickly that's a real time use case. But buying that, driving is the place where seconds matter the most and where you want to inject as much intelligence.

So we were like, okay, if we could. Get a j I to understand your environment and the same way crowdsource the environment of everyone ahead of you. We could actually prevent collisions, right? Because what, what happens in a collision scenario? Why cli? Why do collisions exist? Then? One fundamental thing is they happen because of lack of time, right?

If I had five or 10 seconds more, 20 seconds more. I get a heads up. There won't be any collision, right? But the problem is that, two seconds ahead of me, there's a truck that blocks my point of view, right? Like I don't see anything beyond the next two seconds. So that's my horizon. But if that truck has a Nexar camera and that Nexar camera detecting or the car ahead of it, then normal chain collisions.

If there's someone coming from the side and into an intersection running a red light and someone captures that and we detect it through networking, we can update that within split second, within five milliseconds. And you get all of these alerts that could really protect us. So our vision is to really blanket cover the roads with enough eyes so that we can give you that time.

We can increase your time from two seconds to 20 seconds and eliminate collisions. And that was why it played a big role. The other element where it played a big role is with economics. Because at the end of the day, running AI is expensive, right? If you need to upload billions and billions of images, billions of hours of video, store that and process it.

That's a lot of power. That's a lot of money, right? Our approach was, forget about that. We will push AI to the edge. Which is basically zero cost because every comer does a bid and it doesn't cost you extra. And all of a sudden I can give you a nationwide map that even Google cannot afford because it doesn't cost anything because we centralize the problem.

So Edge AI I is is critical for us both from a use case perspective as well as from a unit economics perspective. 

Grayson Brulte:

The alert is really interesting. Then do the individuals in your network, are they getting a real time alert, collision ahead, black ice ahead? Are they getting real time alerts pushed to their phones or how is that alert system work in real time?

Eran Shir:

Yeah, so we actually have a standout that we open up. We open sourced a standard called Nexagon. That, that has been approved as a standard by the internet standards body. It's called the I E T F, and it's part of a consortium called Automotive Edge Compute Consortium. And what it does is it basically creates a virtual IP server for every tile of road on the planet, right?

We basically break the planet into hexagons. That are like different sizes and each of them gets a virtual IP address that you can subscribe to. You subscribe to someone on Twitter, right? And then you get into a neighborhood, you can say, okay, I wanna be, and you, your car can say, I wanna subscribe to all of the hexagons in the 200 yard radios, and we will curate a feed of everything that's new.

Everything that's happening, everything that notable in that area. So you would know about a construction that is 300 yards ahead that blocks the right lane and continues for 50 yards, right? Or you would know about, okay, there's a stop sign here, or there's a truck there. All this kind of stuff are actually available through our system.

We don't yet commercialize alerts ourselves. We give it more to third parties, especially OEMs to play with. But that's something that I expect us to also commercialize to the broader network in the next couple years. 

Grayson Brulte:

So is it very similar to ways where they say traffic ahead, speed traffic ahead?
Is that a very similar analogy for listeners sitting here? Okay, this is interesting. Is that very similar from an analogy standpoint?

Eran Shir: 

You can think of it as the marriage between Waze and Tesla. So you know, like when you drive a Tesla, it detects all of the things. You see them, like the cone and things like that.

Waze doesn't detect anything, but it crowdsources them by with humans as the mechanical Turks adding the data, right? What we do is basically bring these two together. We say, okay, we don't need human because we have the eyes on the road and edge AI to detect all of the things. So actually you can expand beyond the three use cases that Waze gives you.

You can expand to limitless number of use cases, and you can use the real time networking. In order to bring it all together and share it with the network.

Grayson Brulte:

That's wonderful. You're gathering over 400 million miles of road data being gathered each month on public roads. I wanna repeat that. Over 400 million miles of road data being gathered each month on public roads.

That's a lot of data. How are you using it today? Using that to. Enhance your maps and enhance your AI algorithms. What are you using all that data to do? 

Eran Shir:

Yeah, all of the above. We use it first and foremost for our real time mapping suite. Follow of those products that I mentioned, all of the, okay, what's new around the corner, which signs are new?

Where do I need to patch a pothole, et cetera, et cetera. And then we use it we use it also to train our models. So if I want to introduce a new use case, let's say for next Christmas, I want to tell the tech people on the road, Walking on the sidewalk wearing Santa Claus clothes.

Okay, so that's an important use case, almost as important as Katz Deli and, but it's simple for us because we have these trillions of images that we can kinda use for training, right? So you can find those use cases for whatever use case you're interested in. Then the third thing that we use, the data, especially all of those corner cases, collisions, heartbreaks, et cetera.

Of course anonymized. All the data, I must say, is anonymized. It's important we actually use it to train autonomous vehicles. And that's something that is really exciting, really cool because, Tesla claim to fame is that they collect data from all over and use that to train their autopilot and their.

Claim is that, they have more footprint than the regular AV companies, but we have more footprint, right? We collect more of those events, and so we help AV companies to train their brain and to look into new markets, new cities to expand to. For example, I'll give you an example of a product that we built.

We call it Nexar behavioral Map. And what is it? It is a map that tells you for every road in America, what kind of driver behavior you could expect with what probability, and actually not just driver, also pedestrians, et cetera. So what are the chances that people would change lanes in this road?

And we have the concept of a virtual stop sign. What is a virtual stop sign is when you have a place on the road where many pedestrians pass. Even though it's not regulated, there's no, zebra crossing, there's no, it's not allowed there, maybe even. But that's reality and you want to know about it.

If you're an autonomous vehicle, you really want to know about these kind of peculiarities. And they are very geospecific. 

Grayson Brulte:

Okay. So I'm in a, I've been doing a lot of research, economic modeling, so if you stepped into it and I'm going for it. You mentioned training the AVS with data. I've been doing a lot of economic research around Tesla's Dojo Computer. Yeah. What are your thoughts On Dojo? So Tesla's gathering real world data very similar to your gathering the real world data. Yeah. Tesla has an AI strategy. You have an AI strategy. Yeah. What are your thoughts on vision and then using AI to help train a vehicle and just look at the power of what Dojo has.

Eran Shir:

I am a big believer in camera as the ultimate Uber sensor. We have, for example, we recently shipped our first HD map that is vision only. What does that mean? We you know the concept of of point cloud, right? That you do with LIDAR and stuff. We do an HD point cloud that is SAP 30 centimeter accurate.

Precise with a video stream just running, just capturing a video from a road. And that's a game changer because the unit economics of hd, typically with all the sensing and the lighter and the crunching of it, is in the hundreds of dollars per mile, and we get it down. Orders, magnet, it cheaper because we can use that.

So I think for data gathering, data collection, data modeling, you don't need anything other than a good GPS and a camera, I would say that for like real time safety. I think the jury is still out. If today in the state of AI today, you can handle any situation just with cameras, I think.

That's probably a bit problematic today, I think down the line. I am confident that you'll be able to do that, and the reason why I'm confident is because, I have two cameras when I drive and they serve me pretty well. I think beyond that, something though that's worth mentioning from a philosophical standpoint Th this market of vehicles is super fragmented.

The largest OEM in the world has 8%, 9% market share. And it's super slow, right? Like it's, the turnover is like 10 years, 15 years, whatnot. What does that mean? It means that no one on its own will be able to fully solve this problem. It's going to take a village and so you want to create an ecosystem.

You wanna create an alliance, a consult team, an open standup for a different OEMs to share and collaborate and monetize their data. And I think that's where Tesla is a bit wrong because Tesla is like, we are gonna just. Do it for, just keep it all in the house. We're not gonna collaborate with others.

And I think that strategy is vulnerable to a bunch of the OEMs colluding or collaborating. Under a network like ours to really get to scale to numbers that matter, right? Like I, I mentioned for you, for example, before the use case of available parking spots. This something that we already deliver, we can already deliver in select cities where we have enough density.

And I can tell you, okay, these are all the free parking spots in the last five minutes in Manhattan. And I can give you an SLA on that, but I cannot do it in every city in the US because we don't have enough density. And guess what? Toyota doesn't have enough density for that. Definitely Tesla doesn't have enough Tesla density for that.

So if you want to go to this really value added services, you will have to collaborate. You have to do a marketplace, you have to monetize it and then unlock the those use cases. 

Grayson Brulte:

Collaboration is key to enable the future photo you mentioned. Camera is the ultimate sensor, is that because the economics of cameras are highly scalable and it could fit into the traditional supply chain without having to dramatically raise the price of a vehicle, that's an important element.

Eran Shir:

It's not the only important element. The, I think the other important element is that cameras are running in the band, in the spectrum side, where you can map almost any use case. To repeating vision pattern, right? When I use, when I use radar, right? I don't know whether a sign that I capture a is a stop sign or it's a speed sign speed limit sign, right?

Same thing with lidar. So it's basically blind sensing of depth, but cameras together with AI today. Can do depths, they can do detection. They can unlock cracking. They can unlock a profiling of different agents on the road. They can do anticipation. There are many things that you can do with a simple lidar camera that cost.

Honestly, the camera itself, the sensor cost like five bucks, right? And it's mind boggling, right? So the units you need to build it, you need to do, to put the compute, you need to put, all the memory and storage and all that stuff. But at the end of the day, this is a small sensor that you can embed everywhere.

You can make many of them. You can put eight cameras on a vehicle to do 360, and you can compress the data extremely well. I can either compress it semantically and say, okay, I have this bounding box. This bounding box is a sign. I don't need every I'm cropping everything else. Or I can say, okay, I have here in 15 images.

I compress them through eg. And that's one 10th of the size. So there. We have a lot of know-how of, how to deal with video, how to deal with imagery. We have a whole industry around it. Now we have a lot of AI that was trained on it. It's really the Uber sensor. 

Grayson Brulte:

How important is depth sensing as it relates to the AI in the cameras? Is that a, would you say a critical technology, a critical breakthrough? 

Eran Shir:

Yeah, it is very important. It's important first and foremost for localization. So one thing you should, we should all remember, is that GPS is crap. Especially in urban canyons and places like that, GPS is just all over the map now.

You can fix it. There are, there's RTK, there's a bunch of things that, that you can do. But slam and technologies like that use vision with depth sensing, with ai, with anchoring all this kind of stuff allows you to get to extremely precise localization. That's key. If you want to map the wall, you have to do it properly from a localization perspective.

So that's one element. The adult element is for reconstruction. So I mentioned to you, we help insurance companies and our customers in case of a collision. We do 360, almost like matrix bullet time. You know that. Nain. So we produce that for a collision, right? For if you get, have a Nexar camera and you like that for your insurance, we will produce for you a 360 analysis of the collision, what you've done, what the other vehicles have done, what was their velocity, basically everything other than who's to blame, and that requires depth.

If I wanna measure the velocity, track, the velocity of a pedestrian or a vehicle coming from the side, some stuff like that. Depth is also very important. HD mapping. Super important for that. So depth sensing is definitely a key element we had to develop IP on in order to build breakthroughs in those fields.

Grayson Brulte:

When we look at the insurance markets, you have a real partnership with Cover Whale insurance that's focused on heavy duty class eight trucks. Are you seeing an acceptance from underwriters and carriers across the globe saying the Japanese market, the United States market, The Israeli market, or is it only certain markets where the insurance companies are saying, okay, Eran, let's figure something out.

Or is it just generally accepted across the board that your technology is good for their underwriting practices?

Eran Shir:

I think video telematics and our technology and insurance is a good place for Gibson saying of the future is already here. It's just not evenly distributed because you have different insurance companies with different levels of acceptance.

An eagerness to accept to accept technologies. And many times it has nothing to do with the technology itself. It has everything to do with regulation, local, state, by state regulation, and it has a lot to do with how big you are. So I'll give you a couple of examples. Definitely the higher the premium is, the easier it is to adopt technologies, right?

Because if you have a $20,000 premium like an 18-wheeler has, right? It's no, there's no problem. Taking $500 and spending it on the video telematics if you save one collision of a truck, You basically paid for a thousand different cameras or more. So that's easier from that perspective.

So more commercial, higher premiums, easier to adopt. Second is the question of regulation, right? Because in the consumer space in the US especially the pricing of insurance is reg heavily regulated, especially on the consumer side. For example, in California, you know how they calculate premium in California.

There's a ballot to measure for 1986 that defined the formula of what you can give discounts on car insurance. I kid you not. Okay. So in 1986 there were no dash counts. Yeah so there are very few things that you can actually compensate on or incentivize driver. On and you can't really play with it.

On the other hand, there are other states like Texas where you have more freedom, right? So insurance companies there are more free to adopt technology. It's worth it because they can incentivize the driver, they can give them discounts, they will adopt it more, et cetera, et cetera. So I can go on and on that.

But I'd say in general, It's becoming indispensable, right? Like the way it works is you, your competitors adopt it successfully. They see reduction in loss. They attract the better drivers because when a better driver adopt the camera and you give it a discount, then your overall loss or portfolio became just a bit better. And then everyone else are stuck with the worst drivers. So you get this kinda interesting effect, but it's gonna run through the system, right? There are other countries where that are way ahead of the us. The US is a lagging market. If you go to Korea, a hundred percent of the vehicles have cameras, have dash cams.
In Japan. It's massive. In Taiwan, it's massive. In Russia, we know it's massive, but that's for another reason for corruption. So that's how it goes. And I think the next five years you'll see. All major insurance companies adopting it and building programs for it. 

Grayson Brulte:

Do you feel at some point in the future, not giving a timeline, but some point in the future, either an underwriter or an insurance carrier demands it?

If you want to go through this carrier, you have to have a dash camera. We're not going to take you because insurance companies are very selective, especially as you go in the higher-end markets, the Chubbs and the Pures, and you go higher up in the insurance pecking order, they get a lot pickier on who they're gonna insure or not underinsure. So do we get to a point where they say, no dash cam, no insurance? 

Eran Shir:

It's already happening. So you mentioned cover. That's basically covering stuff. If you wanna, if you wanna do business with us, you have to install a camera. There are other insurance companies as well. We know we have a partnership with the one in Texas that we'll announce.

And there are some international ones. We have one in Israel, for example, that have a program. If you want benefits of that program, you have to install an extra camera. So this is becoming more and more think it's still obviously a niche, but I think this is going to, capture more interest. 

Grayson Brulte:

At the end of the day, what you're describing with the insurance comes down to economics. I'm really curious, what are the economics of Nexar's maps? 

Eran Shir:

Typically to map the road takes 50 to a hundred dollars per mile. If you wanna build a Google Maps competitor from scratch or not from scratch, or that's what it will cost you. Yeah. It will pay for the, those mapping cars and all the sensors and data upload and the drivers and the entire operation.

The way we engineered it, we are talking about two to three orders of magnitude cheaper. Okay, so you can do the math of what that means. It allows us, as a startup, as a company that it's a hundred and some people we're venture backed. We don't have the billions that Google and Apple has.

It allows us to build something that is dramatically better than what they have, right? It's all about the unit economics. Now, as I mentioned in the beginning, the unit economics, Is the barrier. New markets are created when you dramatically change the unit economics, right? Like when things that were not profitable before can now become profitable.

Like I, I mentioned the parking one, right? Like you wouldn't mind. If you live in a city, you wouldn't mind to pay 10 bucks a month in order to get a heads up on where is the nearby vertical parking spot. W we can handle 10 bucks a month for this level of sla, right? Because of our low unit economics.

Google cannot, and that's a key point because all of these use cases, are opening up just because we managed to create this crowd sold network. 

Grayson Brulte:

It seems to me you have multiple revenue stream opportunity. So here I am living in the city and I'll pay you $10 a month.

You multiply that across hundreds of thousand individuals living. In major city and you're licensing the data to, to train autonomous vehicles. Is that the o overall economic strategy to have multiple streams? Of revenue for the company? 

Eran Shir:

Yeah, so we, we had to initially focus because, we're a startup, so we focused on some essential use cases like monitoring construction, monitoring potholes, monitoring this and that, right?

But as we mature, we can unlock more and more use cases. That allowed us to really bring in more and more layers of revenues. But I think that this will converge soon to an sort of a geo oracle because, at the same time as we are building the realtime mapping revolution, There's a much broader revolution that happened in the world of AI in the last year around love language models and ChatGPT and things like that.

Now, today, that revolution is focused on di digital content, right? It's focused on the text and imagery and code, but that, what if you couldn't connect it to the real world? What if you can create ChatGPT for the real world, and then the use cases are not determined by us anymore. They're entertainment by you, the user.
That's unlocking a whole new future. But on that, I think we'll have more to say later in the year. 

Grayson Brulte:

The real-time mapping revolution as you stated, is gonna be really interesting putting that together. In your opinion, what is the future of real-time mapping? 

Eran Shir:

I think the future of real time mapping at the end of the day is first of all, claravoyancy, first of all, knowing what's, what lies ahead.

Like I mentioned, hundred of yards ahead, tens of seconds ahead. That's the ultimate goal because it will be the biggest safety revolution that will go through more than av, more than a as, more than anything. So that's one thing. The other thing is digitizing. Physical space and applying software to it, right?

Like you, you know how they say software is eating the wall. I say software is hitting the roads, right? Like we want to deploy. And I'm not talking just about Nexar, I'm talking about any company, any service that has to do with the physical world. If you wanna deploy software, you need data.

And you need the data to be precise and timely. I think the real time mapping of revolution will allow us to deploy software in more and more use cases to manage the physical wall, manage the Uber scene. I think that is gonna to be exciting when your smart city will really be smart that will be I think an exciting result. 

Grayson Brulte:

Indeed, it will be exciting. And Eran this has been a really exciting, interesting conversation. I thank you so much for taking the time for us today to explain real time mapping and where you're going with Nexar. And as we like to wrap up this insightful conversation, what would you like our listeners take away with them today?

Eran Shir:

I would love for them to first consider a camera because probably most of them do not have a camera. And I think they would benefit, like I benefited last week. And then the other thing, I would invite them to come up with use cases. Cause we, when I talk about things like ChatGPT for the real world and stuff.

A lot of it is what can you imagine? And I would invite people to try our real-time mapping platform at mapping.getnexar.com, think about their use cases and reach out because we have all of these APIs available for developers and whatnot. These are my 2 cents. 

Grayson Brulte:

I like that. I'll summarize it this way. Build. There's an opportunity to build something meaningful. There's an opportunity to save lives. There's an opportunity to build convenience. So go out there and build. Check out the Nexar API. Today is tomorrow. Tomorrow is today. The future is Nexar. Eran, thank you so much for coming on SAE Tomorrow Today.

Eran Shir:

Oh, thank you. It was a pleasure. 

Grayson Brulte:

Thank you for listening to SAE Tomorrow Today. If you've enjoyed this episode and would like to hear more, please kindly rate review and let us know what topics you'd like for us to explore next.

SAE International makes no representations as to the accuracy of the information presented in this podcast. The information and opinions are for general information only. SAE International does not endorse, approve, recommend, or certify any information, product, process, service, or organization presented or mentioned in this podcast.


Listen to the full back catalog on your favorite podcast platform.