June 16, 2026 • 25 mins
This episode of Transportopia explores Imagry's camera-and-AI autonomous bus system and its focus on solving the global shortage of bus drivers and improving public transit. The Israeli-based tech startup with offices in San Jose, CA will soon launch its first US autonomous bus deployments in Florida and Nevada.
Host Rick L'Amie interviews Imagry CEO Eran Ofir about mapless, on-board autonomy, safety testing, real-world deployments, and how autonomous buses could reduce congestion and reshape cities.
Show Notes
IMAGRY https://imagry.co/
VIDEOS L3 Autonomous Bus on Public Roads (Israel) https://youtu.be/jnaTUn7PvFw
Mixed Car/Bus Demo Drives https://youtu.be/TzxnC8D7h5M
San Jose Urban Driving https://youtu.be/hu_xPdBHhX8
NCAP Safety Testing https://youtu.be/XRuFQDkYa5Q
ARTICLES PAVE Campaign — 5 Questions with Eran Ofir https://pavecampaign.org/5-questions-with-eran-ofir-ceo-of-imagry/
The Curbivore — To Map or Not to Map https://www.thecurbivore.com/p/to-map-or-not-to-map
New Polis — The Journey of Autonomous Buses https://newpolis.media/the-journey-of-autonomous-buses-interview-with-eran-ofir-ceo-of-imagry/
S&P Global — Congress Weighs AV Safety Law https://www.spglobal.com/market-intelligence/en/news-insights/articles/2026/2/as-congress-weighs-autonomous-vehicle-safety-law-industry-split-on-proof-points-97424966
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
Those vehicles do not solve any of the problems.
(00:09):
They don't solve the problem of shortage of drivers in public transportation.
They do not solve the problem of congestion in the cities.
They add, if any, they only add more, hundreds of more vehicles into city centers.
The hype and the news and the papers are all about provo taxis.
(00:29):
However, the need and the pain is with buses, with mass public transportation.
This is Transpertopia, a podcast exploring the past, present and future of public transportation,
(00:50):
the built environment and more.
Hello everyone and welcome to Transpertopia.
I'm Rick Lemme.
For years, the autonomous vehicle conversation has been dominated by robo taxis, private
cars for one or two people circling city streets, adding to congestion instead of solving it
(01:14):
in most cases.
But there's a quieter revolution happening in public transit, and it's further along
than many people realize.
My guest today is Aaron Ophir, the CEO of Imagery, an autonomous driving software company
founded in Haifa, Israel, with active deployments in Germany, Israel and Japan, and a U.S. expansion
underway.
(01:35):
Here's what makes Imagery different from other players in this space.
Their buses drive themselves using nothing but cameras and AI, no laser sensors, no detailed
pre-built maps of every road, no connection to the cloud.
The system sees the road the same way you and I do.
It builds an understanding of what's around it in real time and makes decisions three
(01:55):
times faster than a human driver, according to the company.
And unlike everyone else chasing the robo taxi dream, Imagery has chosen to focus entirely
on public transportation.
Buses on real roads serving real riders.
They've been testing and deploying since 2019, and they're the first autonomous bus company
to pass the kind of rigorous crash safety testing that conventional vehicles go through.
(02:18):
And Aaron is here today to make the case this technology could be the most powerful tool
we have for solving the global bus driver shortage and maybe redesigning a city itself.
Aaron, welcome to Transportopia.
Thank you.
Thank you, Rich.
Good morning.
A tradition on our podcast is to chat as if we're sitting at a conference or a coffee
(02:39):
shop having a conversation.
So what are you drinking for our conversation today?
You're in a little different time zone than me by nearly half a day.
But it's always coffee, typically espresso.
That's my thing.
I have three espresso machines at home, two that are manual and one is automatic.
And you know, for like more celebrated events, then it's typically red wine.
(03:06):
I'm drinking what's called high desert sage.
This is a local mixture of herbs from where I am in Santa Fe, New Mexico, here in the
high desert.
Nice.
Another question I like to ask my guest is, what's your first memory of public transportation?
Do you have a time when you remember riding a bus or taking a train in your past that
(03:28):
was remarkable for you?
I wasn't exactly a big user of public transportation.
I grew up in the city and my father, he worked as a driver.
So you know, he always used to drive us in his car.
And until my first car, I had a bike, a racing bike that I used to be on for three hours
(03:56):
of a day.
I would go everywhere with them, even long distance.
And then when I became a student, I, you know, I got my first micro car.
Yeah.
Well, that's a story of so many people everywhere, but it sounds like you did have some alternates
to the car, riding a bike.
So that's good too.
Right.
(04:17):
So for listeners who haven't heard of your company, give me a 30 second version.
What is Imagery actually doing?
No maps, no cameras.
Tell us more about it.
Correct.
So Imagery is a provider, is a developer of full autonomous driving software stack.
We are a software company, basically developing AI based software that can drive any vehicle
(04:42):
autonomously.
So it has everything that is needed, perception, motion planning, control.
And we drive any types of vehicles we drive.
We integrate it with passenger vehicles, with vans and with buses.
So you particularly have focused on public transit.
Based on your background, with all the things you've done in other industries, how did you
(05:07):
decide to focus on public transit as a customer for your software?
So indeed, we went a different way than the hype, right?
Because the hype and the news and the papers are all about robo taxis.
However, the need and the pain and the growth currently is with buses, with mass public
(05:29):
transportation, because there is a global shortage of a dire shortage of 20% in bus
drivers globally, everywhere, whether it's school buses or public transportation operator
buses, and it's growing by 1% per year.
(05:50):
There are places that it's even much worse, like in Japan, where 40% of the bus drivers
are going to retire within the next 10 years.
So it's a national crisis.
And you see lines being canceled everywhere.
So there is a global pain.
And in cities, in big metropolitan areas, there is a lot of demand, especially with
(06:12):
urbanization and with aging.
That's another thing that is currently strong in many places around the world.
When the public transportation operators collapse, they cannot provide the service.
So there is a need.
And we are here to cater for that need.
So that's a perfect timing.
(06:32):
And you know, when you build startups or companies in general, I always say that it's the triple
T that you need to have in order to be successful.
It's team time, timing and technology.
So we have all of them.
We have the right technology and a great team at the right timing.
(06:53):
Team timing and technology, I like that.
That makes sense.
And the global congestion and driver shortage problem isn't going to go away anytime soon.
Your company has been saying that the AV industry has been working for a decade to solve the
wrong problem.
What do you mean by that?
We see all the advancement of autonomous driving.
(07:15):
By the way, if you remember, it only started really like two years ago.
And the reason that everyone were talking about it for at least a decade.
But it didn't happen.
And the reason that it didn't happen is that they did not solve, the silicon companies
could not provide sufficient computing power, edge computing in the vehicles, in order to
(07:41):
run the AI networks that are needed, really heavy ones, in order to drive autonomously
well and safe enough.
So and then you suddenly saw kind of the industry booming with thousands of robo taxis everywhere
and trucking and buses and so on.
(08:01):
But they went the wrong direction because the first thing that happened is that big
tech companies, you know, the Amazon, the Google, the Alphabet, the Uber, they came
in and almost immediately started expanding with vehicles, with passenger vehicles.
But those vehicles do not solve any of the problems.
(08:24):
They don't solve the problem of shortage of drivers in public transportation.
They do not solve the problem of congestion in the cities.
They add, if any, they only add more hundreds of more vehicles into city centers.
They don't make cities become greener.
(08:47):
And every such robo taxi, most of them, by the way, 65, 70% of the time drive empty.
And when they don't drive empty, they carry only one to two passengers.
A bus, buses carry 25 to 40 passengers.
So even though so much funding, including public funding, goes down the path of more
(09:13):
vehicles and more cars and robo taxis on the road, the real solution, like a win-win-win
situation between the governments, the public, and the public transportation operators are
with buses.
Back last December, there was a power station that knocked out power across San Francisco.
Nearly 1,600 Waymo vehicles were stalled in the streets, fire trucks, buses were stuck
(09:37):
behind them.
The problem was that Waymo just couldn't respond fast enough on the ground.
Your system runs entirely on board.
So would your platform have prevented something like that?
So indeed, our system, every vehicle that imagery operates, and we run autonomously
on public roads since 2019, and the US, California, Arizona, Nevada, Germany, Japan, and Israel,
(10:06):
every such vehicle is completely self-sufficient.
We don't need anything coming from the outside.
First of all, we are mapless.
We don't need pre-prepared HD map to be fed into the vehicles, which means that we can
drive everywhere, unlike robo-taxis that can only drive in very specific geo-fenced
areas where they have the maps.
(10:28):
That's one thing.
We are not dependent on high bandwidth communication, because it might be that you are in a tunnel
or in an area where there is no communication currently because of some network problem,
and we won't be affected out of it.
So with that, such a network outage won't affect imagery vehicles, and there is one
(10:54):
more important thing, even beyond the network outage, which is cyber vulnerability.
Because of our vehicles being closed, being self-sufficient, and drive themselves, like
we drive, the same way that humans drive, the same philosophy in terms of development,
(11:17):
we could not be breached.
And by the way, we passed cyber security framework tests in order to prove that, because it's
mandatory, for example, in Europe, and now there is even an investigation in the US Senate
about robo-taxis being remotely driven by people in the Philippines.
(11:40):
And basically, such a thing could not happen with us.
So that's another factor to take into account.
I think those are two key factors.
Explain a little bit about how it is that your self-driving vehicles are essentially
seeing the road like our eyes, the difference between other systems.
(12:01):
There's cameras, there's no mapping.
Tell us a little bit about how it works.
So the entire system was built from the get-go, since 2018.
And it was all built AI-based, even before AI was AI.
And the way that it, because it was developed to walk as a human being, in terms of how
(12:24):
we see the world, in terms of perception with eyes.
We have eight cameras for that, seeing and analyzing everything, 360 degrees, 300 meters.
Every object, every vector of everything that is moving, and so on.
We as humans cannot see what's happening behind our back in real time.
And then we have a distributed neural network architecture, which means that many, many
(12:49):
neural networks are getting the video feed from the cameras, where each network was trained
for many years, hundreds of millions of images, basically, to do object detection and classification
for a different type of object.
So one is looking on traffic lights, and one on traffic signs, and one on lanes, and
pedestrians, and trucks, and bicycles, and so on.
(13:11):
And we take the output of all those neural networks, and we place them one on top of
the other, really quickly, in some 100 milliseconds.
And that map that we created in real time, which is the same thing, the same process
is happening in our brain, gets into another section of our software that we call the cortex.
It's the motion planning part, where, again, with AI, the system was trained for many years
(13:39):
to take the same decisions, the same way that we learned how to behave on the road, and
drive and make decisions.
That's how the system was trained, it's called supervised learning.
There are millions of decisions that AI is making, from what I understand, based on the
training that you gave it.
So it probably even learns as it goes, right?
(14:01):
In different scenarios.
I imagine you've tried to imagine every possible scenario, but it still can possibly learn
even a new one, right?
Exactly.
We have great videos in which we show that the system detected and handled scenarios
that we never trained it.
(14:22):
For example, unknown objects.
Sometimes we drive on a highway, and there is a piano.
We have a picture of that, a piano that fell off a moving truck.
And we need to handle it, even though we never trained any of our neural networks to detect
a piano.
And in Frankfurt, you know, we were driving on snow, and there was a child on a sled.
(14:47):
And again, the system detected a child on a sled, on snow.
And again, we never trained a neural network to detect such things.
So you can see, and there are people that are excited of it, and there are people that
are afraid of it.
But you can see how AI-based system is like a continuous learning type of entity.
(15:15):
And it keeps learning and training itself.
We vet the decisions, and we have annotation tools or basically another software that is
doing vetting in order to tell the system whether in the training process, whether decisions
made are the right ones, are correct, or they're not.
(15:38):
And that's how you keep learning.
You know, it's like me sitting in the car near my high school daughter when, you know,
she learned how to drive.
And I was telling her, well, here, you should have done that, and so on.
So the same process is with the AI.
Let's talk a little bit about the driver shortage.
You cited the statistics earlier.
(15:59):
It's a little complicated because there are going to be drivers, maybe unions, who say,
you're taking jobs away from me.
What do you say to drivers and unions who see this as a job threat?
In the U.S., yes, we hear it.
But the use cases that we face are different.
Even though it's the same product, the same solution, the use cases that we drive in the
(16:25):
U.S., first of all, is airports, which is really the classic case where, you know, autonomous
buses driving within an airport between the terminals and long-term parking and so on.
Then there are like rural areas that they enjoy, they can enjoy better public transportation
(16:46):
that they currently, in many cases, they don't have.
And, you know, the U.S., public transportation is not number one priority in many states.
And now with the midterm elections, there is a lot of federal funding for serving the
underserved, you know, providing access to public transportation for all kinds of communities
(17:10):
that they need it.
And we see sporting events like the Olympics and so on, where there would be autonomous
buses.
And last, maybe I should mention a surprising use case for us in the U.S., many private
operators that are providing corporate shuttles.
(17:30):
So basically, they are bringing employees to factories.
It could be Google, it could be Apple, it could be Boeing.
And it's fixed route shuttles that are taking employees every morning and then every afternoon
to factories and back home.
(17:50):
And there are so many of such companies, and they would like to convert some of these lines
to be autonomous as well.
So I don't think that we really threat anyone in the U.S., you know, with respect of driving.
And it's a known fact that there is a shortage of drivers.
(18:11):
You can ask every school in the U.S. as for their ability to obtain enough drivers for
the, you know, for the school buses.
So we'll wrap up here with a few more questions.
You've been doing your research and development in California, but you don't have a live U.S.
(18:31):
transit deployment yet.
Do you expect that coming soon?
And where would it be?
Yes, our first live U.S. deployment until the end of this year would be in Florida.
And the next day, the one to follow would be in Nevada.
And what would those applications entail?
What kind of passengers would it pick up?
(18:53):
It would be an M2 category vehicle, meaning a van, driving 15, 18, 20 passengers in a
city in an open environment on a public road, like, you know, providing a service, a shuttle
service in city center in both cases.
(19:16):
And in those cases, would those be private providers or would you work with public transportation?
It's private companies, operators that are providing the service, but they were contracted
by the municipalities in order to provide such services in the city.
(19:38):
Where do you see this going in, I guess, the United States and across the world?
What does a city look like in 20 years if autonomous transit vehicles scale the way
you envision it?
So when we talk with regulators, with governments, basically, and with public transportation
operators, we see that they have targets of how many autonomous buses they need to
(20:06):
bring on the road.
And we enjoy strong tailwinds because, first of all, the governments are pushing it with
funds.
That's one.
Second, in almost in every new bid that is getting out in Europe for public transportation,
(20:28):
they already set now a percentage of buses that needs to operate autonomously.
So if you are a PTO, you need to start equipping yourself with autonomous driving capabilities.
So that's another thing.
And another strong tailwind is that all over the world, buses are being converted to electric
(20:50):
buses.
And the advantage that it entails is that those buses typically come AD ready because
their steering system was already built in a way that it can accept commands from a computer.
So the more electric buses exist in fleets of public transportation operator, the easier
(21:13):
and faster and cheaper it is for them to get on that train and start deploying.
And it's very profitable for public transportation operators.
Every such bus that is getting on the road and becoming autonomous, they call it operational
(21:34):
benefits.
And the benefit to public transportation operators is between $120,000 to $150,000 per year per
bus.
So when you do the multiply, it's a lot of money.
Right.
Right.
It's it's going to be sweeping across the industry, not just with, say, government agencies,
(21:55):
but with manufacturers and possibly even your software being used on what I'm understanding
you to say, if there's an electric vehicle, it could be retrofitted or updated to add
your software in some way.
Yes, basically, we do need to assemble our computing, our cameras, cabling and so on.
(22:15):
But we work closely with the OEMs.
We are already integrated with five buses and now with the sixth.
So we have a Spanish bus, an Australian bus, a Turkish bus, Japanese, American, Chinese.
So we work with buses, so we'll have buses that are homologated, meaning that can go
on public roads in all parts of the world.
(22:36):
So we work closely with the bus OEMs so that the integrated buses will come straight from
the bus manufacturers to the customer.
And we just provide the software, basically the license software for that.
Well, this has been an exciting conversation to think about choices besides watching a
(22:58):
Waymo drive around downtown, making people frustrated.
Thanks for sharing with us today and joining us on Transportopia.
Thank you.
Thank you.
Thank you for having me here.
Thank you, Rick.
Take care.
Here's what stays with me from this conversation.
(23:19):
Waymo is now permitted or testing in 26 of the top 30 U.S. metro areas, with plans to
launch in a dozen more cities in 2026.
The robo-taxi movement is real and it's accelerating, but it's also generating a growing list of
complaints, blocked intersections, obstruction of emergency vehicles, incidents in floodwaters,
neighborhoods being circled by driverless cars at 2 in the morning.
(23:43):
Cities are pushing back, regulators are scrambling to keep up.
And through it all, the fundamental question not enough people are asking loudly enough
is what problem are we actually solving?
Research is already showing that with current policies, autonomous vehicles are likely to
increase total vehicle travel, not reduce it.
More cars, more miles, more congestion.
(24:03):
That's the trajectory if we keep building autonomous vehicles for private trips.
Aaron Ophir is asking a different question.
What if the technology served the public bus?
What if it solved the driver shortage instead of replacing the driver?
What if autonomous vehicles actually reduce the number of cars on the road?
Imagery's autonomous buses roll out in Florida and Nevada before the end of this year.
(24:25):
That's not a concept.
That's a pilot.
That's a bus on a public road in an American city with no one behind the wheel, moving
people who need to get somewhere.
That's a story I'll be watching.
Thanks for listening to Transpertopia, the intersection of transportation and utopia.
(24:49):
Like us wherever you get your podcasts, leave us a review, and visit transpertopia.org for
show notes, our blog, and more.
I'm Rick Lemieux.
Goodbye, everybody.
Those vehicles do not solve any of the problems.
(00:09):
They don't solve the problem of shortage of drivers in public transportation.
They do not solve the problem of congestion in the cities.
They add, if any, they only add more, hundreds of more vehicles into city centers.
The hype and the news and the papers are all about provo taxis.
(00:29):
However, the need and the pain is with buses, with mass public transportation.
This is Transpertopia, a podcast exploring the past, present and future of public transportation,
(00:50):
the built environment and more.
Hello everyone and welcome to Transpertopia.
I'm Rick Lemme.
For years, the autonomous vehicle conversation has been dominated by robo taxis, private
cars for one or two people circling city streets, adding to congestion instead of solving it
(01:14):
in most cases.
But there's a quieter revolution happening in public transit, and it's further along
than many people realize.
My guest today is Aaron Ophir, the CEO of Imagery, an autonomous driving software company
founded in Haifa, Israel, with active deployments in Germany, Israel and Japan, and a U.S. expansion
underway.
(01:35):
Here's what makes Imagery different from other players in this space.
Their buses drive themselves using nothing but cameras and AI, no laser sensors, no detailed
pre-built maps of every road, no connection to the cloud.
The system sees the road the same way you and I do.
It builds an understanding of what's around it in real time and makes decisions three
(01:55):
times faster than a human driver, according to the company.
And unlike everyone else chasing the robo taxi dream, Imagery has chosen to focus entirely
on public transportation.
Buses on real roads serving real riders.
They've been testing and deploying since 2019, and they're the first autonomous bus company
to pass the kind of rigorous crash safety testing that conventional vehicles go through.
(02:18):
And Aaron is here today to make the case this technology could be the most powerful tool
we have for solving the global bus driver shortage and maybe redesigning a city itself.
Aaron, welcome to Transportopia.
Thank you.
Thank you, Rich.
Good morning.
A tradition on our podcast is to chat as if we're sitting at a conference or a coffee
(02:39):
shop having a conversation.
So what are you drinking for our conversation today?
You're in a little different time zone than me by nearly half a day.
But it's always coffee, typically espresso.
That's my thing.
I have three espresso machines at home, two that are manual and one is automatic.
And you know, for like more celebrated events, then it's typically red wine.
(03:06):
I'm drinking what's called high desert sage.
This is a local mixture of herbs from where I am in Santa Fe, New Mexico, here in the
high desert.
Nice.
Another question I like to ask my guest is, what's your first memory of public transportation?
Do you have a time when you remember riding a bus or taking a train in your past that
(03:28):
was remarkable for you?
I wasn't exactly a big user of public transportation.
I grew up in the city and my father, he worked as a driver.
So you know, he always used to drive us in his car.
And until my first car, I had a bike, a racing bike that I used to be on for three hours
(03:56):
of a day.
I would go everywhere with them, even long distance.
And then when I became a student, I, you know, I got my first micro car.
Yeah.
Well, that's a story of so many people everywhere, but it sounds like you did have some alternates
to the car, riding a bike.
So that's good too.
Right.
(04:17):
So for listeners who haven't heard of your company, give me a 30 second version.
What is Imagery actually doing?
No maps, no cameras.
Tell us more about it.
Correct.
So Imagery is a provider, is a developer of full autonomous driving software stack.
We are a software company, basically developing AI based software that can drive any vehicle
(04:42):
autonomously.
So it has everything that is needed, perception, motion planning, control.
And we drive any types of vehicles we drive.
We integrate it with passenger vehicles, with vans and with buses.
So you particularly have focused on public transit.
Based on your background, with all the things you've done in other industries, how did you
(05:07):
decide to focus on public transit as a customer for your software?
So indeed, we went a different way than the hype, right?
Because the hype and the news and the papers are all about robo taxis.
However, the need and the pain and the growth currently is with buses, with mass public
(05:29):
transportation, because there is a global shortage of a dire shortage of 20% in bus
drivers globally, everywhere, whether it's school buses or public transportation operator
buses, and it's growing by 1% per year.
(05:50):
There are places that it's even much worse, like in Japan, where 40% of the bus drivers
are going to retire within the next 10 years.
So it's a national crisis.
And you see lines being canceled everywhere.
So there is a global pain.
And in cities, in big metropolitan areas, there is a lot of demand, especially with
(06:12):
urbanization and with aging.
That's another thing that is currently strong in many places around the world.
When the public transportation operators collapse, they cannot provide the service.
So there is a need.
And we are here to cater for that need.
So that's a perfect timing.
(06:32):
And you know, when you build startups or companies in general, I always say that it's the triple
T that you need to have in order to be successful.
It's team time, timing and technology.
So we have all of them.
We have the right technology and a great team at the right timing.
(06:53):
Team timing and technology, I like that.
That makes sense.
And the global congestion and driver shortage problem isn't going to go away anytime soon.
Your company has been saying that the AV industry has been working for a decade to solve the
wrong problem.
What do you mean by that?
We see all the advancement of autonomous driving.
(07:15):
By the way, if you remember, it only started really like two years ago.
And the reason that everyone were talking about it for at least a decade.
But it didn't happen.
And the reason that it didn't happen is that they did not solve, the silicon companies
could not provide sufficient computing power, edge computing in the vehicles, in order to
(07:41):
run the AI networks that are needed, really heavy ones, in order to drive autonomously
well and safe enough.
So and then you suddenly saw kind of the industry booming with thousands of robo taxis everywhere
and trucking and buses and so on.
(08:01):
But they went the wrong direction because the first thing that happened is that big
tech companies, you know, the Amazon, the Google, the Alphabet, the Uber, they came
in and almost immediately started expanding with vehicles, with passenger vehicles.
But those vehicles do not solve any of the problems.
(08:24):
They don't solve the problem of shortage of drivers in public transportation.
They do not solve the problem of congestion in the cities.
They add, if any, they only add more hundreds of more vehicles into city centers.
They don't make cities become greener.
(08:47):
And every such robo taxi, most of them, by the way, 65, 70% of the time drive empty.
And when they don't drive empty, they carry only one to two passengers.
A bus, buses carry 25 to 40 passengers.
So even though so much funding, including public funding, goes down the path of more
(09:13):
vehicles and more cars and robo taxis on the road, the real solution, like a win-win-win
situation between the governments, the public, and the public transportation operators are
with buses.
Back last December, there was a power station that knocked out power across San Francisco.
Nearly 1,600 Waymo vehicles were stalled in the streets, fire trucks, buses were stuck
(09:37):
behind them.
The problem was that Waymo just couldn't respond fast enough on the ground.
Your system runs entirely on board.
So would your platform have prevented something like that?
So indeed, our system, every vehicle that imagery operates, and we run autonomously
on public roads since 2019, and the US, California, Arizona, Nevada, Germany, Japan, and Israel,
(10:06):
every such vehicle is completely self-sufficient.
We don't need anything coming from the outside.
First of all, we are mapless.
We don't need pre-prepared HD map to be fed into the vehicles, which means that we can
drive everywhere, unlike robo-taxis that can only drive in very specific geo-fenced
areas where they have the maps.
(10:28):
That's one thing.
We are not dependent on high bandwidth communication, because it might be that you are in a tunnel
or in an area where there is no communication currently because of some network problem,
and we won't be affected out of it.
So with that, such a network outage won't affect imagery vehicles, and there is one
(10:54):
more important thing, even beyond the network outage, which is cyber vulnerability.
Because of our vehicles being closed, being self-sufficient, and drive themselves, like
we drive, the same way that humans drive, the same philosophy in terms of development,
(11:17):
we could not be breached.
And by the way, we passed cyber security framework tests in order to prove that, because it's
mandatory, for example, in Europe, and now there is even an investigation in the US Senate
about robo-taxis being remotely driven by people in the Philippines.
(11:40):
And basically, such a thing could not happen with us.
So that's another factor to take into account.
I think those are two key factors.
Explain a little bit about how it is that your self-driving vehicles are essentially
seeing the road like our eyes, the difference between other systems.
(12:01):
There's cameras, there's no mapping.
Tell us a little bit about how it works.
So the entire system was built from the get-go, since 2018.
And it was all built AI-based, even before AI was AI.
And the way that it, because it was developed to walk as a human being, in terms of how
(12:24):
we see the world, in terms of perception with eyes.
We have eight cameras for that, seeing and analyzing everything, 360 degrees, 300 meters.
Every object, every vector of everything that is moving, and so on.
We as humans cannot see what's happening behind our back in real time.
And then we have a distributed neural network architecture, which means that many, many
(12:49):
neural networks are getting the video feed from the cameras, where each network was trained
for many years, hundreds of millions of images, basically, to do object detection and classification
for a different type of object.
So one is looking on traffic lights, and one on traffic signs, and one on lanes, and
pedestrians, and trucks, and bicycles, and so on.
(13:11):
And we take the output of all those neural networks, and we place them one on top of
the other, really quickly, in some 100 milliseconds.
And that map that we created in real time, which is the same thing, the same process
is happening in our brain, gets into another section of our software that we call the cortex.
It's the motion planning part, where, again, with AI, the system was trained for many years
(13:39):
to take the same decisions, the same way that we learned how to behave on the road, and
drive and make decisions.
That's how the system was trained, it's called supervised learning.
There are millions of decisions that AI is making, from what I understand, based on the
training that you gave it.
So it probably even learns as it goes, right?
(14:01):
In different scenarios.
I imagine you've tried to imagine every possible scenario, but it still can possibly learn
even a new one, right?
Exactly.
We have great videos in which we show that the system detected and handled scenarios
that we never trained it.
(14:22):
For example, unknown objects.
Sometimes we drive on a highway, and there is a piano.
We have a picture of that, a piano that fell off a moving truck.
And we need to handle it, even though we never trained any of our neural networks to detect
a piano.
And in Frankfurt, you know, we were driving on snow, and there was a child on a sled.
(14:47):
And again, the system detected a child on a sled, on snow.
And again, we never trained a neural network to detect such things.
So you can see, and there are people that are excited of it, and there are people that
are afraid of it.
But you can see how AI-based system is like a continuous learning type of entity.
(15:15):
And it keeps learning and training itself.
We vet the decisions, and we have annotation tools or basically another software that is
doing vetting in order to tell the system whether in the training process, whether decisions
made are the right ones, are correct, or they're not.
(15:38):
And that's how you keep learning.
You know, it's like me sitting in the car near my high school daughter when, you know,
she learned how to drive.
And I was telling her, well, here, you should have done that, and so on.
So the same process is with the AI.
Let's talk a little bit about the driver shortage.
You cited the statistics earlier.
(15:59):
It's a little complicated because there are going to be drivers, maybe unions, who say,
you're taking jobs away from me.
What do you say to drivers and unions who see this as a job threat?
In the U.S., yes, we hear it.
But the use cases that we face are different.
Even though it's the same product, the same solution, the use cases that we drive in the
(16:25):
U.S., first of all, is airports, which is really the classic case where, you know, autonomous
buses driving within an airport between the terminals and long-term parking and so on.
Then there are like rural areas that they enjoy, they can enjoy better public transportation
(16:46):
that they currently, in many cases, they don't have.
And, you know, the U.S., public transportation is not number one priority in many states.
And now with the midterm elections, there is a lot of federal funding for serving the
underserved, you know, providing access to public transportation for all kinds of communities
(17:10):
that they need it.
And we see sporting events like the Olympics and so on, where there would be autonomous
buses.
And last, maybe I should mention a surprising use case for us in the U.S., many private
operators that are providing corporate shuttles.
(17:30):
So basically, they are bringing employees to factories.
It could be Google, it could be Apple, it could be Boeing.
And it's fixed route shuttles that are taking employees every morning and then every afternoon
to factories and back home.
(17:50):
And there are so many of such companies, and they would like to convert some of these lines
to be autonomous as well.
So I don't think that we really threat anyone in the U.S., you know, with respect of driving.
And it's a known fact that there is a shortage of drivers.
(18:11):
You can ask every school in the U.S. as for their ability to obtain enough drivers for
the, you know, for the school buses.
So we'll wrap up here with a few more questions.
You've been doing your research and development in California, but you don't have a live U.S.
(18:31):
transit deployment yet.
Do you expect that coming soon?
And where would it be?
Yes, our first live U.S. deployment until the end of this year would be in Florida.
And the next day, the one to follow would be in Nevada.
And what would those applications entail?
What kind of passengers would it pick up?
(18:53):
It would be an M2 category vehicle, meaning a van, driving 15, 18, 20 passengers in a
city in an open environment on a public road, like, you know, providing a service, a shuttle
service in city center in both cases.
(19:16):
And in those cases, would those be private providers or would you work with public transportation?
It's private companies, operators that are providing the service, but they were contracted
by the municipalities in order to provide such services in the city.
(19:38):
Where do you see this going in, I guess, the United States and across the world?
What does a city look like in 20 years if autonomous transit vehicles scale the way
you envision it?
So when we talk with regulators, with governments, basically, and with public transportation
operators, we see that they have targets of how many autonomous buses they need to
(20:06):
bring on the road.
And we enjoy strong tailwinds because, first of all, the governments are pushing it with
funds.
That's one.
Second, in almost in every new bid that is getting out in Europe for public transportation,
(20:28):
they already set now a percentage of buses that needs to operate autonomously.
So if you are a PTO, you need to start equipping yourself with autonomous driving capabilities.
So that's another thing.
And another strong tailwind is that all over the world, buses are being converted to electric
(20:50):
buses.
And the advantage that it entails is that those buses typically come AD ready because
their steering system was already built in a way that it can accept commands from a computer.
So the more electric buses exist in fleets of public transportation operator, the easier
(21:13):
and faster and cheaper it is for them to get on that train and start deploying.
And it's very profitable for public transportation operators.
Every such bus that is getting on the road and becoming autonomous, they call it operational
(21:34):
benefits.
And the benefit to public transportation operators is between $120,000 to $150,000 per year per
bus.
So when you do the multiply, it's a lot of money.
Right.
Right.
It's it's going to be sweeping across the industry, not just with, say, government agencies,
(21:55):
but with manufacturers and possibly even your software being used on what I'm understanding
you to say, if there's an electric vehicle, it could be retrofitted or updated to add
your software in some way.
Yes, basically, we do need to assemble our computing, our cameras, cabling and so on.
(22:15):
But we work closely with the OEMs.
We are already integrated with five buses and now with the sixth.
So we have a Spanish bus, an Australian bus, a Turkish bus, Japanese, American, Chinese.
So we work with buses, so we'll have buses that are homologated, meaning that can go
on public roads in all parts of the world.
(22:36):
So we work closely with the bus OEMs so that the integrated buses will come straight from
the bus manufacturers to the customer.
And we just provide the software, basically the license software for that.
Well, this has been an exciting conversation to think about choices besides watching a
(22:58):
Waymo drive around downtown, making people frustrated.
Thanks for sharing with us today and joining us on Transportopia.
Thank you.
Thank you.
Thank you for having me here.
Thank you, Rick.
Take care.
Here's what stays with me from this conversation.
(23:19):
Waymo is now permitted or testing in 26 of the top 30 U.S. metro areas, with plans to
launch in a dozen more cities in 2026.
The robo-taxi movement is real and it's accelerating, but it's also generating a growing list of
complaints, blocked intersections, obstruction of emergency vehicles, incidents in floodwaters,
neighborhoods being circled by driverless cars at 2 in the morning.
(23:43):
Cities are pushing back, regulators are scrambling to keep up.
And through it all, the fundamental question not enough people are asking loudly enough
is what problem are we actually solving?
Research is already showing that with current policies, autonomous vehicles are likely to
increase total vehicle travel, not reduce it.
More cars, more miles, more congestion.
(24:03):
That's the trajectory if we keep building autonomous vehicles for private trips.
Aaron Ophir is asking a different question.
What if the technology served the public bus?
What if it solved the driver shortage instead of replacing the driver?
What if autonomous vehicles actually reduce the number of cars on the road?
Imagery's autonomous buses roll out in Florida and Nevada before the end of this year.
(24:25):
That's not a concept.
That's a pilot.
That's a bus on a public road in an American city with no one behind the wheel, moving
people who need to get somewhere.
That's a story I'll be watching.
Thanks for listening to Transpertopia, the intersection of transportation and utopia.
(24:49):
Like us wherever you get your podcasts, leave us a review, and visit transpertopia.org for
show notes, our blog, and more.
I'm Rick Lemieux.
Goodbye, everybody.
Popular Podcasts
Betrayal Weekly
Betrayal Weekly is back for a new season. Every Thursday, Betrayal Weekly shares first-hand accounts of broken trust, shocking deceptions, and the trail of destruction they leave behind. Hosted by Andrea Gunning, this weekly ongoing series digs into real-life stories of betrayal and the aftermath. From stories of double lives to dark discoveries, these are cautionary tales and accounts of resilience against all odds. From the producers of the critically acclaimed Betrayal series, Betrayal Weekly drops new episodes every Thursday. If you would like to share your story, you can reach out to the Betrayal Team by emailing them at betrayalpod@gmail.com and follow us on Instagram at @betrayalpod and @glasspodcasts. Please join our Substack for additional exclusive content, curated book recommendations, and community discussions. Sign up FREE by clicking this link Beyond Betrayal Substack. Join our community dedicated to truth, resilience, and healing. Your voice matters! Be a part of our Betrayal journey on Substack.
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com