Subscribe
Share
Share
Embed
The World is Moving Fast. Futureproof Yourself Weekly.
Your front-row seat to the Future of Business, Tech, Life, and Human Potential.
On Futurecaster with Kimberly Bates, learn through thought-provoking conversations with world-leading founders, AI experts, scientists, engineers, doctors, business leaders, technologists, investors, and everyday builders. Each episode unpacks the real opportunities shaping what’s to come and the blind spots along the way.
Kimberly Bates is a world-renowned futurist and executive brand leader. For over 20 years, she has advised some of the world’s most valuable global companies and iconic brands, from Fortune 100s to fast-moving startups, helping them anticipate change and shape the direction of their industries, customers, and culture. Her work has driven future-ready brand and business transformations, breakthrough products and services, and entirely new business models and revenue streams.
The Futurecaster® Podcast is on a mission to make future thinking available to everyone. If people can see these opportunities for themselves, they can help create them.
©2026 Futurecaster, LLC | Futurecaster© is a registered trademark
I just drove cross country a few months ago in my Tesla to set the full self driving
Cannonball record.
The first time we were successful, was 99.4%.
I forget the percentage.
was well into the nineties percentage FSD engaged, no hands on wheel.
the second and third attempts we hit light rain and FSD disengaged.
I've been in rain of similar severity in a Waymo and it kept going.
So Tesla has work to do, but they can absolutely both solve it.
then the next level question for Tesla versus Waymo is, um, which one of their business
models can actually survive and thrive?
I think they will both be successful, but if there is a race in self-driving, it's not to
deploy a car here or there.
It's...
Since this episode was filmed in January of twenty twenty six, Alex Roy, Warren Honor and
Paul Pham set the first fully autonomous zero intervention cannonball run record from L.A.
to New York in a Model S.
In the middle of winter, through extreme cold snow, ice, slush and rain, FSD drove 100
percent of the three thousand eighty one miles of the journey.
They made it across the country in 58 hours, 22 minutes, nonstop (including charging).
Tesla FSD version 14223 is a game changer for personal autonomy.
Welcome to the FutureCaster Podcast where we give you a front row seat into the future of
business, life, and human potential.
Today, we're speaking to ALEX ROY legendary endurance driver, mobility futurist, and
boundary pusher at the intersection of man and machine.
From setting cross-country speed records to exploring the cultural consequences of
autonomous driving, Alex has always lived at the edge of what's next.
We explore how AI is already reshaping Motorsport today.
Then we dive into what's coming.
Racing BEYOND EARTH.
This episode blends the real, the radical, and the road ahead.
Hi, Alex Roy.
Hi Bates, it's great to see you.
I really want to share your story with the audience.
uh It's a fabulous story.
let's start from the very beginning, how you got interested in racing, and take us all the
way to today and even into tomorrow, where you're sitting right now at the frontier of AI
and the future of motorsport.
Well, thank you.
uh I grew up in a car family.
My dad had a car rental company.
We would spend summers in Europe for three months, rent like a family car.
When I turned 16, I had to learn his permit.
He let me do all the driving.
And so that was an amazing uh bit of driver education for an American teen.
ah I think a pivotal moment for me was watching this nine minute video of French director
Claude Lelouch sending a car across Paris, running every red light.
I saw it when was in high school, it was an illegal VHS, my dad has hundreds of dollars.
I'm going to do that someday.
This is before the internet, so there's no information about how that actually happened.
Reading car magazines and I heard about the Cannonball Run, but I grew up in New York
City, so there wasn't really car culture and certainly you couldn't go go-karting in New
York in the 80s.
The concept of endurance driving,
I followed like 24 hours of Le Mans and Dakar and Baja 1000.
But as a New Yorker, it just really wasn't like, you it wasn't like really an option.
But when my friend, before my father passed away, told me that he'd wanted to enter the
cannonball run race, which was an illegal race across the United States in the seventies,
but that he hadn't done it because my mother threatened to leave him because he had two
young kids.
But so after he passed away, I'm like, I'm like, I'm going to go do that.
Now I'd been to a bunch of racing schools and I'd done some, track days.
Uh, but I, and I wasn't really interested in track racing.
was very interested in all the technology that cannonball drivers use in the seventies to
get across the United States as fast as possible without getting caught.
And, know, in back in 2000, 2001, the technologies were so much better than they were in
the seventies, but especially things like map quest, which is dead today, but was cutting
edge 25 years ago.
And so, you know,
Around that time, I was a tech investor, an angel investor, I was really thinking, what if
I wrote a business plan for how to break the Cannonball run record, but I treated it like
a startup, like a tech startup.
I did a SWOT analysis, 18,700 oh police departments.
Where do I get the geospatial data on the roads?
How much can I...
how much data can I collect about police location?
so treating it like a tech startup combined with a bit of how like Lamont teams, you know,
manage 24 hours of racing and then having been a gamer and obsessed with military
simulations, I was like, how do I tie this all together?
And so in 2003, entered a car rally called the gumball 3000, right?
And that our colleague, Nick Connor here.
Um, but going into it, I was not interested at all in the parties or the fun aspect.
I was interested in it solely to determine how one gets drives a very long distance at the
highest possible speed without getting caught.
And the end goal being to break the cannonball run record.
And so, uh, I arrived at the, on the gumball, you know, 3000 rally dressed like a German
police officer, but that was basically just a cover story for what I was actually doing,
which was testing all the hardware and software I thought would be necessary.
replicate the cannonball run races of the 70s.
And so my car was, you we had a fuel cell, know, laser jammers, radar detector, know, gyro
stabilized binoculars, everything that the folks had in the 70s, but better.
And over the course of five years and dozens of car rallies, determined that it was
possible to cross the United States or Europe or North Africa and parts of Asia at 150 to
180 miles an hour.
dressed like a German police officer and not go to jail.
And so my theory was, well, if we can do this in uniform in a fake police car and we
remove the uniforms, we are unstoppable.
And so a pivotal moment came in 2000, I think it was five.
uh Google acquired Keyhole at some time around that period.
Keyhole was the Department of Defense uh satellite system used for military work.
And when that was superseded by the next better system and became Google Earth, I'm like,
well, this is it.
This is the tool I've been using.
I've been looking for.
so uh drove cross country again, used a FLIR uh thermal camera on the car, recorded the
latitude and longitude of every incoming uh signal that could be police, built a
spreadsheet of all those data points, uh placed them in a data file, exported it to a GPS
unit.
got eight more such units, put them all in the car and went across again.
so having driven that route multiple times and then looked on my desk, one screen was the
driving video.
The other screen was Google earth driver POV matching speeds and rolling through the
terrain.
As if we were doing planning a military operation, it was very easy to correlate the
latitude longitude of incoming like radar signals and laser signals.
to terrain features where a police car can hide.
And so culled that list down to create a data set that I was highly confident was accurate
to police location and went across again.
And over the course of multiple drives, it became apparent that I was correct.
And if we added a spotter plane with triple redundant communications to the car and a
fleet of scout cars, it would be virtually impossible to be stopped by law enforcement.
So that was the beginning of my racing career, but I would call it more of an ersatz
racing career because this was not a form of racing, is legal or even recognized at 17
years ago.
Uh, but the path from that to today was we went public.
We broke the record.
got a book deal and a movie deal.
That was my ROI on this ridiculous business plan and was instantly contacted by people who
had participated in the DARPA challenges in 2005, six, seven.
DARPA uh put on these events where variety of university teams had to try to cross through
urban environments, desert environments, uh autonomously.
And all these guys who emailed me having read the Wired magazine article about us said,
you use the same hardware and software we do, except you're doing it for selfish and
illegal ends.
And yet we're doing it to make driving safer.
You should come speak.
went to MIT.
went to...
the FBI went to like pretty much every government and university that was working anything
autonomous and talked about how we did it.
Began angel investing, launched a podcast called the Autono cast about the future of
transportation and uh spent 10 years investing in the sector.
You know, talked to met everybody in autonomous racing, electric racing, and eventually
was recruited by Argo AI, which was a self-driving car company.
that raised $4 billion.
That was Ford VW, a self-driving car company.
I became an executive there by accident because they had read my book and seen the
Cannonball movie.
And when that ended, I'm like, this is it.
I'm going to spend the rest of my life investing in technologies that move things more
efficiently or faster or safer.
And maybe even all three.
And so today I have a venture capital firm called New Industry Venture Capital, and I
invest in everything I've ever used and the next iteration of that.
I remember coming to see your film at the LA Auto Show and it was amazing.
I have to say thank you for showing that to everyone.
was, if you haven't seen it, you must see it.
Where can people find it?
ah It's the movie is called apex the secret race across America and it's on iTunes and
Amazon Prime
one thing that I noticed in the room that night, there were men in blue, there were a lot
of police officers, and you were talking about how you got away with all of this stuff.
So I'm wondering, in the end, were you pardoned
I did get some curious letters from the FBI and some other agencies, but the Bureau's
interest was primarily in understanding, uh in understanding how someone with no
background in law enforcement or the military could
operationalize a plan to get a car across the United States at high speed and not get
caught.
Their goal was to prevent someone who was a bad actor, say transporting weapons,
explosives, If such a person knew what I knew and taught themselves the techniques and the
technologies that we'd used, how would they be stopped?
And so everything that did not go in the book that my attorneys had omitted uh was shared
in several years of kind of private conversation with various law enforcement agencies.
And they amazingly are a lot of car guys.
And I got a tour of the automotive and technical unit in Quantico, Virginia, where I was
one of the speakers.
the stuff they were doing, they all said the same thing.
They would love to do the cannonball.
And they had technologies and techniques.
that would make it, they could break my record.
This is back in the day,
I'd like to hear from you basically how you think AI is already reshaping motorsport today
from the pit lane to the race strategy room to predictive analytics.
the fundamental misunderstanding and opportunity in applying AI to how cars drive is to uh
understand the box in which almost all development has occurred.
And it's a semantic box, which uh in other words, almost every application of technology,
even rules based pre AI hardware and software uh going back to say early traction control
and and like braking systems have a form of primitive intelligence, but everything from
there.
all the way through today and driverless cars like Waymo's, all of it is designed in the
uh inside a zero sum box.
So the more the technology can do, the less the human should do.
And the end goal of such a way of thinking about technology is the elimination of human
driver altogether.
Now that if safety was the only goal.
of adding technology to cars, that would be a worthy goal.
However, Cars are both transportation, but they're also transformation.
If cars were just transportation, there would be no reason for anything better or more
expensive than a Toyota Prius to exist.
And yet Porsche exists and Ferrari exists.
The reason sports car brands and car racing exists,
certainly up until now and for the foreseeable future, is to demonstrate how much can we
amplify human skill.
Because humans need catharsis and agency and they need, they need and want to buy things,
clothes, tools, cars that both reflect who they believe themselves to be, but also amplify
the best they think they can be.
And that is reflected in motorsport in Formula 1.
And so.
All automotive development with AI thrown into it, unfortunately, is at odds with that
because the better cars get at driving, the less necessity there is for a driver.
And so that is a tragedy and that it's a choice.
It's a cultural and civilizational choice.
And it's the difference between, um, an Iron Man Suit and the WALL- E Chair.
Like, which would you like to have?
Optimally, you have a choice.
If I want to do stuff in my Iron Man Suit great.
But when I want to relax, I want a Wall-E Chair.
Unfortunately, every car maker on earth with one exception is building Wall-E Chairs.
That's what they see the future of cars to be.
the cars we buy.
So that's really unfortunate because the, the human autonomy, which just a fancy word for
freedom and agency, the amplificatory power of technology lies in building an Iron Man
Suit with wheels.
We have not yet seen that.
in the consumer market or even on the horizon.
So that's consumer vehicles.
We have seen in racing the application of technologies to make the cars go faster without
eliminating the input of the human driver.
And have you seen the movie Senna?
it's really a movie about the man, the best driver on earth facing forces beyond his
control and specifically a rival team
In the movie, Senna is just pure skill, like analog car, no tech.
It's just human amplified by car.
And McLaren puts in an active suspension system.
And if I recall correctly, based on the vehicle's location around the track, the
suspension would preemptively reposition itself such that the car's physics are optimized
for every corner.
And so a driver of lesser skill in that car,
could defeat a driver of greater skill in a pure analog car.
Senna and all the great drivers of analog cars, they are always thinking, how could I
position myself not just for maximum speed out of this turn, but what do I have to do
entering the turn to position the car and load the suspension correctly?
So complaints were filed after several races, and that active suspension system was then
banned.
So we've seen in the...
recent decades, a variety of technologies added to Formula 1 cars and also in Le Mans
other racing series, which do amplify the abilities of drivers, but they're often limited
to a single team.
rarely see a spec series, which has everybody benefiting from the same tech.
these, this tech is in Formula One.
It's so wildly expensive to develop it, to deploy on a single car.
It's just not feasible.
And it's really not feasible for anything, anyone, but the biggest teams.
So that is.
fascinating to me because in a perfect world, the pinnacle of Motorsport, everyone would
have access to the cutting edge technology such that driver skill becomes the thing that
matters again.
And yet that's not possible because of the R &D budgets required to develop these
technologies.
That being said, uh there was a uh fascinating article,
think it was in car or automobile magazine in the early nineties.
uh Gordon Murray, the designer of the McLaren F1 and two or three other like world class,
like F1 racing engineers were asked, what will race cars look like in 50 years?
And each of them had a completely different vision, but they were all really ambitious.
One of them was resembled a, like a, like an eagle about to take off, like wings, like
half spread.
it's, it had like 20 wheels and it looked like a train with wings that were spreading.
And the designer suggested that the driver would only have one input.
It was a button, one button.
And that the vehicle would drive itself around the course and the optimal path.
And the button was aggressive pass because they felt that only a human driver would be
able to push beyond what physics and the map could tell the car what to do.
It would still need to be some final human in the loop.
The other vehicles look different, but basically had the same input set, which is a
button.
Maybe there was like, you know, left, right, like rocker switch.
And so they all defaulted to letting the car do most of the critical path planning and
left only like a final, like, you know, risk amplification choice to the human driver who
remained on board to suffer the consequences of a which is a very essential part of human
driving.
There have to be consequences to decisions or you might as well stay home and play video
games.
So
That McLaren, uh, with the suspension in the Senna movie and that article in automobile to
me are like the two most important pieces of content one can digest to understand what the
future racing could be.
And so it is essential that we pick Ironman as the default setting for teaching kids how
the world works such that they can improve themselves and the world, as opposed to
starting off from a WALL-E setting, which is
Everything will be done for you and to do nothing is a viable choice because that lesson
is the, it's the end of progress in any area as we know it.
Uh, this is, you know, this is what I hope the future of all driving and racing will be,
will be a search for better versions of an Ironman suit.
And unbelievably, the best example of AI applied to a sport, uh,
isn't in the real world at all.
There's a movie called Real Steel.
Hugh Jackman with robot boxing.
This is the third most important piece of content for anyone thinking about putting AI
into any type of automated machine or device of any kind.
In the movie, in the future, boxing is illegal.
Human boxing.
Too dangerous.
It's cruel.
And so robots box.
But the robots are the creation of a team with a singular charismatic team leader.
And so the robot designs and fighting styles reflect that team leader, which solves the
problem of whom should we root for?
No one really roots for a robot, but you can root for the robot if you like the guy who
built it.
So in the movie, the robots are...
They can fight fully autonomously.
However, the hero robot is smaller and weaker than the antagonists.
And Hugh Jackman is a retired former boxer.
And as he takes his robot with his son through a series of matches, they face harder
opponents.
And at the end of the film, they face an opponent they cannot defeat with the...
uh
pre program fighting algorithms.
so Hugh Jackman puts on a pair of gloves, haptic gloves, and he allows his robot to
proceed using its built in defense algorithms to block.
But when it comes time to attack, he inputs the attack in the haptic gloves.
And because he's both better than his own robots, you know, offense and behaves
unpredictably.
And against what the opposing robot predicts their bot will do, they win.
And so this X factor of human input, uh human input interfacing harmoniously with an
automated system is that is the future of racing.
If we want racing to survive, as we know it, if we want humans to be in the loop, we have
not yet seen anything like that in cars.
And sadly, you know,
The companies that should understand this best, Porsche, Ferrari, all the F1 teams, to the
best of my knowledge, unless they have a secret program somewhere, none of them are
thinking this way.
When you go to a racing school today, even like a top tier, an F1 school, you go there,
they're recording telemetry in the car, they're recording video, you get back to the pit,
you get back to the classroom, you look at a map, they show you your sector times.
But the real application of AI in racing, like the dream, would be a real time, in-car,
driving instruction system.
You're driving, the vehicle would predict you're about to make a mistake.
It would warn you such that you preempt it or mitigate it.
And if you failed to do so, it would take action to prevent a crash.
And as your skills improve, the boundaries of the vehicles
Performance envelope would expand or contract based on your inputs and over time would
allow you to take almost unlimited action with unlimited inputs.
If you were demonstrably better than the simulation and prediction models of the vehicle
itself.
uh We're a long way from that,
allows different drivers with different skill sets to compete against each other,
I do think that as a training tool, that would be optimal.
In a perfect world.
An AI driven real time driving education system would exist in all passenger cars and in
racing schools.
When you got to the actual track, it would invert.
So the vehicle might have a far outer boundary that would disallow a crash, but just but
inside that boundary, anything goes.
uh Because ultimately, like what is the perfect sports car or race car?
Perfect sports car is a Porsche 911.
that will you do whatever you want, but won't let you hit anything.
And a race car is exactly the same way.
AI has to improve dramatically.
It has to pretty much get to fully driverless operation before you can work your way back
and build a product that would accommodate or allow for what I'm describing.
you really can't build this system until a driverless vehicle can lap a track with a
record time.
And once it can,
Then the envelope of the vehicle having done so is the context inside which you'd build a
user interface for humans to race inside that.
I'm hearing you say is that the best form of intelligence for racing might be when it
helps us to become more superhuman.
And it enhances our own ability.
And it's our intelligence partner in the vehicle during the race, but also gathering all
that data that takes
so long to analyze in the past, but now it's in an instant, it's real time.
the word I try to use, um, is augmented because right now you have driverless cars and
cars with driving assistance.
Driving assistance is by design and semantically a zero sum game, but augmentation is not.
So I hope that people will start building systems that augment and amplify human skill.
And then the words will swap.
Uh,
There is one company today.
It's ironic because it's Toyota that is trying to build an augmented driving system.
they have a system called Guardian, which does function as I described, but is many years
away from deployment in the form I'm describing.
Toyota also
very, you know, as acquired a number of companies, mapping companies.
Um, so they can build high definition maps that would allow for driverless operation.
Ultimately high definition maps of the world will be available, which will allow, you
know, the, you know, any car to accurately predict and avoid crashes the same way an F one
team would have LIDAR mapped every track and no down to like the millimeter.
where there's an imperfection and optimize our racing line to avoid it.
it's the quality of data and is more important than quantity of data.
on your race years ago, you had a plane, right, with you that gave you insights and
foresight and everything, So could drones become a cheaper version of that plane for
racers?
And have they already?
I'm so glad you asked.
I, uh, in 2012, I asked my lawyer Uh, I had him write a brief about the legality of flying
a drone, uh, around the perimeter of Manhattan Island.
Um, I had set the lap of Manhattan record at that point, inspired by rendezvous in Paris.
And I wanted to do it again with a drone.
One flying ahead to recon the route.
And one flying behind to shoot video.
And, uh, legally that was very bad idea, but functionally could have worked if the drones
at the time had collision avoidance software on board, which they did not.
So today, um, if one wanted to replace my spotter plane going cross country, you would
need a military grade drone and you'd need more than one because they're, they're range
limited.
my spotter plane, uh, you know, we left New York at night and got to St.
Louis at dawn.
and drive to Albuquerque at dusk.
So the spotter plane started in St.
Louis and flew to Albuquerque and had to refuel, I think once or twice.
And that was a king error, a twin engine pressurized aircraft for like six or eight
people.
um Nothing smaller than that could have even done the speeds we required.
So drones that could fly a hundred, call it 120 to 160 miles an hour over distance don't
exist in the civilian market.
um
Not yet in the military has a wide variety of drones.
So you need several in a row.
The missing piece is a software platform uh that could integrate the data from the drones
with radar and laser data, plus weather data, plus traffic data.
And,
predict choke points in advance over time, and then integrate with real-time data coming
off the drone and check it against like an HD map to see if a new pothole had emerged.
And then you could plan around that.
These platforms exist uh in the military, like intelligence and like real-time data
aggregation and processing, and then uh suggested decision-making, these things.
I Palantir does things like this.
But there's nothing in the consumer market.
uh
like that yet.
I'm gonna pass it to Nick now
Yeah, all right.
I wanted to just touch, go back to sort of man and machine racing.
know, I know you're kind of you're saying like consequences of mistake.
That's kind of what everyone's going to pay for.
um I think that also the human story is like a really big deal.
I mean, I, I, I believe you're correct.
And I hope that, uh, investors and companies see it, um, because if they don't, we're
going to get more robo race events and what, what is possible in human centered AI
supported racing, um, will be delayed 20 years.
If somebody loses a billion dollars building a really stupid tech focused version.
but yeah, I completely agree with you.
I could see that show in my mind already.
It's 10 years from now, 20 years from now where, you know, some kid has got a crazy theory
about how to build a racing AI and it's make or break change his life.
And if he, if his car wins, there's a $10 million prize from a fund to fund him to build
dual use drones for military and civilian use.
Like that's a great human interest story.
And, you know, the less human input, the less interested I am in watching.
um It's just, it's just the way it is.
mean, in its purest form, mean, Formula One is interesting because you have both a
technology race and a human race at the same time.
And the winning team is the one that nails both.
If they miss one, they're dead.
Yeah, totally.
You know, and if you think like there's, mean, you know, we're talking like the pinnacle
of racing Formula One.
Okay.
But I mean, there's so many racing formats that we love like the, you know, like the Baja
500 and you know, you know, I mean, we're sort of the equipment is wildly substandard for
what you're doing or, know, or, mean, or exactly.
I look even
Even though, know, we did a bunch of gumballs, right?
You know, and I mean, it was the adventure of it and sort of overcoming adversity, not
like having the pinnacle of anything.
it's inevitable that a variety of technologies, hardware and software will proliferate at
all levels of motor sport.
But I think that they will be in the same way that like streaming music has, you know,
increased the amount of music.
I don't know it's increased how much music is listened to, but it's certainly increased
the variety of music.
Any one person can listen to, cause you're no longer bound to read by having to buy the
album.
A basic sub gets you everything.
So, uh, but uh the corollary is that sales of vinyl records have gone up and that is, I
believe, uh, a cultural reaction, um, in pursuit of authenticity and ritual.
cooking a meal for yourself or for people you love.
so ritual is matters.
The ritual of putting on a record.
Is very clearly something that for which people derive value.
And so I believe that as technology, I want to say invades say technology propagates into
car racing and every, every subset of it, supply chain, sourcing, logistics, support,
everything, recruiting everything.
there will be a uh parallel rise in interest in.
a period correct ritual motorsport.
So if I were to organize a cannonball run today, knowing that it is virtually impossible
at current levels of technology, entire technology to break the current cannonball record,
I would organize a period correct race of vehicles from 1979, the maps, the parts,
everything 79, no phones, nothing.
The only exception would be a
paper map that is up to date with the roads, no, nothing will be allowed post 79 except
tires and a paper map.
it's why people like us did the gumball 3000, I think, because we're trying to get closer
to something which no longer, longer existed, know, cross country racing.
I think that vintage racing period, correct with consequences and old hyper expensive
vehicles is going to come back in a big way.
I think Mark honor to have raced and survive some of those races in the same way that
people want to climb Mount Everest.
There is no automation can replace climbing Mount Everest or going to the bottom of the
ocean because the consequences are there.
And
The elimination of those consequences is the elimination of what it means to be
fundamentally human.
And I just don't see that because of wanting to go away.
let's fast forward, uh you say like 20 or 30 years.
I I think this stuff's coming quicker, like much, much quicker.
some short period of time, um, like every car is going to have like the autonomous
capability.
So let me ask you about like Tesla versus Waymo.
Cause I know you know a lot about this.
Waymo sort of, you know, video, but it's also using LIDAR and radar and other.
technologies.
So at Tesla, they think they can crack it just with intelligence and video.
I ride Waymo's almost every day, but I also own a brand new Tesla and I love it.
Tesla that will be able to deploy vehicles driverlessly with camera alone.
I'm convinced of it.
However, there's a term in autonomous vehicle development called an ODD, which just means
operational design domain.
Waymo skeptics will say, oh, Waymo's geo fence is small.
Tesla's operate anywhere.
These are meaningless words because an operational design domain, is how the engineers
discuss the vehicles comprises area of operation, uh the weather, which in which it can
operate, know, shadows, know, extraneous events, uh adherence of, you know, pedestrians to
local law, adherence of local other vehicles to law.
So you go to Singapore, your prediction of behaviors of pedestrians and traffic is very
easy.
But if you go to, you know,
Calcutta or lower Manhattan.
It's very difficult to predict.
the real question is, uh what is the uh perceived level of safety of a Tesla before people
be prepared to ride driverlessly?
The second question for Tesla will be at what point, like what are the acceptable
limitations for a consumer for what that Tesla can do driverlessly?
To be specific.
A Waymo today can operate in a pretty wide range of weather, not heavy snow, but most
other weather.
And it can operate in, you know, simple to moderate complexity areas.
And it can, it can do something that Teslas cannot, which is do pick up and drop at
airport curbside in Phoenix, where I live.
So a Tesla, in order for it to be able to compete directly with Waymo, would have to be
able to do all the same things.
And so with camera alone, it is more difficult to perceive and do path planning and
prediction with fewer sensors.
mean, you just can't see as far as well.
You can still deploy, but the speed at which the vehicle could deploy in marginal weather
is going to have to be lower than a Waymo, which will have more sensors and more
redundancy for perception.
For example,
I just drove cross country a few months ago in my Tesla to set the full self driving
Cannonball record.
And, uh, The first time we were successful, was 99.4%.
I forget the percentage.
was well into the nineties percentage FSD engaged, no hands on wheel.
the second and third attempts we hit light rain and FSD disengaged.
I've been in rain of similar severity in a Waymo and it kept going.
So Tesla has work to do, but they can absolutely both solve it.
then the next level question for Tesla versus Waymo is, um, which one of their business
models can actually survive and thrive?
I think they will both be successful, but if there is a race in self-driving, it's not to
deploy a car here or there.
It's...
Waymo needs to move past, you know, wholly owned fleet, robo taxis, know, fully
verticalized in a handful of cities for them to do that.
They need to take the sensor hardware, which is large and many, and reduce the form factor
of the hardware such that it is conformal and invisible and can fit on preexisting pasture
car designs that are beautiful.
Until then Waymo doesn't have a, that businesses is closed to them.
Luminar, a lidar company, built a system called a halo.
It's a little lighter that goes in the top of the center of the windshield of the Volvo
XC90.
And I think one of the Polestar models, it's as beautiful as you can make a sensor look
that still changes the profile of the car.
So Tesla has already solved this.
The cameras are built into the car.
They're invisible, but Tesla would have to get their perceived safety.
to the level of Waymo and deployment before Waymo miniaturizes their hardware form factor.
That's the race.
Because once Waymo gets there and they can license the hardware and software to any OEM
that wants it, then OEMs can start selling cars that are equivalent to Tesla capability.
And then Tesla is just another car company.
The wild card is that there are...
at least two or three other companies that are trying to build, you know, camera first,
ultra lightweight, end to end neural net AI that could replicate what Tesla's doing.
That could also with very small format cameras, uh, mobile eye out of Israel and wave out
of the UK.
If one or both those companies can solve autonomous driving, um, and get their license to
an OEM before Waymo does.
It becomes very dangerous for Waymo.
uh They still have a business.
They have a fleet on, know, robot taxi business, but they would not have a time advantage.
Their only advantage for any these companies is time.
Long-term, there is no technical advantage.
Every, like the electricity companies and the aviation companies, everyone will solve this
eventually.
Whoever solves it first will have some period to exploit that advantage and try to get
exclusive, you know, partnerships,
ago, you were to parachute AI people into today, they'd probably go like, yep, you've got
like early AGI.
You know, at some point we can accept that like that's gonna come.
I mean, we're probably like sort of 12 months away from
you know, an AI mathematician being better than any human mathematician, for example,
right?
You know, maybe like 18 months away from an AI coder being better than a human coder.
um So like super intelligence, you know, it's on the horizon, it's pretty likely.
And I don't think most people get their head around what that really means.
I mean, super intelligence means
It's an AI that can do everything better than a human can.
So I'm gonna sort of bring that back to Tesla Waymo.
Like when that happens, the benefit of hardware versus intelligence disappears.
uh So obviously none of us know what's gonna happen, but the intelligence to see in...
dark, difficult situations, interpret like different movements, all that kind of stuff,
right?
So, I mean, whether it's Tesla or Waymo or somebody else that wins, there's this future
where nobody's driving a car anymore.
Is that a declarative statement or a question?
I think we're going to move forward to some future point, which is going to be where an
awful lot of people don't know to drive.
I agree with you.
let's put some like timelines on that.
So without question, there is some end state like the, you know, au tono of motive
singularity where the last person who's ever driven a car decides she's not driving
anymore.
And then the last person is driven.
It's over.
That's that day will come
there is only one city in America where you can maybe two.
where you can live without having to own a car, New York City and maybe Miami Beach and
like maybe subsets of some other cities.
But New York's the one because it's got the best transit system and no one else comes
close in the U.S.
uh In Europe, in almost every major city, you may never need to own a car.
Why would you?
What's the point?
So that's doesn't matter.
But let's imagine autonomous vehicles work today perfectly today and we're affordable
today.
uh
What would happen is the average car is owned for, think, 11 years.
So unless they were cheaper, dramatically cheaper and subsidized to replace your
personally owned human driven vehicle, that human driven vehicle will, you'll have it till
the end of your lease or, you the end of your financing.
And so either three years at the short side or seven years for the rest of folks.
That car is still going to be on the road, sold used cheaply to somebody for another
Basically until it's 11 years old, maybe 15.
That's assuming autonomy was available and worked today.
So add 10 years to a timeline for a deploy for the beginning of a ramp to ubiquity.
Now we don't have autonomy today.
And what I mean, autonomy, I mean a vehicle that can drive almost almost anywhere a human
would need to drive.
You don't mean most conditions that such a vehicle will not exist for.
I think seven to 10 years.
uh So, you know, Tesla, they're what they're doing is not magic and enter an end to end
neural net um can grow and learn, but for it to be not just good enough to drive, but
perceived to be safer than me is a much longer term problem than actually getting the tech
to work.
People have to believe it works.
That requires often growing up with it.
So my daughter's seven.
She believes that a Waymo and a Tesla is awesome.
And she already believes they're better than me.
And she's like, dad, why are you in the seat of the Tesla?
Like, isn't it just like a Waymo?
I'm like, not quite yet.
So by the time she's 15 or 20, she is already, for her, it's autonomous vehicles are
context and they're a given.
so that's, you know, 11 to 14 years from now, assuming that the vehicles can work
anywhere.
Cause until they can work anywhere or almost anywhere.
the future is going to be dual mode vehicles.
So, and this is what I believe is going to be the next 20 to 30 years of cars.
There will be the almost every major urban center in America and in the first world will
be high definition mapped.
if it's not HD mapped, it'll be mapped over and over by like low risk cameras like Tesla's
have.
And this, you know, consistent 24 seven mapping will make it possible for driverless
operation in every major city.
in four seasons in the first world.
But outside of the cities, it'll be a mixed bag.
And in order to sell a car to somebody who wants to be able to go anywhere, there'll be
dual mode.
So this is the car I foresee will be available in, you know, seven years.
It's a car that can be driverless inside areas where insurance companies will insure it.
And outside those areas, you have to drive it.
But the driver assistance will become augmented systems.
so cars will have steering wheels.
You can drive it anywhere you want.
If you hit a boundary where driverless operation is possible, you'll have you'll be able
to choose
dual mode vehicles are going to be everything in this country.
Your prediction about autonomous ubiquity is an absolute certainty much sooner in
countries like China, Singapore, Japan.
parts of Europe.
And it depends on the verticalization and like political philosophy guiding that country.
So Kim, you're the futurist.
So with everything that we've discussed, so both future motor racing and the future of
autonomous vehicles and our experiences, what are you seeing 20 years hence?
One of the things that I do believe in, I think 20 years out, is that we're going to have
suborbital racing.
uh Once that technology is cheap enough and affordable, I also think the meaning of rider,
driver is going to stay.
I have this vision of the future
where riding and track also become entertainment and physical performance.
So I will give you a scenario of the future with a character.
Zanthe loves to hit the O-ring.
It's a global entertainment membership track club.
She subscribes to gaming shows and film simulations and chooses the era she plays in.
She physically travels around the track on a hydrogen fuel O-mobile.
Imagine that the O-mobile looks like a motorcycle, but like a donut with like a big O.
And her temperature optimized Tesla suit lets her physically feel the thrill of all her
senses.
The Omobile doubles as a personal gym as it helps to burn a thousand calories per hour via
vibrations and electric pulses.
She subscribes to licensed holographic characters created by what I'm calling Zollywood
Studios in the future.
They move from simulation to simulation.
She can delete or add new characters to her story worlds.
Holographic scenery fills the O track.
uses holographic lasers and weapons to attack her enemies.
Zanthi likes to choose a different ending every time she visits this world.
So I can imagine that we're going to have our personal mobility racing devices in the
future, and it's going to become also uh human performance and entertainment as well, not
just automotive racing in the future.
All gaming simulation, everything you're talking about the, racing on the O device.
Kim, you're talking about, I want all of that.
But if the consequences aren't actually physical injury or death, then the consequences
have to be one to 24 hours.
You can't play.
You have to watch everyone else
Alex, thanks for coming on the show today.
adore you and just brilliant thinking as usual.
Thanks for having me.
love you both.
Wait, no, I forgot to tell you.
You see that picture over my shoulder right there?
That's me and my daughter and my Morgan three wheeler risking our lives for the
consequences that make life great.
Ciao.
If you're enjoying the show and appreciate what we do, the simplest way you can support us
is by hitting the subscribe button and sharing this episode with others.
And by doing this, our commitment to you is we'll do our very best to make each and every
episode better than the last.
See you in the future.