December 19, 2023 • 33 mins
Ryan Myrehn and Paul Mitchell speak with a pioneer in autonomous vehicles, Sebastian Thrun. He's the CEO of Kitty Hawk and the founder of Google's self-driving car team, as well as an educator and co-founder of Udacity. Sebastian guides us through the future of personal flying machines, the ethics of AI, and how the work of the Indy Autonomous Challenge will make the roads safer for everyone.
From Straw Hut Media
From Straw Hut Media
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:04):
The world around us is changing fasterthan ever before. Ideas once only imagined
in science fiction are becoming a reality. Throughout the course of our amazing twenty
three episode season, we'll speak tosome of the greatest minds in robotics and
artificial intelligence to discuss the groundbreaking workthat's fueling it all. I'm your host,
Ryan Marine joined me and my cohost Paul Mitchell, the president of
(00:26):
the Indian Autonomous Challenge and see whywe call this the inside Track. Today
we're speaking with a pioneer in autonomousvehicles, Sebastian Thrune. He's the CEO
of Kitty Hawk and the founder ofGoogle's self driving car team, as well
as an educator and co founder ofFeudacity. Sebastian guides is through the future
of personal flying machines, the ethicsof AI, and how the work of
(00:48):
the Indie Autonomous Challenge will make theroads safer for everyone. Sebastian, it
is a thrill to talk to you, to have you here on the show.
I know from talking to Paul overthe last few years, your vote
of confidence in the Indie Autonomous Challengehad a lot to do with getting it
off the ground. And for you, this AI in mobility, This has
(01:12):
been a big part of your lifefor the past several decades. So for
you to see how that has evolved, what pride do you take in playing
a pretty major part in getting itto the place where it is today.
Look, it takes many, manypeople to invent a new technology like this,
and I happen to be around whenDARPRA launches very first challenge two thousand
(01:34):
and five that we had to buildcost right the desert. I happened to
be around when Google decided to getinto the game and build Weemo, which
I have founded. And I'm superexcited about the INDI five found Out Challenge.
I think that's a great thing forhumanity and it's a wonderful thing to
do. Let's go back to theearly days and to DARPA, which in
many ways is the genesis of whatthe Indie Autonomous Challenge is about. What
(01:57):
drew you first of all to thatcompetition. Look, there was a long
time ago, two thousand and five, the US government had reserved I think
two million bucks as price money forsomeone who could build a car that could
drive itself one hundred and forty milesto a desert trail, and I got
drawn in not because I cared aboutdesert driving, but the care about human
life. Earlier in my life,when I was eighteen, I lost my
(02:19):
best friend to a traffic accident,and I felt reminded that by and large,
humans are not the safest drivers onEarth, and I was intrigued thinking
I could make a machine that coulddrive safer than people would drive. You
know, Sebastian, when you andI were talking in twenty seventeen, In
twenty eighteen about the concept of highspeed autonomous racing, I remember you talked
(02:45):
about kind of the first frontier ofautomation and self driving at lower speeds,
trying to navigate the urban suburban environment, and that the next frontier was,
you know, moving to high speeds, extreme speeds one hundred miles an hour,
you know, plus and testing whetheror not the hardware, the software
(03:07):
stacks can handle those environments. Soyou know, we've been doing that now
with competitions throughout the US, andwe're going to be running in Monza in
Europe here in June. But wheredo you see the contribution of high speed
automation playing the greatest role in savinglives in being part of the overall goal
(03:32):
that we all have of autonomous mobility. Yeah, look, that's correct.
I always believe that the faster thatyou can go is the hot It's going
to be a headled safely. Soif you want to be challenge technology,
going at the envelope of it's possible, physically possible certainly challenges technology the most.
The Dopogren challenge, I think wasan average but tendy miles an hour
(03:54):
was very slow. The Google ofAMO costs can dive highway speeds, but
they try to stay away from extremedriving. But when it comes to retesting
the technology, I think extreme drivingdriving at the envelope that's physically possible,
it's going to be the hardest test. And I love that you're doing this
and to be living in the extremelikely and the autonomous challenges. It's an
(04:15):
extended edge case. Basically, everytime the cars are on the track,
not only is it testing the technology, but also doesn't that say a lot
to the consumer public as well,to prove the efficacy of this technology in
an environment that they wouldn't expect tohave to be a part of themselves.
Yeah, it depends on where theworks, but I honestly believe eventually a
(04:40):
robotic car will surpass the ability ofa human driver, especially if we let
their robotic car control wheels and brakesindividually, which people can't do as easily.
And given that level of freedom,and I think the technology that's being
developed, we'll trickle down into safercars for normal drivers, because it's been
always extreme racing technology that let ustest out what is feasible and possible,
(05:03):
and we then took this technology andbrought it back to normal cars for normal
safety. One thing we've been workingon and thinking about some of the teams
and some of the sponsors and technologytechnologists that are involved in India Autonomous Challenge
is also how do you transition thetechnology to human race car drivers in parallel
(05:27):
with trying to advance the capabilities ofthe of the AI driver. And we
often think of it in the contextof you know, advanced fighter pilots,
right like if a human race cardriver has the ability to have three hundred
and sixty degree perception, you know, either through their their their visor or
or even their race control team cankind of provide indications to them if if
(05:49):
vehicles can you know, take overin the event that you know, unfortunately
the driver misses a red flag warningor or you know, is out of
control. That some of these technologiescould allow human pilots of race cars to
be closer to what we see.And you know, let's say the advanced
fighter pilots that are in F twentytwo. Yeah. Look, I would
(06:15):
say that even today we have technologyin cars that make us massively safe.
While that goes way beyond human capabilities, the most notable is abs. The
ability to see if a tire isslipping and then release it very quickly to
make sure it doesn't constantly slip.Is the technology that saved uncounted numbers of
lives. It's the racing field thatdevelops these technologies first. And I believe
(06:40):
that a computer system and principle doesn'tjust have a three six view that people
don't really have, but I alsocan look at details of the vehicle that
people can pay attention to, like, for example, is a tire slipping
right now, and really help makingthe driving system much more safe. And
that's nothing new. It's been aroundfor decades now and i'ven't benefit from it.
Everybody benefits from it. And Isee this to progress. I think
(07:01):
also get to the point where itmight someone be impossible to pressure a car,
it might be impossible to hit epestia, It might just be impossible to
do these things by having a smartcar that looks a lot of shorter and
prevents us from making these mistakes.One thing I've been wanting to ask you
about for a while is is justyour perspective on this whole national global debate
(07:23):
on AI and and for somebody who'sbeen using advanced technologies and AI and machine
learning in the in the applicant andapplying it to mobility, whether it's ground
based or aircraft. What is yourperspective on on AI as it relates to
(07:45):
its contribution to society and how wecan utilize it in ways that are productive
versus you know, do you haveany fear for the role that AI may
play with the chat, GPT andother kind of human communication based AAIYE tools
that are emerging. Great question.I mean, in the early days of
(08:05):
self diving cars, we already usedmachine learning to train these cars so you
could learn from us and from themselves. And now we're entering a world where
we have large scale called large languagemodels and and ms that are being trained
on all corpuses of text and thenshow a credible productivity and capability to form
(08:26):
original pros that translate languages, todraw images and creative art. I would
say we're at a big inflection pointfor humanity, and we had a point
now where these models, these systemscan do everything we ask a high school
student to do within a plus infact, they can do but anything we
ask a college graduate to do,like a psychology major or maybe a marketing
(08:50):
major. Some of them can evenpass the Bar exam ninety percentile. Here
in California, about forty percent ofcandidates fail the bar exams, So it's
obviously better than you average a lawschool graduate. And that doesn't mean that
they're going to replace all these people. Instead, they're going to make all
of us much more powerful. They'regoing to enable us to do something in
our that we've previously taken us amonth to do or a week to do.
(09:13):
That's nothing new, it's actually historicallywe did this already. We went
to a transformation society with the adminof the steam engine and modern agriculture over
the last one hundred and fifty hundredyears. Prior to that, most people
in the world worked in farming.A farmer could make food for something like
four people their family, and ourfarmer can make food for four hundred people.
(09:37):
And that has not been the endof civilization as we know. It's
been the opposite. I'd say we'reall better off. We now live in
a world that's safer, and welive longer, we live healthier, we
have more food, we have moreeverything because of those revolutions. I think
that's just the next step. Thenext step is going to be that everybody
working offices, from accountant to lawyerto what have you, will become one
(09:58):
hundred times as effective. And Iassume that you would see that extending to
mobility and human safety and vehicles andaircraft, etc. That some of these
learning models that are doing amazing thingsin the social sciences and literature and marketing
will emerge further in the field ofself driving cars and vertical takeoff aircraft and
(10:24):
other platforms. Absolutely, if yougoing to San Francisco, you're surrounded by
completely empty cars that drive themselves orcars that have passengers, and the very
first instantiations technology provides safety. It'ssafer than human drivers. I think they're
y more cars and I have todrive roughly thirty million miles without any major
(10:46):
accident caused by those cars. That'skind of amazing compared to human safety levels.
But then the second stage, you'regoing to see accessibility. You're going
to see people who wouldn't use carsto day, blind people, very very
old, people very very young,and be able to partic in transportation with
this new model, and you'll seethe even more cost effective I for,
for example, I'm a pilot.I'm mesually a jet pilot. I have
(11:09):
my own jet, and when Ifly these days, all I ever do
is after I take off, Ipush one button and then I'm being told
by ATC to the left of fiftydegrees, climb up one thousand feet,
and that means I'm going to diala number into a computer and that's all
I ever do. So the machinebasically flies itself, but all practical purposes.
It can even land itself in anemergency. That's something that I think
(11:31):
we're going to see all around seeeverybody now in the future being able to
use these technologies to be safer,faster. You can sleep while you've been
car don't drive it anymore. Youcan you can watch a television movie and
get transportations safer and cheaper than today. Back to Paul's comment about the conversation
surrounding AI right now, one ofthe big angles I suppose in that conversation
(11:54):
is the ethical question. And youbring up the point where the vent of
the steam engine or modern agriculture changesociety, And certainly, if you look
through a broad lens, it changedit in an objectively good way. But
that transition wasn't always an easy one, and there were inflection points along the
(12:15):
way in which the human condition insome instances did suffer from this process of
modernization. So what lessons can welearn from industrialization, for example, and
that process that took several hundred yearsto apply these new technologies in a way
that is more uniformly good for humanityat large. I think it's a great
(12:37):
question Ryan, that you're asking here, and it's a question we should all
be talking about, in part becausethis evolution is so much faster than the
past evolutions we've seen. What itused to take fifteen years will not take
five years, to be quite frank, And it's important that no one has
left behind. It's important that weas a society to use these technologies responsibly
for the betterment of everybody, notjust a few. It's a very broad
(12:58):
dialogue. That's a way right now. I don't think technology can be stopped.
I think what we can do isreally think about what are the implications
overall? These technologies will produce morewealth, more things, more goods,
more abandons. How can make surethat the abandonce that's being produced is fairly
distributed among all of us? Inyour opinion, who bears responsibility for ensuring
(13:22):
that that is the outcome? Isit up to individual agents, is it
up to the corporate world, ordoes there need to be some kind of
oversight from a government structure locally orinternationally. I think it's everybody's responsibility.
It's us, the people that areaffected, as the technologists, as the
consumers, as the voters who voterepresentatives for us into office. Is us
(13:46):
the office holders have been elected todo this, and we've got to beick
smart about it. It's really easyto say it's just outlawed. It just
happened in Italy, at least,said Chad GPT's outlawed. If you go
back in history, you find thatthe Turkish Empire in the Middle East illegalized
the printing press, and that didn'tturn out well for them because that precluded
(14:09):
the dissimnation of innovation and new information. So we have to have a broad
dialogue. But we should not putour head in the scent and say you
don't want it. Should instead saywhat's the most productive, best used that
is fair to everybody and helps usas human race to become better. I
want to go back to your yourwork in the arc of your of your
(14:33):
corporate and technology endeavors. You playedthis critical role in the early and then
the autonomous vehicle and working with theteam at Google to get way more going,
and then you've transitioned into the flyingcar or the vertical takeoff aircraft space.
(14:54):
And I'm just curious as we thinkabout and I don't know if you
can see behind us, but youknow, we've got these one hundred and
eighty one hundred and ninety mile perhour, and I don't know if you've
been tracking the progress, but wenow are passing each other at one hundred
and eighty miles an hour on oldlisteness and transitioning to road courses kind of
(15:15):
you know, how do you seethe challenges of transitioning from the passenger car
environment to aircraft that are traveling athigher speeds. What drew you to that?
It sounds like you're a pilot,so maybe it's just your own personal
passion. And then what is yourthought about how soon are we going to
(15:37):
see that technology? In some casesaugment probably not replaced, but certainly augment
are focused on ground based mobility thatis still we're so dependent on for let's
say short or mid distance trips.I mean, historically the growing made sense
for transportation because gravit it just pulledus all down and it's really hard to
(16:00):
stay up in the air. ButI'd say for quite a while now it's
been clear that we can get thesame efficiency fuel efficiency, cost efficiency in
the air that you can get onthe ground. And what do you gain
by going to the air? Numberone, pretty much all obstacles are gone,
Like all this stuff that you avoidabout the ground, like trees and
houses and curbs and whatever, they'reall gone. Maybe with the exception of
(16:25):
a few birds. There's no catsand dogs in the sky. Then you
get to go in a much muchbigger volume. So instead of having been
paved roads, the roads in thesky they just don't require any pavement,
there's no infrastructure necessary except for theplace where they take off a land and
with a VTOL vertical take off alanding helicopter like you can land and take
(16:47):
off pretty much anywhere. If there'sno alidlines, you just go and do
this in your front yard and yourdriveway, and then once you're in the
air, you go on a straightline. You go much faster, much
more efficiently. We've shown a kiddingtalk that the fuel consumption for a one
passenger EVE to the vehicle is roughlya third bit of a Tesla electric car
(17:07):
per mile, and then even attwice as speed, which probably quadruples the
drag of the vehicles. So it'sit's actually more fuel efficient to be in
the air and on the ground.So from a safety perspective, from an
efficiency perspective, from a cost perspective, from a green perspective, air based
transportation bill completely throvolved GUND transportation goingforward. Can you sour? One of
(17:33):
our thesis in this is that everyI agree with everything you just said,
that the importance of the edge casein air traffic is that much more important
because of the velocity in which anaccident happens. The way in which an
accident happens. Now, I thinka lot of that can be managed with
(17:56):
things like the carbon fiber monocoques thatare in raised cars. And we know
so much about how to protect thepassenger the critical cargo. But one of
the luxuries you have with passenger carsgoing twenty thirty miles an hour is they
can crash and you just sort ofput them back together and you know,
hopefully nobody gets hurt in it.So in your mind, how do you
(18:17):
test those ede cases like these areexpensive aircraft, and you know, how
do you convince yourself and convince theconsumer that getting into this type of transportation
option is going to keep them safe. Look, we already have their transportation
that's insanely safe today. If youtook the excellent rates for commercial traffic in
(18:41):
the nineteen eighties and applied it tothe volumes we see in the sky,
is today we have a major jetcrash every three to four weeks. Wow,
I can't remember the last commercial crashin the United States of a large
aircraft. I know there was one. Of course, we had a terrorist
attack to one pilot. Fully thereafter. In fact, I'd say roughly
(19:06):
half of the large jet crashes todayseem to be more of a terrorist nature
that someone as standing very very stupidintentionally. It's very very safe, And
why is it so safe. It'ssafe for a number of reasons. One
is the AFA is doing a reallytremendously great job in mandating redundant sea and
certification and standards that go way beyondthe automotive centers today. And secondly,
(19:27):
the sky is actually very patient.It's very large. It's very hard to
run into stuff in the sky.Compare yourself on the ground. You might
be on the surface road you gowith sixty five miles an hour, you've
oncoming traffic and you pass each otherwith the direction differential speed of one hundred
and thirty miles an hour maybe withlike six to ten foot clearance. That
is scary. You would never dothis in the sky. In the sky
(19:49):
right now, aircraft have clearances thatare hundreds of thousands of feet away,
and we can maintain those using computercontrol very very very easily. So point
I think the sky will outsafe outsave the ground at someone. At some
point you're can go to a grandcheland say, look, it will be
unsafe. Take a car, itwill be safe. Take a plane.
(20:11):
What is the timeline? How soondo you think that this is going to
be commonplace versus something that almost soundslike science fiction to those that haven't been
following what you've been working on.Well, it's about I'd say these ten
very credible companies right now, somerecently when public through a stack building those
vehicles, working with the essay,and as in Europe on certification, and
(20:33):
I expect that we go to seethe first ones on the market maybe in
let's say three years from now,roughly two, three years, four years
from now, and those will notbe very economical. They'll still be very
expensive, hard to make. Thecost millions of dollars. But a lot
of thinking has gone recently in thequestion can we make them cheaper? Can
we make them as cheap as acar? A car, in a low
end car, like your cheapest caryou can find, is about five dollars
(20:55):
a pound. Okay, a raisedcar is probably fifty bucks a pounds,
and we will take the depending onwhere you are on the spectrum. A
plane today is five hundred bucks apound. We have to get to the
point where these airplanes cost five bucksa pound. And what would this mean?
I would mean that for ten thousandbucks for the price of the jet
ski, you could buy a sevenaircraft that can fly one hundred miles and
get you. We're going to goover in your city. You talked earlier
(21:18):
about how motorsports has produced innovation forthe auto industry. Certainly your experience with
the DARPA Grand Challenge is that competitionbreeds innovation. That's certainly a big part
of the mantra for the Indie AutonomousChallenge too. Is there some similar type
of competition that you can envision forautonomy in the air You're like you're talking
(21:41):
about that can drive this forward?I mean, the obvious one that comes
to mind is that've had bull races, which I like the Devil races.
Yeah, yeah, and that kindof They're kind of interesting because there are
any challenge technology and pilots alike.Obviously, doing an air base race has
this additional feeling if you make aweek grave mistake, you might actually go
up in flames, even more sothan you'd be in the ground. Based
(22:03):
like the INDE five hundreds and thosetend to push safety. They tend to
help us understand in extreme situations,but otherwise technologies and rules of the vote
to help us keep safe. Well, I want to build a air race
like the Vtport Race, probably onlywithout people inside. I wouldn't want to
(22:25):
put my children into a vehicle likethis and feel good about it. But
in due time, I'm pretty surethat we can build save systems that will
give every single person a chance tofly from A to B. Another passion
that you've had is education with udacityin your efforts to attract just want more
(22:48):
people and make it easier for themto engage, particularly in science and technology
and engineering. Obviously, one ofthe focuses of India Autonomy Challenge, and
we've been speaking with some of thestudents that are involved in the challenge is
motivating young people and motivating students todrop everything and focus on these technical challenges
(23:15):
in a way that they kind ofget the bug and they now want to
do this for the rest of theircareer and they can see the value in
what they're doing in a way thatmaybe you can't get in a lab environment
or just watching a professor online.Did you see that in DARPA with not
(23:37):
only yourself but the other I mean, you were part of a whole team
of students that you were leading andcompetitors, and how important are these kinds
of hands on, real world experienceswith the technology that you can see progressing,
versus having to trust that the labwork you're doing and the research what
(24:00):
you're doing will contribute to something inthe future. I recommend every student in
the world just go do it.It's so much fun. There's something very
concrete you want to accomplish. Don'tlet the wise old people tell you it
can't be done. There wrong.Everything can be done if you just put
your brains behind it. And tome as a researcher, as a long
(24:23):
time stime for professor, the sourceof all the theories that we came up
with on the academic side were alwaysgrounded in experiments just like these, where
we went and built something in thephysical world and really try it out and
learn from how to build it better, how to make it stronger, what
technology work better, and so on. So highly recommend it. It's something
where if you get engulfed, youshould find yourself not sleeping, you should
(24:47):
find yourself every waking moment thinking aboutit, including weekends. I know it's
not very social for most people,but it's been a ton of fun and
if you finally build something that drivesitself at one hundred and fifty miles an
hour, you can do it andsay I made that. I made that.
We uh. We did get thequestion from students when we told them
(25:08):
that we had this garage space atManza. They wanted to know what the
hours are, and of course theyalways want very early hours, very late
hours, but we had to makeit very clear they could not sleep in
the in the garage. They hadto leave for at least two or three
hours. But I think I'm prettysure they're going to sleep in the gar
Yeah. I think some of themare sleeping in their cars just outside the
(25:29):
park and then they come back inin the morning. I remember a very
first robot competition. I had arobot, a small robot office robot,
and it was standing in front ofme and had worked like three days without
sleeping, and it was starting thecompetition running around and then looked up of
us gone and entirely disappeared. Ihad like evacuated, and I was completely
confused, and then my teammates toldme, Sebastian, you just took a
(25:52):
nap on your keyboard. He wonthe competition and you just woke up,
right, I must have been liketwenty at the time. I'm super interested.
Actually if we go back to thoseearly days, because in my research,
your your interest in robotics, inautomation, this goes back a long
(26:14):
way. If I'm not mistaken,back in the late nineties, you were
working on a completely automated tour guide. If I'm not mistaken, is that
right? Right? And now you'redatingly that's kind of embarrassing. I want
be careful here. Yes, ninetiesdoesn't exist that. Yes, see,
we were live back then, myselfincluded well me too. My research went
from indoor to outdoors, and theearly years I built a lot of robots
(26:37):
that would be inside buildings. Oneof the early highlights was a robot report
of the Smithsonian Museum, right oneof the museums in Washington, d C.
Where we had a tour guide operationalas a robot that would greet kids
and people say hey, can Igive you a tour? And it became
a spectacle. It became an eventin Washington, d C. To see
the robotic tour guide. I stillremember at some point in the evening in
(27:00):
these museums, the tour guides arehuman dozents and're typically retired people who do
this as a hobby. They don'ttake any pay. And one of the
ladies walked up to the robot andit was really angry and said, so
you're not going to replace me?And no one else saw it, but
I saw it, like the blinkof my eye, that she was having
this conversation with this robot and overlike sitting in the corner and doing nothing.
(27:22):
It's looked like a guilty feeling robot. And there was a moment when
I realized, my god, you'reactually doing something. It could change people's
lives. If the only you hadchat GPT at the time, you could
have embedded it in the robot andit would have been, oh my god,
it would have gotten rid of allthe jobs really quickly. Have you
seen people's reactions to robotics change inthe last three decades or so or is
(27:45):
that same visceral reaction you can't replaceme? Does that still kind of ring
true? Is it almost universal?I think most of us are still waiting
for the retolution to unfold. We'veseen it now in AI, in robotics,
there's an amazing work that people don'tnormally see naturally, which is in
factories cunlayer bells and packaging and pallettingand welding and stuff like that. Automotive
(28:08):
that stuff is established, But inhome robotics were still waiting except for me,
for the lumba vacuum Kean, andwe're still waiting for this robot that
can just keep invocation at night.And that's around the corner, and there's
no set up companies that are seriouslybuilding affordable, safe and reliable machines they
can clean up our house. Andwhen that happens, it'd be interesting to
(28:29):
see how people react. I thinkby and largely be a good thing because
it helps us keeping our environments clean. It gives us more time back,
and it does the work we don'twant to do ourselves that's going to happen
next let's say ten years or so. Your love affair with robotics, has
it been driven by the use casesand the application of robotics to a particular
(28:51):
challenge, and then you apply yourunderstanding and your fascination with robotics versus just
the robot itself and trying to perfectthe hardware in the software. Yes,
so I've I always love the ideaof understanding human intelligence, and I always
felt from my earliest days on asan undergrad, robots are the perfect way
(29:15):
to study intelligence because they're embedded,they have ability to manipulate the world,
to perceive the physical world, andthey're the closest thing we can build to
ourselves, sort of like I know, not having kids or something. If
you can really understand every little bitof information that's throwing through the robot's brains.
From that, I've always been askingmyself that what's the right way to
(29:37):
build these robots to make the worlda better place for everybody? How can
I build a vote that will impressmy grandmother who doesn't understand technology and still
has human needs like normal beings andself diving Cars emerged quickly as the winning
answer because cars is something we alluse every day in this world. We
lose over a million people a yearin traffic accidents. It's a massive number
(30:02):
we often don't even pay attention to. That is kind It's like a million
plus funerals life's lost. It's likea big city a vipe dot every year.
So for me, there was kindof the opportunity to say, this
is something that we can do bringtogether this gentlemine curiosity on AI into something
that's practical and good for people.Last thing for me, we seem to
(30:25):
be on the precipice of a newage of AI, of robotics, of
automation, and so many avenues appearto be open for exploration. Are there
any though that you have some reticenceabout exploring or is it open season?
I think given the recent advances inthe eye, I'm very excited to ask
(30:48):
the question, how can we leverageAI to make people more effective and people
more efficient and give people more capabilitiesIn doing so? There are potential abuses,
the biggest being, of course,cyber kind of threats to good people
against good people from bad actors,bad countries like North Korea. There's something
(31:11):
of Russia that I worry about,and that's something we should all worry about.
And then we talked about earlier,how can we make sure when this
technology is being used that everybody canbenefit, not just a small number of
individuals, but all this. Ithink if you look at humanity, humanity
is three hundred thousand years old,right, we are three hundred thousand years
(31:32):
old. It is a long time, and almost everything that has advanced us.
The things we're reit proud of formmake yourself onto your car, to
your airplane, to what have youhas been invented in the last one hundred
and fifty years, even like electricityas a phenomenal, your light switch and
very mundane thinks your water toilet arenot older than two or three hundred years.
So that's something I think that's beenin fiction point for humanity to see
(31:56):
this technology advance us so rapidly,and I would say we're all better off
as a result. We all betteroff with a microwaveit moment, a roof
our top, and heat and allthese wonderful things that we all all enjoy
compared to where we were like three, four or five ndred years ago.
So for me, AI is justanother step, another progression in that that
don't give us the ability to bemore creative, more engaged, build more
(32:19):
things, try more things, andadvance society for that. So me,
it's a very very positivelopment, fascinatingconversation. So much more I'm sure we
could go into, but we haveto be conscious of your time. Sebastia,
thank you so much for joining usand sharing so many of your thoughts.
Thanks for joining us this week onthe inside track that was Sebastian through
(32:43):
and sharing his insight into the futureof AI and the new roads were likely
to see in the future. Nextweek, we'll be sitting down with David
Watkins, Senior Vice President of theIndiana Economic Development Corporation, to discuss how
state funded races can truly put Indianaon the technol lot map. And thanks
to the Indiana Economic Development Corporation forsponsoring this show.
The world around us is changing fasterthan ever before. Ideas once only imagined
in science fiction are becoming a reality. Throughout the course of our amazing twenty
three episode season, we'll speak tosome of the greatest minds in robotics and
artificial intelligence to discuss the groundbreaking workthat's fueling it all. I'm your host,
Ryan Marine joined me and my cohost Paul Mitchell, the president of
(00:26):
the Indian Autonomous Challenge and see whywe call this the inside Track. Today
we're speaking with a pioneer in autonomousvehicles, Sebastian Thrune. He's the CEO
of Kitty Hawk and the founder ofGoogle's self driving car team, as well
as an educator and co founder ofFeudacity. Sebastian guides is through the future
of personal flying machines, the ethicsof AI, and how the work of
(00:48):
the Indie Autonomous Challenge will make theroads safer for everyone. Sebastian, it
is a thrill to talk to you, to have you here on the show.
I know from talking to Paul overthe last few years, your vote
of confidence in the Indie Autonomous Challengehad a lot to do with getting it
off the ground. And for you, this AI in mobility, This has
(01:12):
been a big part of your lifefor the past several decades. So for
you to see how that has evolved, what pride do you take in playing
a pretty major part in getting itto the place where it is today.
Look, it takes many, manypeople to invent a new technology like this,
and I happen to be around whenDARPRA launches very first challenge two thousand
(01:34):
and five that we had to buildcost right the desert. I happened to
be around when Google decided to getinto the game and build Weemo, which
I have founded. And I'm superexcited about the INDI five found Out Challenge.
I think that's a great thing forhumanity and it's a wonderful thing to
do. Let's go back to theearly days and to DARPA, which in
many ways is the genesis of whatthe Indie Autonomous Challenge is about. What
(01:57):
drew you first of all to thatcompetition. Look, there was a long
time ago, two thousand and five, the US government had reserved I think
two million bucks as price money forsomeone who could build a car that could
drive itself one hundred and forty milesto a desert trail, and I got
drawn in not because I cared aboutdesert driving, but the care about human
life. Earlier in my life,when I was eighteen, I lost my
(02:19):
best friend to a traffic accident,and I felt reminded that by and large,
humans are not the safest drivers onEarth, and I was intrigued thinking
I could make a machine that coulddrive safer than people would drive. You
know, Sebastian, when you andI were talking in twenty seventeen, In
twenty eighteen about the concept of highspeed autonomous racing, I remember you talked
(02:45):
about kind of the first frontier ofautomation and self driving at lower speeds,
trying to navigate the urban suburban environment, and that the next frontier was,
you know, moving to high speeds, extreme speeds one hundred miles an hour,
you know, plus and testing whetheror not the hardware, the software
(03:07):
stacks can handle those environments. Soyou know, we've been doing that now
with competitions throughout the US, andwe're going to be running in Monza in
Europe here in June. But wheredo you see the contribution of high speed
automation playing the greatest role in savinglives in being part of the overall goal
(03:32):
that we all have of autonomous mobility. Yeah, look, that's correct.
I always believe that the faster thatyou can go is the hot It's going
to be a headled safely. Soif you want to be challenge technology,
going at the envelope of it's possible, physically possible certainly challenges technology the most.
The Dopogren challenge, I think wasan average but tendy miles an hour
(03:54):
was very slow. The Google ofAMO costs can dive highway speeds, but
they try to stay away from extremedriving. But when it comes to retesting
the technology, I think extreme drivingdriving at the envelope that's physically possible,
it's going to be the hardest test. And I love that you're doing this
and to be living in the extremelikely and the autonomous challenges. It's an
(04:15):
extended edge case. Basically, everytime the cars are on the track,
not only is it testing the technology, but also doesn't that say a lot
to the consumer public as well,to prove the efficacy of this technology in
an environment that they wouldn't expect tohave to be a part of themselves.
Yeah, it depends on where theworks, but I honestly believe eventually a
(04:40):
robotic car will surpass the ability ofa human driver, especially if we let
their robotic car control wheels and brakesindividually, which people can't do as easily.
And given that level of freedom,and I think the technology that's being
developed, we'll trickle down into safercars for normal drivers, because it's been
always extreme racing technology that let ustest out what is feasible and possible,
(05:03):
and we then took this technology andbrought it back to normal cars for normal
safety. One thing we've been workingon and thinking about some of the teams
and some of the sponsors and technologytechnologists that are involved in India Autonomous Challenge
is also how do you transition thetechnology to human race car drivers in parallel
(05:27):
with trying to advance the capabilities ofthe of the AI driver. And we
often think of it in the contextof you know, advanced fighter pilots,
right like if a human race cardriver has the ability to have three hundred
and sixty degree perception, you know, either through their their their visor or
or even their race control team cankind of provide indications to them if if
(05:49):
vehicles can you know, take overin the event that you know, unfortunately
the driver misses a red flag warningor or you know, is out of
control. That some of these technologiescould allow human pilots of race cars to
be closer to what we see.And you know, let's say the advanced
fighter pilots that are in F twentytwo. Yeah. Look, I would
(06:15):
say that even today we have technologyin cars that make us massively safe.
While that goes way beyond human capabilities, the most notable is abs. The
ability to see if a tire isslipping and then release it very quickly to
make sure it doesn't constantly slip.Is the technology that saved uncounted numbers of
lives. It's the racing field thatdevelops these technologies first. And I believe
(06:40):
that a computer system and principle doesn'tjust have a three six view that people
don't really have, but I alsocan look at details of the vehicle that
people can pay attention to, like, for example, is a tire slipping
right now, and really help makingthe driving system much more safe. And
that's nothing new. It's been aroundfor decades now and i'ven't benefit from it.
Everybody benefits from it. And Isee this to progress. I think
(07:01):
also get to the point where itmight someone be impossible to pressure a car,
it might be impossible to hit epestia, It might just be impossible to
do these things by having a smartcar that looks a lot of shorter and
prevents us from making these mistakes.One thing I've been wanting to ask you
about for a while is is justyour perspective on this whole national global debate
(07:23):
on AI and and for somebody who'sbeen using advanced technologies and AI and machine
learning in the in the applicant andapplying it to mobility, whether it's ground
based or aircraft. What is yourperspective on on AI as it relates to
(07:45):
its contribution to society and how wecan utilize it in ways that are productive
versus you know, do you haveany fear for the role that AI may
play with the chat, GPT andother kind of human communication based AAIYE tools
that are emerging. Great question.I mean, in the early days of
(08:05):
self diving cars, we already usedmachine learning to train these cars so you
could learn from us and from themselves. And now we're entering a world where
we have large scale called large languagemodels and and ms that are being trained
on all corpuses of text and thenshow a credible productivity and capability to form
(08:26):
original pros that translate languages, todraw images and creative art. I would
say we're at a big inflection pointfor humanity, and we had a point
now where these models, these systemscan do everything we ask a high school
student to do within a plus infact, they can do but anything we
ask a college graduate to do,like a psychology major or maybe a marketing
(08:50):
major. Some of them can evenpass the Bar exam ninety percentile. Here
in California, about forty percent ofcandidates fail the bar exams, So it's
obviously better than you average a lawschool graduate. And that doesn't mean that
they're going to replace all these people. Instead, they're going to make all
of us much more powerful. They'regoing to enable us to do something in
our that we've previously taken us amonth to do or a week to do.
(09:13):
That's nothing new, it's actually historicallywe did this already. We went
to a transformation society with the adminof the steam engine and modern agriculture over
the last one hundred and fifty hundredyears. Prior to that, most people
in the world worked in farming.A farmer could make food for something like
four people their family, and ourfarmer can make food for four hundred people.
(09:37):
And that has not been the endof civilization as we know. It's
been the opposite. I'd say we'reall better off. We now live in
a world that's safer, and welive longer, we live healthier, we
have more food, we have moreeverything because of those revolutions. I think
that's just the next step. Thenext step is going to be that everybody
working offices, from accountant to lawyerto what have you, will become one
(09:58):
hundred times as effective. And Iassume that you would see that extending to
mobility and human safety and vehicles andaircraft, etc. That some of these
learning models that are doing amazing thingsin the social sciences and literature and marketing
will emerge further in the field ofself driving cars and vertical takeoff aircraft and
(10:24):
other platforms. Absolutely, if yougoing to San Francisco, you're surrounded by
completely empty cars that drive themselves orcars that have passengers, and the very
first instantiations technology provides safety. It'ssafer than human drivers. I think they're
y more cars and I have todrive roughly thirty million miles without any major
(10:46):
accident caused by those cars. That'skind of amazing compared to human safety levels.
But then the second stage, you'regoing to see accessibility. You're going
to see people who wouldn't use carsto day, blind people, very very
old, people very very young,and be able to partic in transportation with
this new model, and you'll seethe even more cost effective I for,
for example, I'm a pilot.I'm mesually a jet pilot. I have
(11:09):
my own jet, and when Ifly these days, all I ever do
is after I take off, Ipush one button and then I'm being told
by ATC to the left of fiftydegrees, climb up one thousand feet,
and that means I'm going to diala number into a computer and that's all
I ever do. So the machinebasically flies itself, but all practical purposes.
It can even land itself in anemergency. That's something that I think
(11:31):
we're going to see all around seeeverybody now in the future being able to
use these technologies to be safer,faster. You can sleep while you've been
car don't drive it anymore. Youcan you can watch a television movie and
get transportations safer and cheaper than today. Back to Paul's comment about the conversation
surrounding AI right now, one ofthe big angles I suppose in that conversation
(11:54):
is the ethical question. And youbring up the point where the vent of
the steam engine or modern agriculture changesociety, And certainly, if you look
through a broad lens, it changedit in an objectively good way. But
that transition wasn't always an easy one, and there were inflection points along the
(12:15):
way in which the human condition insome instances did suffer from this process of
modernization. So what lessons can welearn from industrialization, for example, and
that process that took several hundred yearsto apply these new technologies in a way
that is more uniformly good for humanityat large. I think it's a great
(12:37):
question Ryan, that you're asking here, and it's a question we should all
be talking about, in part becausethis evolution is so much faster than the
past evolutions we've seen. What itused to take fifteen years will not take
five years, to be quite frank, And it's important that no one has
left behind. It's important that weas a society to use these technologies responsibly
for the betterment of everybody, notjust a few. It's a very broad
(12:58):
dialogue. That's a way right now. I don't think technology can be stopped.
I think what we can do isreally think about what are the implications
overall? These technologies will produce morewealth, more things, more goods,
more abandons. How can make surethat the abandonce that's being produced is fairly
distributed among all of us? Inyour opinion, who bears responsibility for ensuring
(13:22):
that that is the outcome? Isit up to individual agents, is it
up to the corporate world, ordoes there need to be some kind of
oversight from a government structure locally orinternationally. I think it's everybody's responsibility.
It's us, the people that areaffected, as the technologists, as the
consumers, as the voters who voterepresentatives for us into office. Is us
(13:46):
the office holders have been elected todo this, and we've got to beick
smart about it. It's really easyto say it's just outlawed. It just
happened in Italy, at least,said Chad GPT's outlawed. If you go
back in history, you find thatthe Turkish Empire in the Middle East illegalized
the printing press, and that didn'tturn out well for them because that precluded
(14:09):
the dissimnation of innovation and new information. So we have to have a broad
dialogue. But we should not putour head in the scent and say you
don't want it. Should instead saywhat's the most productive, best used that
is fair to everybody and helps usas human race to become better. I
want to go back to your yourwork in the arc of your of your
(14:33):
corporate and technology endeavors. You playedthis critical role in the early and then
the autonomous vehicle and working with theteam at Google to get way more going,
and then you've transitioned into the flyingcar or the vertical takeoff aircraft space.
(14:54):
And I'm just curious as we thinkabout and I don't know if you
can see behind us, but youknow, we've got these one hundred and
eighty one hundred and ninety mile perhour, and I don't know if you've
been tracking the progress, but wenow are passing each other at one hundred
and eighty miles an hour on oldlisteness and transitioning to road courses kind of
(15:15):
you know, how do you seethe challenges of transitioning from the passenger car
environment to aircraft that are traveling athigher speeds. What drew you to that?
It sounds like you're a pilot,so maybe it's just your own personal
passion. And then what is yourthought about how soon are we going to
(15:37):
see that technology? In some casesaugment probably not replaced, but certainly augment
are focused on ground based mobility thatis still we're so dependent on for let's
say short or mid distance trips.I mean, historically the growing made sense
for transportation because gravit it just pulledus all down and it's really hard to
(16:00):
stay up in the air. ButI'd say for quite a while now it's
been clear that we can get thesame efficiency fuel efficiency, cost efficiency in
the air that you can get onthe ground. And what do you gain
by going to the air? Numberone, pretty much all obstacles are gone,
Like all this stuff that you avoidabout the ground, like trees and
houses and curbs and whatever, they'reall gone. Maybe with the exception of
(16:25):
a few birds. There's no catsand dogs in the sky. Then you
get to go in a much muchbigger volume. So instead of having been
paved roads, the roads in thesky they just don't require any pavement,
there's no infrastructure necessary except for theplace where they take off a land and
with a VTOL vertical take off alanding helicopter like you can land and take
(16:47):
off pretty much anywhere. If there'sno alidlines, you just go and do
this in your front yard and yourdriveway, and then once you're in the
air, you go on a straightline. You go much faster, much
more efficiently. We've shown a kiddingtalk that the fuel consumption for a one
passenger EVE to the vehicle is roughlya third bit of a Tesla electric car
(17:07):
per mile, and then even attwice as speed, which probably quadruples the
drag of the vehicles. So it'sit's actually more fuel efficient to be in
the air and on the ground.So from a safety perspective, from an
efficiency perspective, from a cost perspective, from a green perspective, air based
transportation bill completely throvolved GUND transportation goingforward. Can you sour? One of
(17:33):
our thesis in this is that everyI agree with everything you just said,
that the importance of the edge casein air traffic is that much more important
because of the velocity in which anaccident happens. The way in which an
accident happens. Now, I thinka lot of that can be managed with
(17:56):
things like the carbon fiber monocoques thatare in raised cars. And we know
so much about how to protect thepassenger the critical cargo. But one of
the luxuries you have with passenger carsgoing twenty thirty miles an hour is they
can crash and you just sort ofput them back together and you know,
hopefully nobody gets hurt in it.So in your mind, how do you
(18:17):
test those ede cases like these areexpensive aircraft, and you know, how
do you convince yourself and convince theconsumer that getting into this type of transportation
option is going to keep them safe. Look, we already have their transportation
that's insanely safe today. If youtook the excellent rates for commercial traffic in
(18:41):
the nineteen eighties and applied it tothe volumes we see in the sky,
is today we have a major jetcrash every three to four weeks. Wow,
I can't remember the last commercial crashin the United States of a large
aircraft. I know there was one. Of course, we had a terrorist
attack to one pilot. Fully thereafter. In fact, I'd say roughly
(19:06):
half of the large jet crashes todayseem to be more of a terrorist nature
that someone as standing very very stupidintentionally. It's very very safe, And
why is it so safe. It'ssafe for a number of reasons. One
is the AFA is doing a reallytremendously great job in mandating redundant sea and
certification and standards that go way beyondthe automotive centers today. And secondly,
(19:27):
the sky is actually very patient.It's very large. It's very hard to
run into stuff in the sky.Compare yourself on the ground. You might
be on the surface road you gowith sixty five miles an hour, you've
oncoming traffic and you pass each otherwith the direction differential speed of one hundred
and thirty miles an hour maybe withlike six to ten foot clearance. That
is scary. You would never dothis in the sky. In the sky
(19:49):
right now, aircraft have clearances thatare hundreds of thousands of feet away,
and we can maintain those using computercontrol very very very easily. So point
I think the sky will outsafe outsave the ground at someone. At some
point you're can go to a grandcheland say, look, it will be
unsafe. Take a car, itwill be safe. Take a plane.
(20:11):
What is the timeline? How soondo you think that this is going to
be commonplace versus something that almost soundslike science fiction to those that haven't been
following what you've been working on.Well, it's about I'd say these ten
very credible companies right now, somerecently when public through a stack building those
vehicles, working with the essay,and as in Europe on certification, and
(20:33):
I expect that we go to seethe first ones on the market maybe in
let's say three years from now,roughly two, three years, four years
from now, and those will notbe very economical. They'll still be very
expensive, hard to make. Thecost millions of dollars. But a lot
of thinking has gone recently in thequestion can we make them cheaper? Can
we make them as cheap as acar? A car, in a low
end car, like your cheapest caryou can find, is about five dollars
(20:55):
a pound. Okay, a raisedcar is probably fifty bucks a pounds,
and we will take the depending onwhere you are on the spectrum. A
plane today is five hundred bucks apound. We have to get to the
point where these airplanes cost five bucksa pound. And what would this mean?
I would mean that for ten thousandbucks for the price of the jet
ski, you could buy a sevenaircraft that can fly one hundred miles and
get you. We're going to goover in your city. You talked earlier
(21:18):
about how motorsports has produced innovation forthe auto industry. Certainly your experience with
the DARPA Grand Challenge is that competitionbreeds innovation. That's certainly a big part
of the mantra for the Indie AutonomousChallenge too. Is there some similar type
of competition that you can envision forautonomy in the air You're like you're talking
(21:41):
about that can drive this forward?I mean, the obvious one that comes
to mind is that've had bull races, which I like the Devil races.
Yeah, yeah, and that kindof They're kind of interesting because there are
any challenge technology and pilots alike.Obviously, doing an air base race has
this additional feeling if you make aweek grave mistake, you might actually go
up in flames, even more sothan you'd be in the ground. Based
(22:03):
like the INDE five hundreds and thosetend to push safety. They tend to
help us understand in extreme situations,but otherwise technologies and rules of the vote
to help us keep safe. Well, I want to build a air race
like the Vtport Race, probably onlywithout people inside. I wouldn't want to
(22:25):
put my children into a vehicle likethis and feel good about it. But
in due time, I'm pretty surethat we can build save systems that will
give every single person a chance tofly from A to B. Another passion
that you've had is education with udacityin your efforts to attract just want more
(22:48):
people and make it easier for themto engage, particularly in science and technology
and engineering. Obviously, one ofthe focuses of India Autonomy Challenge, and
we've been speaking with some of thestudents that are involved in the challenge is
motivating young people and motivating students todrop everything and focus on these technical challenges
(23:15):
in a way that they kind ofget the bug and they now want to
do this for the rest of theircareer and they can see the value in
what they're doing in a way thatmaybe you can't get in a lab environment
or just watching a professor online.Did you see that in DARPA with not
(23:37):
only yourself but the other I mean, you were part of a whole team
of students that you were leading andcompetitors, and how important are these kinds
of hands on, real world experienceswith the technology that you can see progressing,
versus having to trust that the labwork you're doing and the research what
(24:00):
you're doing will contribute to something inthe future. I recommend every student in
the world just go do it.It's so much fun. There's something very
concrete you want to accomplish. Don'tlet the wise old people tell you it
can't be done. There wrong.Everything can be done if you just put
your brains behind it. And tome as a researcher, as a long
(24:23):
time stime for professor, the sourceof all the theories that we came up
with on the academic side were alwaysgrounded in experiments just like these, where
we went and built something in thephysical world and really try it out and
learn from how to build it better, how to make it stronger, what
technology work better, and so on. So highly recommend it. It's something
where if you get engulfed, youshould find yourself not sleeping, you should
(24:47):
find yourself every waking moment thinking aboutit, including weekends. I know it's
not very social for most people,but it's been a ton of fun and
if you finally build something that drivesitself at one hundred and fifty miles an
hour, you can do it andsay I made that. I made that.
We uh. We did get thequestion from students when we told them
(25:08):
that we had this garage space atManza. They wanted to know what the
hours are, and of course theyalways want very early hours, very late
hours, but we had to makeit very clear they could not sleep in
the in the garage. They hadto leave for at least two or three
hours. But I think I'm prettysure they're going to sleep in the gar
Yeah. I think some of themare sleeping in their cars just outside the
(25:29):
park and then they come back inin the morning. I remember a very
first robot competition. I had arobot, a small robot office robot,
and it was standing in front ofme and had worked like three days without
sleeping, and it was starting thecompetition running around and then looked up of
us gone and entirely disappeared. Ihad like evacuated, and I was completely
confused, and then my teammates toldme, Sebastian, you just took a
(25:52):
nap on your keyboard. He wonthe competition and you just woke up,
right, I must have been liketwenty at the time. I'm super interested.
Actually if we go back to thoseearly days, because in my research,
your your interest in robotics, inautomation, this goes back a long
(26:14):
way. If I'm not mistaken,back in the late nineties, you were
working on a completely automated tour guide. If I'm not mistaken, is that
right? Right? And now you'redatingly that's kind of embarrassing. I want
be careful here. Yes, ninetiesdoesn't exist that. Yes, see,
we were live back then, myselfincluded well me too. My research went
from indoor to outdoors, and theearly years I built a lot of robots
(26:37):
that would be inside buildings. Oneof the early highlights was a robot report
of the Smithsonian Museum, right oneof the museums in Washington, d C.
Where we had a tour guide operationalas a robot that would greet kids
and people say hey, can Igive you a tour? And it became
a spectacle. It became an eventin Washington, d C. To see
the robotic tour guide. I stillremember at some point in the evening in
(27:00):
these museums, the tour guides arehuman dozents and're typically retired people who do
this as a hobby. They don'ttake any pay. And one of the
ladies walked up to the robot andit was really angry and said, so
you're not going to replace me?And no one else saw it, but
I saw it, like the blinkof my eye, that she was having
this conversation with this robot and overlike sitting in the corner and doing nothing.
(27:22):
It's looked like a guilty feeling robot. And there was a moment when
I realized, my god, you'reactually doing something. It could change people's
lives. If the only you hadchat GPT at the time, you could
have embedded it in the robot andit would have been, oh my god,
it would have gotten rid of allthe jobs really quickly. Have you
seen people's reactions to robotics change inthe last three decades or so or is
(27:45):
that same visceral reaction you can't replaceme? Does that still kind of ring
true? Is it almost universal?I think most of us are still waiting
for the retolution to unfold. We'veseen it now in AI, in robotics,
there's an amazing work that people don'tnormally see naturally, which is in
factories cunlayer bells and packaging and pallettingand welding and stuff like that. Automotive
(28:08):
that stuff is established, But inhome robotics were still waiting except for me,
for the lumba vacuum Kean, andwe're still waiting for this robot that
can just keep invocation at night.And that's around the corner, and there's
no set up companies that are seriouslybuilding affordable, safe and reliable machines they
can clean up our house. Andwhen that happens, it'd be interesting to
(28:29):
see how people react. I thinkby and largely be a good thing because
it helps us keeping our environments clean. It gives us more time back,
and it does the work we don'twant to do ourselves that's going to happen
next let's say ten years or so. Your love affair with robotics, has
it been driven by the use casesand the application of robotics to a particular
(28:51):
challenge, and then you apply yourunderstanding and your fascination with robotics versus just
the robot itself and trying to perfectthe hardware in the software. Yes,
so I've I always love the ideaof understanding human intelligence, and I always
felt from my earliest days on asan undergrad, robots are the perfect way
(29:15):
to study intelligence because they're embedded,they have ability to manipulate the world,
to perceive the physical world, andthey're the closest thing we can build to
ourselves, sort of like I know, not having kids or something. If
you can really understand every little bitof information that's throwing through the robot's brains.
From that, I've always been askingmyself that what's the right way to
(29:37):
build these robots to make the worlda better place for everybody? How can
I build a vote that will impressmy grandmother who doesn't understand technology and still
has human needs like normal beings andself diving Cars emerged quickly as the winning
answer because cars is something we alluse every day in this world. We
lose over a million people a yearin traffic accidents. It's a massive number
(30:02):
we often don't even pay attention to. That is kind It's like a million
plus funerals life's lost. It's likea big city a vipe dot every year.
So for me, there was kindof the opportunity to say, this
is something that we can do bringtogether this gentlemine curiosity on AI into something
that's practical and good for people.Last thing for me, we seem to
(30:25):
be on the precipice of a newage of AI, of robotics, of
automation, and so many avenues appearto be open for exploration. Are there
any though that you have some reticenceabout exploring or is it open season?
I think given the recent advances inthe eye, I'm very excited to ask
(30:48):
the question, how can we leverageAI to make people more effective and people
more efficient and give people more capabilitiesIn doing so? There are potential abuses,
the biggest being, of course,cyber kind of threats to good people
against good people from bad actors,bad countries like North Korea. There's something
(31:11):
of Russia that I worry about,and that's something we should all worry about.
And then we talked about earlier,how can we make sure when this
technology is being used that everybody canbenefit, not just a small number of
individuals, but all this. Ithink if you look at humanity, humanity
is three hundred thousand years old,right, we are three hundred thousand years
(31:32):
old. It is a long time, and almost everything that has advanced us.
The things we're reit proud of formmake yourself onto your car, to
your airplane, to what have youhas been invented in the last one hundred
and fifty years, even like electricityas a phenomenal, your light switch and
very mundane thinks your water toilet arenot older than two or three hundred years.
So that's something I think that's beenin fiction point for humanity to see
(31:56):
this technology advance us so rapidly,and I would say we're all better off
as a result. We all betteroff with a microwaveit moment, a roof
our top, and heat and allthese wonderful things that we all all enjoy
compared to where we were like three, four or five ndred years ago.
So for me, AI is justanother step, another progression in that that
don't give us the ability to bemore creative, more engaged, build more
(32:19):
things, try more things, andadvance society for that. So me,
it's a very very positivelopment, fascinatingconversation. So much more I'm sure we
could go into, but we haveto be conscious of your time. Sebastia,
thank you so much for joining usand sharing so many of your thoughts.
Thanks for joining us this week onthe inside track that was Sebastian through
(32:43):
and sharing his insight into the futureof AI and the new roads were likely
to see in the future. Nextweek, we'll be sitting down with David
Watkins, Senior Vice President of theIndiana Economic Development Corporation, to discuss how
state funded races can truly put Indianaon the technol lot map. And thanks
to the Indiana Economic Development Corporation forsponsoring this show.
Popular Podcasts
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.
My Favorite Murder with Karen Kilgariff and Georgia Hardstark
My Favorite Murder is a true crime comedy podcast hosted by Karen Kilgariff and Georgia Hardstark. Each week, Karen and Georgia share compelling true crimes and hometown stories from friends and listeners. Since MFM launched in January of 2016, Karen and Georgia have shared their lifelong interest in true crime and have covered stories of infamous serial killers like the Night Stalker, mysterious cold cases, captivating cults, incredible survivor stories and important events from history like the Tulsa race massacre of 1921. My Favorite Murder is part of the Exactly Right podcast network that provides a platform for bold, creative voices to bring to life provocative, entertaining and relatable stories for audiences everywhere. The Exactly Right roster of podcasts covers a variety of topics including historic true crime, comedic interviews and news, science, pop culture and more. Podcasts on the network include Buried Bones with Kate Winkler Dawson and Paul Holes, That's Messed Up: An SVU Podcast, This Podcast Will Kill You, Bananas and more.
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.