back to indexSteve Viscelli: Trucking and the Decline of the American Dream | Lex Fridman Podcast #237
link |
The following is a conversation with Steve Vasile,
link |
formerly a truck driver and now a sociologist
link |
at the University of Pennsylvania
link |
who studies freight transportation.
link |
His first book, The Big Rig,
link |
Trucking and the Decline of the American Dream,
link |
explains how long haul trucking
link |
went from being one of the best blue collar jobs
link |
to one of the toughest.
link |
His current ongoing book project,
link |
Driverless, Autonomous Trucks
link |
and the Future of the American Trucker,
link |
explores self driving trucks
link |
and their potential impacts on labor and on society.
link |
This is the Lex Friedman podcast.
link |
To support it, please check out our sponsors
link |
in the description.
link |
And now, here's my conversation with Steve Vasile.
link |
You wrote a book about trucking
link |
called The Big Rig, Trucking and the Decline
link |
of the American Dream,
link |
and you're currently working on a book
link |
about autonomous trucking called Driverless,
link |
Autonomous Trucks and the Future of the American Trucker.
link |
I have to bring up some Johnny Cash to you
link |
because I was just listening to this song.
link |
He has a ton of songs about trucking,
link |
but one of them I was just listening to,
link |
it's called All I Do is Drive,
link |
where he's talking to an old truck driver.
link |
It goes, I asked them if those trucking songs
link |
tell about a life like his.
link |
He said, if you want to know the truth about it,
link |
here's the way it is.
link |
All I do is drive, drive, drive,
link |
try to stay alive, that's the chorus,
link |
and keep my mind on my load,
link |
keep my eye upon the road.
link |
I got nothing in common with any man
link |
who's home every day at five.
link |
All I do is drive, drive, drive,
link |
drive, drive, drive, drive.
link |
So I got to ask you,
link |
same thing that he asked the trucker.
link |
You worked as a trucker for six months
link |
while working on the previous book.
link |
What's it like to be a truck driver?
link |
I think that captures it.
link |
Can you take me through the whole experience,
link |
what it takes to become a trucker,
link |
what actual day to day life was on day one,
link |
week one, and then over time how that changed?
link |
Well, the book is really about how that changed over time.
link |
So my experience, and I'm an ethnographer, right?
link |
So I go in, I live with people,
link |
I work with people, I talk to them,
link |
try to understand their world.
link |
Ethnographer, by the way, what is that?
link |
The science and art of capturing the spirit of a people?
link |
I think that would be a good way to capture it,
link |
try to understand what makes them unique as a society,
link |
maybe as a subculture, what kind of makes them tick,
link |
that might be different than the way you and I are wired.
link |
And really sort of thickly describe it
link |
would be at least one component of it.
link |
That's sort of the basic essential.
link |
And then for me, I want to exercise what C. Wright Mills
link |
called the sociological imagination,
link |
which is to put that individual biography
link |
into the long historical sweep of humanity,
link |
if at all possible.
link |
My goals are typically more modest than C. Wright Mills's.
link |
And to then put that biography in the larger social structure
link |
to try to understand that person's life
link |
and the way they see the world,
link |
their decisions in light of their interests
link |
relative to others and conflict and power
link |
and all these things that I find interesting.
link |
And conflict and power.
link |
In the context of society and in the context of history.
link |
And the small tangent, what does it take to do that,
link |
to capture this particular group, the spirit, the music,
link |
the full landscape of experiences
link |
that a particular group goes through
link |
in the context of everything else?
link |
You only have a limited amount of time
link |
and you come to the table probably with preconceived notions
link |
that are then quickly destroyed, all that whole process.
link |
So I don't know if it's more art or science,
link |
but what does it take to be great at this?
link |
I do think my first book was a success
link |
relative to my goals of trying to really get at the heart
link |
of sort of the central issues
link |
and the lives being led by people.
link |
If I have a resource, a talent,
link |
it's that I'm a good listener.
link |
I can talk with anybody.
link |
My wife loves to remark on this
link |
that I can sort of sit down with anyone.
link |
I think I learned that from my dad who worked at a factory
link |
and actually had a lot of truckers go through
link |
the gate that he operated.
link |
And he always had a story, a joke for everybody,
link |
kind of got to know everyone individually.
link |
And he just taught me that essentially
link |
everyone has something to teach you.
link |
And I try to embody that.
link |
Like that's the rule is for me is every single person
link |
I interact with can teach me something.
link |
I gotta ask you, I'm sorry to interrupt
link |
because I'm clearly of the two of us, the poorer listener.
link |
I think you're a great listener.
link |
I've been listening to the podcast.
link |
I think you're a great listener.
link |
I really appreciate that.
link |
You've done a large number of interviews,
link |
like you said, of truckers for this book.
link |
I'm just curious, what are some lessons you've learned
link |
about what it takes to listen to a person enough,
link |
maybe guide the conversation enough
link |
to get to the core of the person, the idea,
link |
again, the ethnographer goal to get to the core?
link |
Yeah, I think it doesn't happen in the moment, right?
link |
So I'm a ruminator.
link |
I just sit with the data for years.
link |
I sat with the trucking data for almost 10 full years
link |
and just thought about the problems and the questions
link |
using everything that I possibly could.
link |
And so in the moment, my ideal interview is I open up
link |
and I say, tell me about your life as a trucker.
link |
And they never shut up.
link |
And they keep telling me the things that I'm interested in.
link |
Now, it never works out that way
link |
because they don't know what you're interested in, right?
link |
And so a lot of it is the, as you know,
link |
I think you're a great interviewer, prep, right?
link |
So you try to get to know a little bit about the person
link |
and sort of understand kind of the central questions
link |
you're interested in that they can help you explore.
link |
And so I've done hundreds of interviews with truck drivers
link |
at this point and I should really go back
link |
and read the original ones.
link |
They were probably terrible.
link |
What's the process like?
link |
You're sitting down, do you have an audio recorder
link |
and also taking notes or do you do no audio, just notes or?
link |
Yeah, audio recorder and social scientists
link |
always have to struggle with sampling, right?
link |
Like who do you interview?
link |
Where do you find them?
link |
How do you recruit them?
link |
I just happened to have a sort of natural place to go
link |
that gave me essentially the population
link |
that I was interested in.
link |
So all these long haul truck drivers that I was interested in
link |
they have to stop and get fuel
link |
and get services at truck stops.
link |
So I picked a truck stop at the juncture
link |
of a couple of major interstates,
link |
went into the lounge that drivers have to walk through
link |
with my clipboard and everybody who came through,
link |
I said, hey, are you on break?
link |
And that was sort of the first criteria was,
link |
do you have time, right?
link |
And if they said, yes, I said,
link |
I'd say, I'm a graduate student at Indiana University.
link |
I'm doing a study,
link |
trying to understand more about truck drivers.
link |
Will you sit down with me?
link |
And I think the first,
link |
I think I probably asked like 104 or 103 people
link |
to get the first 100 interviews.
link |
That's pretty good odds.
link |
It's amazing, right?
link |
For any response rate like that for interviewing,
link |
these are people who sat down and gave me an hour
link |
or sometimes more of their time,
link |
just randomly at a truck stop.
link |
And it just tells you something about like,
link |
truckers have something to say.
link |
They're alone a lot.
link |
And so I had to figure out how to kind of
link |
turn the spigot on, you know?
link |
And I got pretty good at it, I think, yeah.
link |
So they have good stories to tell
link |
and they have an active life in the mind
link |
because they spend so much time on the road
link |
just basically thinking.
link |
Yeah, there's a lot of reflection,
link |
a lot of struggles, you know?
link |
And they take different forms.
link |
One of the things that they talk about
link |
is the impact on their families.
link |
They say truckers have the same rate of divorce
link |
as everybody else.
link |
And that's because trucking saves so many marriages
link |
because you're not around and ruins so many.
link |
And so it ends up being a wash.
link |
So I had this experience.
link |
I met another person and he recognized me from a podcast.
link |
And he said, you know, I'm a fan of yours
link |
and a fan of Joe Rogan, but you guys never talk.
link |
You always talk to people with Nobel Prizes.
link |
You always talk to these kinds of people.
link |
You never talk to us regular folk.
link |
And that guy really stuck with me.
link |
First of all, the idea of regular folk is a silly notion.
link |
I think people that win Nobel Prizes
link |
are often more boring than the people, these regular folks
link |
in terms of stories, in terms of richness of experience,
link |
in terms of the ups and downs of life.
link |
And, you know, that really stuck with me
link |
because I set that as a goal for myself
link |
to make sure I talked to regular folk.
link |
And you did just this talking, again, regular folk.
link |
It's human beings.
link |
All of them have experiences.
link |
If you were to recommend to talk some of these folks
link |
with stories, how would you find them?
link |
Yeah, so I do do this sometimes for journalists
link |
who will come and they want to write about
link |
sort of what's happening right now in trucking, you know.
link |
And I send them to truck stops.
link |
I say, you know, yeah, there's a town
link |
called Effingham, Illinois.
link |
And it's just this place where, you know,
link |
bunch of huge truck stops, tons of trucks
link |
and really nothing else out there.
link |
You know, it's in the middle of corn country.
link |
And, you know, again, truckers in this, you know,
link |
sadly, I think, you know, the politics of the day,
link |
it's changing a little bit.
link |
I think there's a little, the polarization is getting
link |
to the trucking industry in ways that, you know,
link |
maybe we're seeing in other parts of our social world.
link |
But truckers are generally, you know,
link |
real open sort of friendly folks.
link |
Some of them ultimately like to work alone and be alone.
link |
That's a relatively small subset, I think.
link |
But all of them are generally, you know,
link |
kind of open, you know, trusting,
link |
willing to have a conversation.
link |
And so, you know, you go to the truck stop
link |
and you go in the lounge and they're usually,
link |
there's usually a booth down there
link |
and somebody is sitting at their laptop
link |
or on their phone and willing to strike up a conversation.
link |
You should try that.
link |
You should, you know.
link |
That 100% will try this.
link |
Just again, we're just going from tangent to tangent.
link |
We'll return to the main question,
link |
but what do they listen to?
link |
Do they listen to talk radio?
link |
Do they listen to podcasts, audio books?
link |
Do they listen to music?
link |
Do they listen to silence?
link |
Some, I mean, and some still listen to the CB,
link |
which, you know, it's a ever dwindling group.
link |
They'll call it the Original Internet Citizens Band.
link |
You know, back in the 70s,
link |
they thought it was going to be the medium of democracy.
link |
And they love to just get on there
link |
and, you know, cruise along one truck after the other
link |
Usually, you know, it's guys who know each other
link |
from the same company or happen to run into each other.
link |
But other than that, it's everything under the sun.
link |
You know, and that's, it's probably one of the stereotypes
link |
and it's, I think it was more true in the past, you know,
link |
about the sort of heterogeneity of truck drivers.
link |
They're a really diverse group now.
link |
You know, there's definitely a large,
link |
still a large component of rural white guys
link |
who work in the industry,
link |
but there's a huge growing chunk of the industry
link |
that's immigrants, people of color, and even some women.
link |
Still huge barriers to women entering it,
link |
but it's a much more diverse place than most people think.
link |
So let's return to your journey as a truck driver.
link |
What did it take to become a truck driver?
link |
What were the early days like?
link |
Yeah, so this is, I mean, this is a central part
link |
of the story, right, that I uncovered.
link |
And the good part was that I went in
link |
without knowing what was gonna happen.
link |
So I was able to experience it as a new truck driver would.
link |
It's one of the important stories in the book
link |
is how that experience is constructed by employers
link |
to sort of, you know, help you think the way
link |
that they would like you to think about the job
link |
and about the industry and about the social relations of it.
link |
So it's super intimidating.
link |
I say in the book, you know, pretty handy guy,
link |
you know, familiar with tools, machines,
link |
like, you know, comfortable operating stuff,
link |
like from time I was a kid.
link |
The truck was just like a whole nother experience.
link |
I mean, as I think most people think about it,
link |
it's this big, huge vehicle, right?
link |
It's really long, it's 70 feet long,
link |
it can weigh 80,000 pounds.
link |
You know, it does not stop like a car.
link |
It does not turn like a car.
link |
But at least when I started, and this has changed
link |
and it's part of the technology story of trucking,
link |
the first thing you had to do was learn how to shift it.
link |
And it doesn't shift like a manual car.
link |
The clutch isn't synchronized.
link |
So you have to do what's called double clutch.
link |
And it's basically the foundational skill
link |
that a truck driver used to have to learn.
link |
So you would, you know, accelerate, say you're in first gear,
link |
you push in the clutch,
link |
you pull the shifter out of first gear,
link |
you let the clutch out,
link |
and then you let the RPMs of the engine drop an exact amount.
link |
Then you push the clutch back in
link |
and you put it in second gear.
link |
If your timing is off,
link |
those gears aren't gonna go together.
link |
So if you're in an intersection,
link |
you're just gonna get this horrible grinding sound
link |
as you coast, you know, to a dead stop in the,
link |
you know, underneath the stoplight or whatever it is.
link |
So the first thing you have to do is learn to shift it.
link |
And so at least for me and a lot of drivers
link |
who are going to private company CDL schools,
link |
what happens is it's kind of like a bootcamp.
link |
They ship me three states away from home,
link |
send you a bus ticket and say,
link |
''Hey, we'll put you up for two weeks.''
link |
You sit in a classroom,
link |
you sort of learn the theory of shifting,
link |
the theory of kind of how you fill out your log book,
link |
rules of the road, you know, you do that maybe half the day.
link |
And then the other half you're in this giant parking lot
link |
with one of these old trucks and just like, you know,
link |
destroying what's left of the thing, you know,
link |
and it's lurching and belching smoke
link |
and just making horrible noises and like rattling.
link |
I mean, in these things, like there's a lot of torque.
link |
And so if you do manage to get it into gear,
link |
but the engine's lugging,
link |
I mean, it can throw you right out of the seat, right?
link |
So it's this, it's like, you know, it's bull,
link |
you're trying to ride and it's super intimidating.
link |
And the thing about it is that for everybody there,
link |
it's almost everybody there, it's super high stakes.
link |
So trucking has become a job of last resort
link |
for a lot of people.
link |
And so they, you know, they lose a job in manufacturing,
link |
they get too old to do construction any longer, right?
link |
The knees can no longer handle it.
link |
And they get replaced by a machine,
link |
their job gets, you know, offshored
link |
and they end up going to trucking because it's a place
link |
where they can maintain their income.
link |
And so it's super high stress.
link |
Like they've left their family behind,
link |
maybe they quit another job.
link |
They're typically being charged a lot of money.
link |
So that first couple of weeks,
link |
like you might get charged $8,000 by the company
link |
that you have to pay back if you don't get hired.
link |
And so the stakes are high and this machine is huge
link |
and it's intimidating.
link |
And so it's super stressful.
link |
I mean, I watched, you know, grown men break down crying
link |
about like how they couldn't go home and tell their son
link |
that they had been telling they were gonna, you know,
link |
go become a long haul truck driver that they'd failed.
link |
And it's kind of this super high stress system.
link |
And it's designed that way partly
link |
because as one of my trainers later told me,
link |
it's basically a two week job interview.
link |
Like they're testing you, they're seeing like, you know,
link |
how's this person gonna respond when it's tough, you know,
link |
when they have to do the right thing and it's slow
link |
and, you know, they need to learn something,
link |
are they gonna rush, you know,
link |
or are they gonna kind of stay calm, figure it out,
link |
you know, nose to the grindstone.
link |
Cause when you're a new truck driver, you're unsupervised,
link |
you know, and that's what they're really looking for
link |
is that kind of quality of conscientious work
link |
that's gonna carry through to the job.
link |
Well, so the truck is such an imposing part
link |
of a traffic scenario.
link |
So you said like turning, it stresses me out every time
link |
I look at a truck cause they, I mean,
link |
the geometry of the problem is so tricky.
link |
And so if you combine the fact that they have to,
link |
like everybody, basically all the cars in the scene
link |
are staring at the truck and they're waiting,
link |
often in frustration.
link |
And in that mode, you have to then shift gears perfectly
link |
and move perfectly.
link |
And if, when you're new, especially,
link |
like you'll probably, for somebody like me,
link |
it feels like it would take years to become calm
link |
and comfortable in that situation
link |
as opposed to be exceptionally stressed under the eyes
link |
of the road, everybody looking at you, waiting for you.
link |
Is that the psychological pressure of that?
link |
Is that something that was really difficult?
link |
Again, just, I saw people freeze up, you know,
link |
in that intersection as, you know, horns are blaring
link |
and the truck's grinding, you know, gears
link |
and you just can't, you know, and they just shut down.
link |
They're like, this isn't for me, I can't do it.
link |
You're right, it takes years.
link |
If, you know, trucking is not considered
link |
a skilled occupation, but, you know, my six months there,
link |
and I was a pretty good rookie, but when I finished,
link |
I was still a rookie, even shifting, definitely backing,
link |
tight corners and situations, you know,
link |
I could drive competently, but the difference between me
link |
and someone who had, you know, two, three years
link |
of experience was, it was a giant gulf between us.
link |
And between that and the really skilled drivers
link |
who've been doing it for 20 years, you know,
link |
it's still another step beyond that.
link |
So it is highly skilled.
link |
Would it be fair to break trucking into the task
link |
of driving a truck into two categories?
link |
One is like the local stuff, getting out of the parking lot,
link |
getting into, you know, driving down local streets
link |
and then highway driving, those two tasks.
link |
What are the challenges associated with each task?
link |
You kind of emphasized the first one.
link |
What about the actual like long haul highway driving?
link |
Yeah, so, I mean, and they are very different, right?
link |
And the key with the long haul driving is really a set of,
link |
the way I came to understand it was a set of habits, right?
link |
We have a sense of driving, particularly men, I think,
link |
have a sense of driving as like being really skilled,
link |
is like the goal and you can kind of maneuver yourself
link |
out of in and out of tight spaces with great speed
link |
and breaking and acceleration, you know.
link |
For a really good truck driver,
link |
it's about understanding traffic and traffic patterns
link |
and making good decisions
link |
so you never have to use those skills.
link |
And the really good drivers, you know,
link |
the mantra is always leave yourself an out, right?
link |
So always have that safe place that you can put that truck
link |
in case that four wheeler in front of you
link |
who's texting loses control.
link |
You know, what are you gonna do in that situation?
link |
And what really good truck drivers do on the highway
link |
is they just keep themselves
link |
out of those situations entirely.
link |
They see it, they slow down, you know, they avoid it.
link |
And then the local driving is really something
link |
that takes just practice and routine to learn.
link |
You know, this quarter turn,
link |
it feels like the back of the truck sometimes is on delay
link |
when you're backing it up.
link |
So it's like, all right, I'm gonna do a quarter turn
link |
of the wheel now to get the effect that I want
link |
like five seconds from now
link |
in where that tail of that trailer is gonna be.
link |
And there's just no,
link |
I mean, some people have a natural talent for that,
link |
you know, spatial visualization
link |
and kind of calculating those angles and everything,
link |
but there's really no escaping the fact
link |
that you've gotta just do it over and over again
link |
before you're gonna learn how to do it well.
link |
Do you mind sharing how much you were getting paid,
link |
how much you were making as a truck driver
link |
in your time as a truck driver?
link |
Yeah, I started out at 25 cents a mile
link |
and then I got bumped up to 26 cents a mile.
link |
So we had a minimum pay,
link |
which was sort of a new pay scheme
link |
that the industry had started to introduce to, you know,
link |
because there's lots of unpaid work and time.
link |
And so we had a minimum pay of $500 a week
link |
that you would get
link |
if you didn't drive enough miles to exceed that.
link |
You get paid in sort of,
link |
so you get paid when you turn the bills in,
link |
which is the paperwork that goes with the load.
link |
So, you know, you have to get that back to your company
link |
and then that's how they bill the customer.
link |
And so you might get a bunch of those bills
link |
that kind of bunch up in one week.
link |
So, you know, I might get a paycheck for, you know, $1,200.
link |
And I mean, I was a poor graduate student.
link |
So this was real, real money to me.
link |
And so I had this sort of natural incentive to,
link |
you know, earn a lot or to maximize my pay.
link |
Some weeks were that minimum, 500, very few.
link |
And then some I'd get 1200, 1300 bucks.
link |
Pay has gone up, you know,
link |
typical drivers now starting in the 30s, you know,
link |
in the kind of job that I was in.
link |
30 cents per mile, 30 to 35.
link |
So can we try to reverse engineer that math,
link |
how that maps to the actual hours?
link |
So the hours connected to driving are so widely dispersed,
link |
as you said, some of them don't count as actual work,
link |
That's a very interesting discussion
link |
that we'll then continue
link |
when we start talking about autonomous trucking.
link |
But, you know, you're saying all these cents per mile
link |
What, how does that map to like average hourly wage?
link |
Yeah, so, I mean, and this is kind of the,
link |
this is also an interesting technology story in the end.
link |
And it's the technology story that didn't happen.
link |
So pay per mile was, you know,
link |
invented by companies when you couldn't surveil drivers,
link |
you didn't know what they were doing, right?
link |
And you wanted them to have some skin in the game.
link |
And so you'd say, you know, here's the load,
link |
it's going from, you know, for me,
link |
I might start in, you know, the Northeast,
link |
maybe in upstate New York with a load of beer,
link |
and say, here's this load of beer,
link |
bring it to this address in Michigan,
link |
we're gonna pay you by the mile, right?
link |
If I was being paid by the hour,
link |
I might just pull over at the diner and have breakfast.
link |
So you're paid by the mile,
link |
but increasingly over time,
link |
the typical driver is spending more and more time
link |
doing non driving tasks.
link |
There's lots of reasons for that.
link |
One of which is railroads captured a lot of freight
link |
that goes long distances now.
link |
Another one is traffic congestion.
link |
And the other one is that drivers are pretty cheap.
link |
And they're almost always the low people
link |
on the totem pole in some segments.
link |
And so their time is used really inefficiently.
link |
So I might go to that brewery
link |
to pick up that load of Bud Light.
link |
And, you know, their dock staff may be busy
link |
loading up five other trucks.
link |
And they'll say, you know, go over there and sit and wait,
link |
and we'll call you on the CB when the dock's ready.
link |
So you wait there a couple hours, they bring you in,
link |
you know, you never know what's happening in the truck.
link |
Sometimes they're loading it with a forklift,
link |
maybe they're throwing 14 pallets on there full of kegs.
link |
But sometimes it'll take them hours, you know,
link |
and you're sitting in that truck.
link |
And you're essentially unpaid.
link |
You know, then you pull out, you've got control
link |
over what you're gonna get paid
link |
based on how you drive that load.
link |
And then on the other end,
link |
you got a similar situation of kind of waiting, so.
link |
So if that's the way truck drivers are paid,
link |
then there's a low incentive for the optimization
link |
of the supply chain to make them more efficient, right?
link |
To utilize truck labor more efficiently.
link |
So that's a technology problem that,
link |
one of several technology problems that could be addressed.
link |
I mean, so what did, if we just linger on it,
link |
what are we talking about in terms of dollars per hour?
link |
Is it close to minimum wage?
link |
Is it, you know, there's something you talk about,
link |
there was a conception or a misconception
link |
that truckers get paid a lot for their work.
link |
Do they get paid a lot for their work?
link |
And I think that's part of the complexity.
link |
So, you know, what interested me as an ethnographer
link |
about this was, you know, I'm interested
link |
in the kind of economic conceptions
link |
that people have in their heads
link |
and how they lead to certain decisions in labor markets.
link |
You know, why some people become an entrepreneur
link |
and other people become a wage laborer,
link |
or, you know, why some people wanna be doctors
link |
and other people wanna be truck drivers.
link |
That conception, right, is getting shaped
link |
in these labor markets is the argument of the book.
link |
And the fact that drivers can hear,
link |
or potential drivers can hear about these, you know,
link |
workers who make $100,000 plus,
link |
which happens regularly in the trucking industry.
link |
There are many truck drivers who make more
link |
than $100,000 a year, you know, is an attraction.
link |
But the industry is highly segmented.
link |
And so the entry level segment,
link |
and we can probably get into this,
link |
but, you know, the industry is dominated
link |
by a few dozen really large companies
link |
that are self insured and can train new drivers.
link |
So if you want those good jobs,
link |
you've gotta have several years,
link |
up until recently, now the labor market's becoming tighter,
link |
but you had to have several years of accident free,
link |
you know, perfectly clean record driving
link |
The other part of the segment, you know,
link |
those drivers often don't make minimum wage.
link |
But this leads to one of the sort of central issues
link |
that has been in the courts,
link |
and in the legislature, in some states,
link |
is, you know, what should truck drivers get paid for?
link |
Right, the industry, you know,
link |
for the last 30 years or so has said, essentially,
link |
it's the hours that they log for safety reasons
link |
for the Department of Transportation, right?
link |
Now, since the drivers are paid by the mile,
link |
they try to minimize those,
link |
because those hours are limited by the federal government.
link |
So the federal government says,
link |
you can't drive more than 60 hours in a week
link |
as a long haul truck driver.
link |
And so you wanna drive as many miles as you can
link |
in those 60 hours, and so you under report them, right?
link |
And so what happens is the companies say,
link |
well, that guy, you know, he only said he logged 45 hours
link |
of work that week, or 50 hours of work.
link |
That's all we have to pay him minimum wage for.
link |
When in fact, typical truck driver in these jobs will work,
link |
according to most people, would sort of define it as like,
link |
okay, I'm at the customer location, I'm waiting to load,
link |
I'm doing some paperwork, you know,
link |
inspecting the truck, I'm fueling it,
link |
just waiting to, you know, get put in the dock,
link |
80 to 90 hours would be sort of a typical work week
link |
for one of these drivers.
link |
And just when you look at that,
link |
they don't make minimum wage oftentimes.
link |
Right, just to be clear, what we're dancing around here
link |
is that a little bit over, a little bit under minimum wage
link |
is nevertheless most truck drivers seem to be making
link |
close to minimum wage.
link |
Like this is the, so like we maybe haven't made that clear.
link |
There's a few that make quite a bit of money,
link |
but like you're as an entry and for years,
link |
you're operating essentially minimum wage
link |
and potentially far less than minimum wage
link |
if you actually count the number of hours
link |
that are taken out of your life
link |
due to your dedication to trucking.
link |
Well, if you count like the hours taken out of your life,
link |
then you gotta go, you know, maybe a full 24.
link |
That's right, yeah, from family,
link |
from the high quality of life parts of your life.
link |
Yeah, and there's a whole nother set of rules
link |
that the Department of Labor has,
link |
which basically say that a truck driver
link |
who's dispatched away from home for more than a day
link |
should get minimum wage 24 hours a day.
link |
And that could be a state minimum wage,
link |
but typically what it would work out to for most drivers
link |
is that, you know, the minimum wage for a truck driver
link |
should be 50s of thousands, you know, 55, $60,000
link |
should be the minimum wage of a truck driver.
link |
And you've probably heard about the truck driver shortage.
link |
If, you know, which I hope we can talk about,
link |
if the minimum wage for truck drivers
link |
is as it should be on the books at, you know,
link |
around $60,000, we wouldn't have a shortage of truck drivers.
link |
And to me, 60,000 is not a lot of money
link |
for this kind of job.
link |
Cause you're, this isn't, this is essentially two jobs
link |
and two jobs where you don't get to sleep
link |
with your wife or see your kids at night.
link |
That's 60,000 is a very little money for that.
link |
But you're saying if it was 60,000,
link |
you wouldn't even have the shortage.
link |
If that was the minimum.
link |
If that was the minimum.
link |
And I think that's what,
link |
now we have drivers who start in the 30s.
link |
And I mean, so we're talking two, three jobs really,
link |
when you look at the total hours
link |
that people are working at, you know,
link |
they can work over a hundred.
link |
If they're a trainer, you know,
link |
training other truck drivers,
link |
well over a hundred hours a week.
link |
So a job of last resort.
link |
Maybe you can jump around from tangent to tangent.
link |
This is such a fascinating and difficult topic.
link |
I heard that there's a shortage of truck drivers.
link |
So there's more jobs than truck drivers
link |
willing to take on the job.
link |
Is that the state of affairs currently?
link |
I mean, I think the way that you just put that is right.
link |
We don't have a shortage of people
link |
who are currently licensed to do the jobs.
link |
So I'm working on a project for the state of California
link |
to look at the shortage of agricultural drivers.
link |
And the first thing that the DMV commissioner of the state
link |
wanted to look at was, you know,
link |
is there actually a shortage of licensed drivers?
link |
He's like, I've got a database here
link |
of all the people who have a commercial driver's license
link |
who could potentially have the credential to do this.
link |
There are about 145,000 jobs in California
link |
that require a class A CDL,
link |
which would be that commercial driver's license
link |
that you need for the big trucks.
link |
About 145,000 jobs.
link |
The industry in their, you know,
link |
regular promotion of the idea that there's a shortage
link |
is always projecting forward and says,
link |
you know, we're gonna need 165,000 or so
link |
in the next 10 years.
link |
They're currently like 435,000 people licensed
link |
in the state of California to drive one of these big trucks.
link |
So it is not at all an absence of people who,
link |
I mean, and again, going back
link |
to what we were talking about before,
link |
getting that license is not something
link |
that you just walk down to the DMV and take the test.
link |
Like this is somebody who probably quit another job,
link |
was unemployed, and took months to go to a training school,
link |
paid for that training school oftentimes,
link |
left their family for months,
link |
invested in what they thought was gonna be
link |
a long term career, and then said,
link |
you know what, forget it, I can't, I can't do it.
link |
So yeah, so it's not just skill,
link |
it's like they were psychologically invested
link |
potentially for months, if not years,
link |
into this kinds of position as perhaps a position
link |
that if they lose their current job, they could fall too.
link |
Okay, so that's an indication
link |
that there's something deeply wrong with the job,
link |
if so many licensed people are not willing to take it.
link |
What are the biggest problems
link |
of the job of truck driver currently?
link |
Yeah, the job, the problems with the job
link |
and the labor market, right?
link |
But let's start with the job, which is, you know, again,
link |
just so much time that's not compensated directly
link |
for the amount of time.
link |
And that's just psychologically,
link |
and this was a big part of what I studied
link |
for the first book was, you know,
link |
that conception of like, what's my time worth, right?
link |
And like, what truck drivers love is oftentimes,
link |
is that tangible outcome based compensation.
link |
So they say, you know what, you know, honest days work,
link |
I work hard, I get paid for what I do,
link |
I drive 500 miles today,
link |
that's what I'm gonna get paid for.
link |
And then you get to that dock,
link |
and they tell you, sorry, the load's not ready,
link |
go sit over there, and you stew.
link |
And that weight can break you psychologically
link |
because your time every second becomes more worthless.
link |
Yeah, and again, the industry is gonna say, for instance,
link |
okay, well, you know, they've got skin in the game, right?
link |
That argument about sort of compensation
link |
based on sort of output, right?
link |
But that's a holdover from when you couldn't
link |
Now they all have, you know, satellite linked computers
link |
in the trucks that tell these large companies,
link |
this driver was, you know, at this GPS location
link |
for four and a half hours, right?
link |
So if you wanted to compensate them for that time directly,
link |
and the trucker can't control what's happening
link |
on that customer location, you know,
link |
they're waiting for that, you know, firm,
link |
that customer to tell them, hey, pull in there.
link |
And so what it becomes is just a way to shift
link |
the inefficiencies and the cost of that onto that driver.
link |
Now it's competitive for customers.
link |
So if you're Walmart, you might have your choice
link |
of a dozen different trucking companies
link |
that could move your stuff.
link |
And if one of them tells you, hey, you're not moving
link |
our trucks in and out of your docks fast enough,
link |
we're gonna charge you for how long our truck
link |
is sitting on your lot.
link |
If you're Walmart, you're gonna say,
link |
I'll go see what the other guy says, right?
link |
And so companies are gonna allow that customer
link |
to essentially waste that driver's time, you know,
link |
in order to keep that business.
link |
Can you try to describe the economics,
link |
the labor market of the situation?
link |
You mentioned freight and railroad.
link |
What is the sort of the dynamic financials,
link |
the economics of this that allow for such low salaries
link |
to be paid to truckers?
link |
Like what's the competition?
link |
What's the alternative to transporting goods via trucks?
link |
Like what seems to be broken here
link |
from an economics perspective?
link |
Yeah, so it's, well, nothing.
link |
It's a perfect market, right?
link |
I mean, so for economists, this is how it should work, right?
link |
But the inefficiencies, like you said,
link |
sorry to interrupt, are pushed to the truck driver.
link |
Doesn't that like spiral, doesn't that lead to
link |
a poor performance on the part of the truck driver
link |
and just like make the whole thing more and more inefficient
link |
and it results in lower payment
link |
to the truck driver and so on.
link |
It just feels like in capitalism,
link |
you should have a competing solution
link |
in terms of truck drivers.
link |
Like another company that provides transportation via trucks
link |
that creates a much better experience for truck drivers,
link |
making them more efficient, all those kinds of things.
link |
How is the competition being suppressed here?
link |
Yeah, so it is, the competition is based on who's cheaper.
link |
And this is the cheapest way to move the freight.
link |
Now, there are externalities, right?
link |
I mean, so this is the explanation
link |
that I think is obvious for this, right?
link |
There are lots of costs that,
link |
whether it's that driver's time,
link |
whether it's the time without their family,
link |
whether it's the fact that they drive through congestion
link |
and spew lots of diesel particulates into cities
link |
where kids have asthma and make our commutes longer
link |
rather than more efficiently use their time
link |
by sort of routing them around congestion
link |
and rush hour and things like that.
link |
This is the cheapest way to move freight.
link |
And so it's the most competitive.
link |
A big part of this is public subsidy of training.
link |
So when those workers are not paying for the training,
link |
you and I often are.
link |
So if you lose your job because of foreign trade
link |
or you're a veteran using your GI benefits,
link |
you may very well be offered training,
link |
publicly subsidized training to become a truck driver.
link |
And so all of these are externalities
link |
that the companies don't have to pay for.
link |
And so this makes it the most profitable way to move freight.
link |
So trucks is way cheaper than trains?
link |
Well, over the long,
link |
so one of the big stories for these companies
link |
is that the average length of haul,
link |
which becomes very important for self driving trucks,
link |
the average length of haul has been steadily declining.
link |
Over the last 15 years or so,
link |
and this is industry collected data
link |
from sort of the big firms that report it,
link |
but roughly been cut in half from typically
link |
about a thousand miles to under 500.
link |
And under 500 is what a driver can move in a day, right?
link |
So you can get loaded, drive and unload,
link |
around 400 miles or something like that.
link |
I wanna steal a good question from the Penn Gazette
link |
interview you did, which people should read.
link |
It's a great interview.
link |
Was there a golden age for long haul truckers in America?
link |
And if so, this is just a journalistic question.
link |
And if so, what enabled it and what brought it to an end?
link |
Wow, I might have to have you read my answer to that.
link |
That was a few years ago,
link |
be interesting to compare what I'll say, but.
link |
I mean, one bigger question to ask, I guess,
link |
is like Johnny Cash wrote a lot of songs about truckers.
link |
There used to be a time when perhaps falsely,
link |
perhaps it's part of the kind of perception
link |
that you study with the labor markets and so on.
link |
There was a perception of truckers being,
link |
first of all, a lucrative job
link |
and second of all, a job to be desired.
link |
Yeah, so I mean, this is,
link |
the trucking industry to me is fascinating,
link |
but I think it should be fascinating to a lot of people.
link |
So the golden age was really two different kinds
link |
of markets as well, right?
link |
Today we have really good jobs and some really bad jobs.
link |
We had the Teamsters Union
link |
that controlled the vast majority of employee jobs.
link |
And even where they had something called
link |
the National Master Freight Agreement.
link |
And this was Jimmy Hoffa who led the union
link |
through its sort of critical period by the mid 60s
link |
had unified essentially the entire nation's
link |
trucking labor force under one contract.
link |
Now you were either covered by that contract
link |
or your employer paid a lot of attention to it.
link |
And so by the end of the 1970s,
link |
the typical truck driver was making
link |
well more than $100,000,
link |
typical truck driver was making more than $100,000
link |
in today's dollars and was home every night.
link |
That was without a doubt and even more
link |
than unionized auto workers, steel workers,
link |
10, 20% more than those workers made.
link |
That was the golden age for sort of job quality,
link |
wages, teamster power.
link |
They were without a doubt the most powerful union
link |
in the United States at that time.
link |
At the same time in the 1970s,
link |
you had the mythic long haul trucker.
link |
And these were the guys who were kind of on the margins
link |
of the regulated market,
link |
which is what the teamsters controlled.
link |
A lot of them were in agriculture,
link |
which was never regulated.
link |
So in the new deal, when they decided to regulate trucking,
link |
they didn't regulate agriculture
link |
because they didn't wanna drive up food prices,
link |
which would hurt workers in urban areas.
link |
So they essentially left agricultural truckers out of it.
link |
And that's where a lot of the kind of outlaw,
link |
you know, asphalt cowboy imagery that we get.
link |
And I grew up, I know you didn't grow up in the US
link |
at this sort of, you know, as a young child.
link |
And I'm a bit older than you, but in the late 70s,
link |
there were movies and TV shows and CBs were a craze.
link |
And it was all these kind of outlaw truckers
link |
who were out there hauling some unregulated freight.
link |
They weren't supposed to be trying to avoid the bears,
link |
you know, who are the cops.
link |
And, you know, with all this salty language
link |
and these like, you know, terms that only they understood
link |
and, you know, the partying at diners and popping pills,
link |
you know, the California turnarounds.
link |
So asphalt cowboys, truly.
link |
So it's like another form of cowboy movies.
link |
Oh, absolutely, absolutely.
link |
And I think that sort of masculine ethos of like,
link |
you got 40,000 pounds of something you care about,
link |
I'm your guy, you know,
link |
you needed to go from New York to California,
link |
don't worry about it, I got it.
link |
That's appealing and it's tangible, right?
link |
And you think about people who don't wanna be paper pusher
link |
and sit in the, I deal with office politics,
link |
like just give me what you care about
link |
and I'll take care of it, you know, just pay me fair,
link |
you know, and that appeals.
link |
You mentioned unions, Teamsters, Jimmy Hoffa.
link |
Big question, maybe difficult question.
link |
What are some pros and cons of unions historically
link |
and today in the trucking space?
link |
Yeah, well, if you're a worker, there are a lot of pros.
link |
And I don't, you know,
link |
and this was one of the things I talked to truckers about
link |
Yeah, what's their perception of Jimmy Hoffa,
link |
for example, of unions?
link |
Yeah, so, and this was probably one of the central
link |
hypotheses that I had going in there.
link |
And it may sound, you know,
link |
someone who does hard science, right?
link |
You may hear a social scientist, you know,
link |
sort of use that terminology,
link |
even other social scientists.
link |
Yeah, you know, they don't like it,
link |
but I do like to think that way.
link |
And my initial hypothesis was that, you know,
link |
and it's very simple,
link |
that, you know, the tenure of the driver in the industry
link |
would have a strong effect on how they viewed unions.
link |
That, you know, somebody who had experienced unions
link |
would be more favorable
link |
and someone who had not would not be, right?
link |
And that turned out to be the case without a doubt.
link |
But in an interesting way,
link |
which was that even the drivers
link |
who were not part of the union,
link |
who in the kind of public debate of deregulation
link |
were portrayed as these kind of small business truckers
link |
who were getting shut out by the big regulated monopolies
link |
and the Teamsters Union, you know,
link |
the corrupt Teamsters Union.
link |
Even those drivers longed for the days of the Teamsters
link |
because they recognized the overall market impact
link |
That trucking just naturally
link |
tended toward excessive competition
link |
that meant that there was no profit to be made
link |
and oftentimes you'd be operating at a loss.
link |
And so even these, you know,
link |
the asphalt cowboy owner operators from back in the day
link |
would tell me when the Teamsters were in power,
link |
I made a lot more money.
link |
And, you know, this is, you know, unions,
link |
at least those kinds of unions, like the Teamsters,
link |
you know, there's, I think a lot of misconceptions today
link |
sort of popularly about what unions did back then.
link |
They tied wages to productivity.
link |
Like that was the central thing that the Teamsters Union did.
link |
And, you know, there were great accounts
link |
of sort of Jimmy Hoffa's perspective
link |
for all his portrayal as sort of corrupt
link |
and criminal, and there's, you know, I'm not disputing that.
link |
He broke a lot of laws.
link |
He was remarkably open about who he was and what he did.
link |
He actually invited a pair, a husband and wife team
link |
of Harvard economists to follow him around
link |
and like opened up the Teamsters books to them
link |
so that they could see how he was, you know,
link |
thinking about negotiating with the employers.
link |
And the Teamsters, and this goes back to the beginning,
link |
and this goes back well before Hoffa,
link |
back to the, you know, 1800s,
link |
they understood that workers did better
link |
if their employers did better.
link |
And the only way the employers would do better
link |
was if they controlled the market.
link |
And so oftentimes the corruption in trucking
link |
was initiated by employers who wanted to limit competition
link |
and they knew they couldn't limit competition
link |
without the support of labor.
link |
And so you'd get these collusive arrangements
link |
between employers and labor to say no new trucking companies.
link |
There are 10 of us, that's enough.
link |
We control Seattle, we're gonna set the price
link |
and we're not gonna be undercut.
link |
When there's a shortage of trucks around, it's great,
link |
rates go up, but you get too many trucks.
link |
It's very often that you end up operating at a loss
link |
just to keep the doors open.
link |
You know, you don't have any choice.
link |
You can't, it's what economists called derived demand.
link |
You can't like make up a bunch of trucking services
link |
and store it in a warehouse, right?
link |
You gotta keep those trucks moving to pay the bills.
link |
Can we also lay out the kind of jobs that are in trucking?
link |
What are the best jobs in trucking?
link |
What are the worst jobs in trucking?
link |
What are we, how many jobs are we talking about today?
link |
And what kind of jobs are there?
link |
So there are a number of different segments
link |
and the first part would be, you know, are you offering,
link |
the first question would be,
link |
are you offering services to the public
link |
or are you moving your own freight, right?
link |
So are you a retailer, say Walmart or, you know,
link |
a paper company or something like that
link |
that's operating your own fleet of trucks?
link |
That's private trucking.
link |
For hire are the folks who, you know,
link |
offer their services out to other customers.
link |
So you have private and for hire.
link |
In general, for hire pays less.
link |
Is that because of the, something you talk about
link |
with employee versus contractor situation
link |
or are they all tricked or led to become contractors?
link |
That can become a part of it as a strategy,
link |
but the fundamental reason is competition.
link |
So those private carriers aren't in competition
link |
with other trucking fleets, right?
link |
For their own in house services.
link |
So, you know, they tend to, and this, you know,
link |
the question of why private versus for hire
link |
because for hire is cheaper, right?
link |
And so if you need that, if that trucking service
link |
is central to what you do and you cannot afford disruptions
link |
or volatility in the price of it, you keep it in house.
link |
You should be willing to pay more for that
link |
because it's more valuable too
link |
and you keep it in house and that.
link |
So that's an interesting distinction.
link |
What about, and this is kind of moving towards
link |
our conversation of what can and can't be automated.
link |
How else does it divide the different trucking jobs?
link |
So the next big chunk is kind of
link |
how much stuff are you moving, right?
link |
And so we have what's called truckload
link |
and truckload means, you know, you can fill up a trailer
link |
either by volume or by weight and then less than truckload.
link |
Less than truckload, the official definition
link |
is like less than 10,000 pounds.
link |
You know, this is gonna be a couple of pallets of this,
link |
a couple of pallets of that.
link |
The process looks really different, right?
link |
So that truckload is, you know, point A to point B,
link |
I'm buying, you know, a truckload of bounty paper towels,
link |
I'm bringing it into, you know, my distribution center,
link |
go pick it up at the bounty plant,
link |
bring it to my distribution center, right?
link |
Nowhere in between do you stop.
link |
At least process that freight.
link |
Less than truckload, what you've got is terminal systems.
link |
And this is what you had under regulation too.
link |
And so these terminal systems, what you do is you do
link |
a bunch of local pickup and delivery,
link |
maybe with smaller trucks,
link |
and you pick up two pallets of this here,
link |
four pallets of this there, you bring it to the terminal,
link |
you combine it based on the destination,
link |
you then create a full truckload, you know, trailer,
link |
and you send it to another terminal
link |
where it gets broken back down,
link |
and then out for local delivery.
link |
That's gonna look a lot like if you send a package by UPS,
link |
right, they pick all these parcels, right,
link |
figure out where they're all going,
link |
put them on planes or in trailers
link |
going to the same destination,
link |
then break them out to put them
link |
in what they call package cars.
link |
So, before I ask you about autonomous trucks,
link |
let's just pause for your experience as a trucker.
link |
Did it get lonely?
link |
Like, can you talk about some of your experiences
link |
of what it was actually like?
link |
Did it get lonely?
link |
Yeah, no, I mean, it was, I didn't have kids at the time.
link |
Now I have kids, I can't even imagine it.
link |
You know, I've been married for five years at the time.
link |
My wife hated it, I hated it.
link |
You know, I describe in the book
link |
the experience of being stuck,
link |
if I remember correctly, it was like Ohio
link |
at this truck stop in the middle of nowhere
link |
and like, you know, sitting on this concrete barrier
link |
and just watching fireworks in the distance
link |
and like eating Chinese food on the 4th of July.
link |
And you know, my wife calls me from like the family barbecue
link |
and our anniversary is July 8th.
link |
And she's like, are you gonna be home?
link |
And I'm like, I don't know, you know.
link |
I have a cousin whose husband drove truck
link |
as a truck driver would say, drove truck for a while.
link |
And he told me before I went into it,
link |
he was like, the advantage you have is that you know
link |
that you're not gonna be doing this long term.
link |
Like, and Lex, I can't even like,
link |
the emotional content of some of these interviews,
link |
I mean, I would sit down at a truck stop with somebody
link |
I had never met before and you know, you open the spicket
link |
and the last question I would ask drivers
link |
was that by the time I really sort of figured out
link |
how to do it, the last question I would ask them is,
link |
you know, what advice would you give to somebody,
link |
like your nephew, you know, a family friend asks you
link |
about what it's like to be a driver and should they do it?
link |
What advice would you give them?
link |
And this question, some of these, you know,
link |
grizzled old drivers, you know, tough, tough guys,
link |
would that question would like, some of them would break down
link |
and they would say, I would say to them,
link |
you better have everything
link |
that you ever wanted in life already.
link |
Because I've had a car that I've had for 10 years,
link |
it's got 7,000 miles on it.
link |
I own a boat that hasn't seen the water in five years.
link |
My kids, I didn't raise them.
link |
Like I'd be out for two weeks at a time,
link |
I'd come home, my wife would give me two kids to punish,
link |
a list of things to do, you know, on Saturday night
link |
and I might leave out Sunday night or Monday morning.
link |
You know, I come home dead tired,
link |
my kids don't know who I am.
link |
And you know, it was just like,
link |
it was heartbreaking to hear those stories.
link |
And then before you know it, you know,
link |
life is short and just the years run away.
link |
Hard question to ask in that context,
link |
but what's the best,
link |
what was the best part of being a truck driver?
link |
Was there moments that you truly enjoyed on the road?
link |
There was, there's definitely a pride and mastery of,
link |
you know, even basic competence
link |
of sort of piloting this thing safely.
link |
There's a lot of responsibility to it.
link |
That thing's dangerous and you know it.
link |
So there's some pride there.
link |
For me personally, and I know for a lot of other drivers,
link |
it's just like seeing these behind the scenes places
link |
that you know exist in our economy.
link |
And I think we're all much more aware of them now
link |
after COVID and supply chain mess that we have.
link |
I don't know if we'll talk about that,
link |
but you know, you get to see those places.
link |
You know, you get to see those ports.
link |
You get to see the place where they make the cardboard boxes
link |
that the Huggies diapers go in.
link |
Or the warehouse full of Bud Light.
link |
I moved Bud Light from like upstate New York
link |
and the first load like went to Atlanta, you know?
link |
And then a couple months later,
link |
I circled back through that same brewery
link |
and I brought a load of Bud Light out to Michigan.
link |
And I was like, holy shit, all the Bud Light,
link |
like, you know, for this whole giant swath
link |
of the United States comes from this one plant,
link |
this cavernous plant with like kegs of beer.
link |
And you see that part of the economy
link |
and it's like, you're almost like you're an economic tourist.
link |
And I think all, everybody kind of appreciates that.
link |
Like kind of, it's almost like a behind the scenes tour.
link |
That wears off after a few months, you know?
link |
You start to see new things less and less frequently.
link |
At first, everything's novel and sort of life on the road.
link |
And then it becomes just endless miles of white lines
link |
and yellow lines and truck stops.
link |
And the days just blur together.
link |
You know, it's one loading dock.
link |
It's one loading dock after another.
link |
So you lose the magic of being on the road.
link |
Yeah, it's very rare the driver that doesn't.
link |
You mentioned COVID and supply chain.
link |
While being this, for a brief time,
link |
this member of the supply chain,
link |
what have you come to understand about our supply chain,
link |
United States and global and its resilience
link |
against strategies, catastrophic in the world?
link |
Like COVID, for example.
link |
Yeah, I mean, we have built really long,
link |
really lean supply chains.
link |
And just by definition, they're fragile.
link |
You know, the current mess that we have,
link |
it's not gonna clear by Christmas.
link |
It will be lucky if it clears by next Christmas.
link |
Can you describe the current mess in supply chain
link |
that you're referring to?
link |
Yeah, so we've got pile ups of ships
link |
off the coast of California, Long Beach,
link |
and LA in particular, in bad shape.
link |
You know, last I checked, it was around 60 ships,
link |
all of which are holding thousands of containers
link |
full of stuff that retailers were hoping
link |
was gonna be on shelves for the holiday season.
link |
Meanwhile, the port itself has stacks and stacks
link |
of containers that they can't get rid of.
link |
The truckers aren't showing up to pick up
link |
the containers that are there,
link |
so they can't offload the ships that are waiting.
link |
And why aren't the truckers picking it up?
link |
Partly because there's a long history of inefficiency
link |
in making them wait,
link |
but it's because the warehouses are full.
link |
So we've had all these perverse outcomes
link |
that no one really expected.
link |
Like in the middle of all these shortages,
link |
people are stockpiling stuff.
link |
So there are suppliers who used to keep two months
link |
of supply of bottled water on hand.
link |
And after going through COVID and not having supply
link |
to send to their customers,
link |
they're like, we need three months.
link |
Well, our system is not designed for major storage of goods
link |
to go up 50% in a category.
link |
If you're a warehouse operator,
link |
you know, you wanna be 90% plus.
link |
You don't want a lot of open bays sitting around.
link |
So we don't have 10% extra capacity in warehouses.
link |
We don't have 10% of them.
link |
Trucking capacity can fluctuate a bit,
link |
but you don't have that kind of slack.
link |
And now, I mean, and we saw this
link |
right when people shifted consumption.
link |
And I get a little mad when people talk about panic buying
link |
as kind of the reason that we had all these shortages.
link |
It really, like it's preventing us from understanding,
link |
you know, the real problem there,
link |
which is that lean supply chain.
link |
Sure, there was some panic buying, you know,
link |
no doubt about it,
link |
but we had an enormous shift in people's behavior.
link |
So with my sister and brother in law,
link |
I own a couple of small businesses and we serve food, right?
link |
So we get, you know, food from Cisco.
link |
Cisco couldn't get rid of food, right?
link |
Because nobody's eating out.
link |
So they've got, you know, 50 pounds sacks of flour,
link |
you know, sitting in their warehouse
link |
that they can't get rid of.
link |
They've got cases of lettuce and meat and everything else
link |
that's just gonna go bad.
link |
So that panic buying certainly exacerbated some things
link |
like toilet paper and whatever,
link |
but we saw just a massive change in demand.
link |
And our supply chains are based on historical data, right?
link |
So, you know, that stuff leaves Asia,
link |
you know, months before you wanna have it on the shelves
link |
and you're predicting based on last year, you know,
link |
what you want on that shelf.
link |
And so it's a, you know, I guess at its best,
link |
it's a beautiful symphony of lots of moving parts,
link |
but now everyone can't get on the same page of music.
link |
But it's not resilient to changes
link |
in on mass human behavior.
link |
So even like I read somewhere,
link |
maybe you can tell me if it's true in relation to food,
link |
it's just the change of human behavior
link |
between going out to restaurants versus eating at home.
link |
As a species, we consume a lot less food that way.
link |
Apparently what I read in restaurants,
link |
like there's a lot of food just thrown out.
link |
It's part of the business model.
link |
And so like you then have to move a lot more food
link |
through the whole supply chain.
link |
And now because you're consuming, you know,
link |
there's leftovers at home,
link |
you're consuming much more of the food you're getting
link |
when you're eating at home,
link |
that's creating these bottleneck situations,
link |
problems as you're referring to,
link |
too much in a certain place, not enough in another place.
link |
And it's just the supply chain is not robust
link |
those kinds of dynamic shifts in who gets what where.
link |
Yeah, I mean, so, and I have worked in agriculture a bit
link |
on sort of the supply side, you know,
link |
and there are product categories, right?
link |
Where 30% of the crop raised does not get used, right?
link |
Just gets plowed under or wasted.
link |
But here's the importance of this
link |
in sort of getting this right, you know, like that,
link |
not that like panic buying, you know,
link |
blame the irrational consumer, you know,
link |
look at the hard sort of truth
link |
of the way we've set up our economy.
link |
And I'll ask you this, Lex, I know you're a hopeful,
link |
optimistic person.
link |
I mean, I write about problems all the time.
link |
And so people think I'm sort of like a,
link |
just a Debbie Downer, you know, pessimist,
link |
but I'm a glass half full kind of guy.
link |
Like I want to identify problems so we can solve them.
link |
So let me ask you this,
link |
we've got these long lean supply chains.
link |
In the future, do you see more environmental problems
link |
that could disrupt them,
link |
more geopolitical problems that could disrupt trade
link |
from Asia, you know, other institutional failures?
link |
Do those things seem, you know,
link |
potentially more likely in the future
link |
than they have been in say the last 20 years?
link |
Yeah, it almost absolutely seems to be the case.
link |
So you then have to ask the question of
link |
how do we change our supply chains?
link |
Whether it's making more resilient
link |
or make them less densely connected,
link |
you know, building a, it's like a, what is it?
link |
You know, the Tesla model for in the automotive sector
link |
of like trying to build everything,
link |
like trying to get the factory to do as much as possible
link |
with as little reliance on widely distributed sources
link |
of the supply chain as possible.
link |
So maybe like rethinking how much we rely
link |
on the infrastructure of the supply chain.
link |
Yeah, I mean, you know, there's some basic,
link |
and I assume, right, that there are a lot of folks
link |
in corporate boardrooms looking at risk
link |
and saying that didn't go well,
link |
and maybe it could have even gone worse.
link |
Maybe we need to think about reshoring, right?
link |
At the very least, one of the things
link |
that I'm hearing about anecdotally
link |
is that they're starting stuff up, you know,
link |
when they can, right?
link |
Which is, that's probably not sustainable, right?
link |
I mean, at some point, somebody in that corporate boardroom
link |
is gonna say, you know, guys, inventory is getting
link |
kind of heavy and the cost of that is like,
link |
do we, can we really justify that much longer
link |
to the shareholders, right?
link |
We can back off and start, you know,
link |
back, things are back to normal, let's lean out.
link |
Well, my hope is that there's a technology solution
link |
to a lot of aspects of this.
link |
So one of them on the supply chain side
link |
is collecting a lot more data,
link |
like having much more integrated
link |
and accurate representation of the inventory
link |
all over the place,
link |
and the available transportation mechanisms,
link |
the trucks, the all kinds of freight,
link |
and how in the different models
link |
of the possible catastrophes that can happen,
link |
like how will the system respond?
link |
So having a really solid model that you're operating under
link |
as opposed to just kind of being in emergency response mode
link |
under poor, incomplete information,
link |
which is what seems like is more commonly the case,
link |
except for things like you said, Walmart and Amazon,
link |
they're trying to internally get their stuff together
link |
on that front, but that doesn't help
link |
the rest of the economy.
link |
So another exciting technological development
link |
as you write about, as you think about is autonomous trucks.
link |
So these are often brought up in different contexts
link |
as the examples of AI and robots taking our jobs.
link |
Should we be concerned?
link |
I think they've really come to epitomize
link |
this anxiety over automation, right?
link |
It's such a simple idea, right?
link |
Truck that drives itself,
link |
classic blue collar job that pays well,
link |
guy maybe with not a lot of other good options, right?
link |
To sort of make that same income easily,
link |
and you build a robot to take his job away, right?
link |
So I think 2016 or so,
link |
that was the sort of big question out there,
link |
and that's actually how I started studying it, right?
link |
I just wrapped up the book,
link |
just so happened that somebody was working at Uber,
link |
Uber had just bought auto, saw the book and was like,
link |
hey, can you come out and talk to our engineering teams
link |
about what life is like for truck drivers
link |
and maybe how our technology could make it better.
link |
And at that time, there were a lot of different ideas
link |
about how they were gonna play out, right?
link |
So while the press was saying,
link |
all truckers are gonna lose their jobs,
link |
there were a lot of people in these engineering teams
link |
who thought, okay, if we've got an individual owner operator
link |
and they can only drive eight or 10 hours a day,
link |
they hop in the back, they get their rest,
link |
and the asset that they own works for them, right?
link |
Sort of perfect, right?
link |
And at that time, there were a bunch of reports
link |
that came out and sort of basically what people did
link |
was they took the category of truck driver.
link |
Some people took a larger category from BLS
link |
of sales and delivery workers
link |
that was about three and a half million workers
link |
and others took the heavy duty truck driver category,
link |
which was at the time about 1.8 million or so.
link |
And they picked a start date and a slope
link |
and said, let's assume that all these jobs
link |
are just gonna disappear.
link |
And really smart researcher, Annette Bernhardt
link |
at the Labor Center at UC Berkeley
link |
was sort of looking around for people
link |
who were sort of deeply into industries
link |
to complicate those analyses, right?
link |
And reached out to me and was like,
link |
what do you think of this?
link |
And I said, the industry is super diverse.
link |
I haven't given a ton of thought, but it can't be that.
link |
It's not that simple, it never is.
link |
And so she was like, will you do this?
link |
And I was like ready to move on to another topic.
link |
I had like been in trucking for 10 years
link |
and that's how I started looking at it.
link |
And it is, it's a lot more complicated.
link |
And the initial impacts, and here's the challenge I think,
link |
and it's not just a research challenge,
link |
it's the fundamental public policy challenge
link |
is we look at the existing industry
link |
and the impacts, the potential impacts,
link |
they're not, you know, nothing.
link |
For some communities and some kinds of drivers,
link |
they're gonna be hard.
link |
And there are a significant number of them.
link |
Nowhere near what people thought.
link |
You know, I estimate it's like around 300,000,
link |
but that's a static picture of the existing industry.
link |
And here's the key with this is, at least in my conclusion
link |
is this is a transformative technology.
link |
We are not going to swap in self driving trucks
link |
for human driven trucks and all else stays the same.
link |
This is gonna reshape our supply chains.
link |
It's gonna reshape landscapes.
link |
It's gonna affect our ability to fight climate change.
link |
This is a really important technology in this space.
link |
Do you think it's possible to predict the future
link |
of the kind of opportunities it will create,
link |
how it will change the world?
link |
So like when you have the internet,
link |
you can start saying like all the kinds of ways
link |
that office work, all jobs will be lost
link |
because it's easy to network.
link |
And then software engineering allows you to automate
link |
a lot of the tasks like Microsoft Excel does, you know.
link |
But it opened up so many opportunities,
link |
even with things that are difficult to imagine,
link |
like with the internet, I don't know, Wikipedia,
link |
which is widely making accessible information.
link |
And that increased the general education globally by a lot,
link |
all those kinds of things.
link |
And then the ripple effects of that
link |
in terms of your ability to find other jobs
link |
is probably immeasurable.
link |
So is it just a hopeless pursuit to try to predict
link |
if you talk about these six different trajectories
link |
that we might take in automating trucks,
link |
but like as a result of taking those trajectories,
link |
is it a hopeless pursuit to predict
link |
what the future will result in?
link |
Because it's the wrong question.
link |
The question is, what do we want the future to be
link |
and let's shape it, right?
link |
And I think this is, and this is the only point
link |
that I really wanna make in my work
link |
for the foreseeable future,
link |
is that we have got to get out of this mindset
link |
that we're just gonna let technology kind of go
link |
and it's a natural process and whatever pops out
link |
will fix the problems on the backside.
link |
And we've got to recognize that one,
link |
that's not what we do, right?
link |
You know, and self driving vehicles
link |
is just such a perfect example, right?
link |
We would not be sitting here today
link |
if the Defense Department, right?
link |
If Congress in 2000 had not written into legislation
link |
funding for the DARPA challenges,
link |
which followed, actually I think the funding came
link |
a couple of years later,
link |
but the priority that they wrote in 2000
link |
was let's get a third of all ground vehicles
link |
in our military forces unmanned, right?
link |
And this was before aerial unmanned vehicles
link |
had really sort of proven their worth.
link |
They would come to be incredibly,
link |
like, you know, just blow people's minds
link |
in terms of their additional capabilities,
link |
the lower costs, you know,
link |
keeping soldiers out of harm's way.
link |
Now, of course they raised other problems
link |
and considerations that I think we're still wrestling with,
link |
but that was even before that they had this priority.
link |
We would not be sitting here today
link |
if Congress in 2000 had not said, let's bring this about.
link |
So they already had that vision, actually.
link |
I didn't know about that.
link |
So for people who don't know the DARPA challenges
link |
is the events that were just kind of like
link |
these seemingly small scale challenges
link |
that brought together some of the smartest roboticists
link |
in the world, and that somehow created enough of a magic
link |
where ideas flourished,
link |
both engineering and scientific,
link |
that eventually then was the catalyst
link |
for creating all these different companies
link |
that took on the challenge.
link |
Some failed, some succeeded,
link |
some are still fighting the good fight.
link |
And that somehow just that little bit of challenge
link |
was the essential spark of progress
link |
that now resulted in this beautiful up and down wave
link |
of hype and profit and all this kind of weird dance
link |
where the B word, billions of dollars
link |
have been thrown around and we still don't know.
link |
And the T word, trillions of dollars
link |
in terms of transformative effects of autonomous vehicles.
link |
And all that started from DARPA
link |
and that initial vision of, I guess, as you're saying,
link |
of automating part of the military supply chain.
link |
I did not know that.
link |
That's interesting.
link |
So they had the same kind of vision for the military
link |
as we're not talking about a vision for the civilian,
link |
whether it's trucking or whether it's autonomous vehicle,
link |
sort of a ride sharing kind of application.
link |
Yeah, I mean, what an incredible spark, right?
link |
And just the story of what it produced, right?
link |
I mean, your own work on self driving, right?
link |
I mean, you've studied it as an academic, right?
link |
How many great researchers and minds have been harnessed
link |
by this outcome of that spark, right?
link |
And I think this is sort of theoretically about technology,
link |
right, this is what makes it sort of so great
link |
is that this is what makes us human, in my opinion, right?
link |
Is that you conceive of something in your mind
link |
and then you bring it into reality, right?
link |
I mean, that's what is so great about it.
link |
Sometimes you're too dumb to realize how difficult it is
link |
And then eventually you're in too deep.
link |
You might as well solve the problem.
link |
Well, and maybe we're in that situation right now
link |
with self driving.
link |
But, you know, and so let me throw this out there.
link |
I'd be curious to hear your thoughts on it.
link |
But truck drivers always ask me, like, is this for real?
link |
Like, is this really, like, it's harder than they think,
link |
like, right, and they can't really do this.
link |
And, you know, at first I was like, look, you know,
link |
this is like the defense department
link |
and like basically the top computer science
link |
and robotics departments in the world.
link |
And now Silicon Valley with billions of dollars in funding
link |
and just, you know, some of the smartest, hardest working,
link |
most visionary people focused on what is clearly,
link |
you know, a gigantic market, right?
link |
And what I tell them is like,
link |
if self driving vehicles don't happen,
link |
I think this will be the biggest technology failure story
link |
I don't know of anything else that is just galvanized.
link |
I mean, you've had people in garages or weird inventors
link |
work on things their whole lives and come really close
link |
and it never happens and it's a great failure story, right?
link |
But never have we had like whole,
link |
I mean, we're talking about GM, right?
link |
And these are not, you know, these are not tech companies,
link |
right, these are industrial giants, right?
link |
What were in the 20th century,
link |
the pinnacle of industrial production in the world
link |
in human history, right?
link |
And they're focused on it now.
link |
So if we don't pull this off, it's like, wow.
link |
It's fascinating to think about.
link |
I've never thought of it that way.
link |
There was a mass hysteria on a level
link |
in terms of excitement and hype
link |
on a level that's probably unparalleled in technology space.
link |
Like I've seen that kind of hysteria just studying history
link |
when you talk about military conflict.
link |
So we often wage war with a dream of making a better world
link |
and then realize it costs trillions of dollars
link |
and then we step back and like, and go, wait a minute,
link |
what do we actually get for this?
link |
But in the space of technology,
link |
it seems like all these kinds of large efforts
link |
This, you're right.
link |
It seems like, it seems like even GM and Ford
link |
and all these companies now are a little bit like,
link |
hey, or Toyota and even Tesla,
link |
like, are we sure about this?
link |
And it's fascinating to think about
link |
when you tell the story of this,
link |
this could be one of the big first, perhaps,
link |
but by far the biggest failures of the dream
link |
in the space of technology.
link |
That's really interesting to think about.
link |
I was a skeptic for a long time
link |
because of the human factor.
link |
Because for business to work in the space,
link |
you have to work with humans
link |
and you have to work with humans at every level.
link |
So in the truck driving space,
link |
you have to work with the truck driver,
link |
but you also have to work with the society
link |
that has a certain conception of what driving means.
link |
And also you have to have work with businesses
link |
that are not used to this extreme level of technology,
link |
you know, in the basic operation of their business.
link |
So I thought it would be really difficult
link |
to move to autonomous vehicles in that way.
link |
But then I realized that there's certain companies
link |
that are just willing to take big risks
link |
and really innovate.
link |
I think the first impressive company to me was Waymo
link |
or what was used to be the Google self driving car.
link |
And I saw, okay, here's a company
link |
that's willing to really think longterm
link |
and really try to solve this problem, hire great engineers.
link |
Then I saw Tesla with Mobileye when they first had.
link |
I thought, actually Mobileye is the thing that impressed me.
link |
When I sat down, I thought,
link |
because I'm a computer vision person,
link |
I thought there's no way a system could keep me in lane
link |
long enough for it to be a pleasant experience for me.
link |
So from a computer vision perspective,
link |
I thought there'd be too many failures.
link |
It'd be really annoying.
link |
It'd be a gimmick, a toy.
link |
It wouldn't actually create a pleasant experience.
link |
And when I first was gotten Tesla with Mobileye,
link |
the initial Mobileye system,
link |
it actually held to lane for quite a long time
link |
to where I could relax a little bit.
link |
And it was a really pleasant experience.
link |
I couldn't exactly explain why it's pleasant
link |
because it's not like I still have to really pay attention,
link |
but I can relax my shoulders a little bit.
link |
I can look around a little bit more.
link |
And for some reason that was really reducing the stress.
link |
And then over time, Tesla with a lot of the revolutionary
link |
stuff they're doing on the machine learning space
link |
made me believe that there's opportunities here to innovate,
link |
to come up with totally new ideas.
link |
Another very sad story that I was really excited about
link |
is Cadillac SuperCruise system.
link |
It is a sad story because I think I vaguely read in the news
link |
they just said they're discontinuing SuperCruise,
link |
but it's a nice innovative way
link |
of doing driver attention monitoring
link |
and also doing lane keeping.
link |
And it just innovation could solve this
link |
in ways we don't predict.
link |
And same with in the trucking space,
link |
it might not be as simple as like journalists envision
link |
a few years ago, where everything's just automated.
link |
It might be gradually helping out the truck driver
link |
in some ways that make their life more efficient,
link |
more effective, more pleasant,
link |
remove some of the inefficiencies
link |
that we've been talking about in totally innovative ways.
link |
And that I still have that dream
link |
that I believe to solve the fully autonomous driving problem
link |
we're still many years away,
link |
but on the way to solving that problem,
link |
it feels like there could be,
link |
if there's bold risk takers and innovators in this space,
link |
there's an opportunity to come up
link |
with like subtle technologies that make all the difference.
link |
That's actually just what I realized
link |
is sometimes little design decisions
link |
make all the difference.
link |
It's the Blackberry versus the iPhone.
link |
Why is it that you have a glass and you're using your finger
link |
for all of the work versus the buttons
link |
makes all the difference.
link |
This idea that now that you have a giant screen
link |
so that every part of the experience
link |
is now a digital experience.
link |
So you can have things like apps that change everything.
link |
You can't, when you first thinking about
link |
do I want a keyboard or not on a smartphone,
link |
you think it's just the keyboard decision.
link |
But then you later realize by removing the keyboard,
link |
you're enabling a whole ecosystem of technologies
link |
that are inside the phone.
link |
And now you're making the smartphone into a computer.
link |
And that same way,
link |
who knows how you can transform trucks, right?
link |
By like automating some parts of it,
link |
maybe adding some displays,
link |
maybe allows you to,
link |
maybe giving the truck driver some control
link |
in the supply chain to make decisions
link |
all those kinds of things.
link |
So where are you on the spectrum of hope
link |
for the role of automation in trucking?
link |
I think automation is inevitable.
link |
And again, I think this is really going to be transformative
link |
and it's gonna be,
link |
I've studied the history of trucking technology
link |
There's not a lot of great stuff written
link |
and you kind of have to,
link |
there isn't a lot of data and places to know
link |
sort of volumes of stuff and how they're changing, et cetera.
link |
But the big revolutionary changes in trucking
link |
are because of constellations of factors.
link |
It's not just one thing, right?
link |
So Daimler builds a motorized truck
link |
and I think it's 1896, right?
link |
Intercity trucking.
link |
So basically what they use that truck for
link |
is just to swap out horses, right?
link |
They basically do the same thing.
link |
The service doesn't really change, you know?
link |
And then World War I really spurs the development
link |
of sort of bigger, larger trucks,
link |
like spreads air filled tires.
link |
And then we start paving roads, right?
link |
And paved roads, right?
link |
Air filled tires and the internal combustion engine.
link |
Now you got a winning mix.
link |
Now it met with demand for people who wanted to get out
link |
from under the thumb of the railroads, right?
link |
So there was all of this pent up demand
link |
to get cheaper freight from the countryside
link |
into cities and between cities
link |
that typically had to go by rail.
link |
And so now, you know, 40 years
link |
after that internal combustion engine,
link |
it becomes this absolutely essential, right?
link |
This necessary but not sufficient piece of technology
link |
to create the modern trucking industry in the 1930s.
link |
And I think self driving is gonna be,
link |
self driving trucks are gonna be part of that.
link |
And the idea, I guess we credit Jeff Bezos.
link |
The idea is, you know, okay, so Sam Walton,
link |
if we can do it like a slight tangent
link |
on sort of the importance of trucking to business strategy
link |
and sort of how it has transformed our world.
link |
The central insight that Sam Walton had
link |
that made him the giant that he was
link |
in influencing the way that so many people get stuff
link |
was a trucking insight.
link |
And so if you look at the way that he developed his system,
link |
you build a distribution center
link |
and then you ring it with stores.
link |
Those stores are never further out
link |
from that distribution center
link |
than a human driven truck
link |
can drive back and forth in one day.
link |
And so rather than the way all of his competitors
link |
were doing it with sending trucks all over the place
link |
and having people sleep overnight
link |
and sort of making the trucking service fit
link |
where they had stores,
link |
he designed the layout of the stores
link |
to fit what trucks could do.
link |
And so transportation and logistics
link |
become Walmart's edge
link |
and allows them to dominate the space.
link |
That's the challenge that Amazon has now.
link |
They've mastered the digital part of it.
link |
And now they got to figure out
link |
how do we dominate the actual physical movement
link |
that complements that.
link |
Others are obviously gonna follow.
link |
But the capabilities of these trucks
link |
is completely different
link |
than the capability of a human driven truck.
link |
So if you're Smith packing
link |
and you've got a bunch of meat in a warehouse
link |
and it's going to grocery distribution centers,
link |
you have that trucker probably come in the night before
link |
and you make him wait
link |
so that he has a full 10 hour break,
link |
which is what the law requires
link |
so that he can get to the furthest reaches
link |
that he can of one of those stores.
link |
So he can drive his full 11 hours
link |
and bring that meat
link |
so it doesn't have to sit overnight
link |
in that refrigerated trailer.
link |
And so their system is based on that.
link |
Now, what happens when that truck
link |
can now travel two times as far, right?
link |
Three times as far.
link |
Now you don't need the warehouses where they were.
link |
Now you can go super lean with your inventory.
link |
Instead of having meat here, meat there, meat there,
link |
you can put it all right here.
link |
And if it's cheap enough,
link |
substitute those transportation costs
link |
for all that warehousing costs, right?
link |
So this is gonna remake landscapes
link |
in the same way that big box supply chains did, right?
link |
And then of course, the further compliment of that is,
link |
how do you then get it to two people at their door, right?
link |
And the big box supply chain,
link |
it moves very few items in really large quantities
link |
to very few locations pretty slowly, right?
link |
Ecommerce aspires to do something completely different,
link |
move huge varieties of things in small quantities,
link |
virtually everywhere as fast as possible, right?
link |
And so that is like that intercity trucking
link |
under the, in the era of railroad monopolies, right?
link |
The demand for that is potentially enormous, right?
link |
And so there's such a,
link |
so right now I think a lot of the business plans
link |
for sort of automated trucks, right?
link |
And sort of the way that the journalistic accounts portray it
link |
is like, okay, if we swap out a human for a computer,
link |
what are the labor costs per mile?
link |
And like, oh, here's the profitability
link |
of self driving trucks, uh uh.
link |
Like this is transformative technology.
link |
We're gonna change the way we get stuff.
link |
So we could actually get a lot more trucks period
link |
with like with autonomous trucks
link |
because they would enable a very different kind
link |
of transportation networks you think.
link |
Yeah, here's, and this is where it's like, uh oh.
link |
Like, yeah, so we really thought
link |
we were gonna be electrifying trucks.
link |
If they're going twice as far,
link |
if they're moving three times as much,
link |
if they're going three times as far, right?
link |
What does that mean for how far we are
link |
behind on batteries, right?
link |
We've got sort of these, you know, ideas about like, man,
link |
we, you know, here's how far,
link |
how close we could get to meet this demand.
link |
That demand is gonna radically change, right?
link |
These trucks are, you know, so then we've got to think
link |
about, all right, if it's not batteries, you know,
link |
how are we powering these things?
link |
And how many of them are they're gonna be?
link |
Like right now we've got 5 million containers
link |
that move from LA and Long Beach to Chicago on rail.
link |
Rail is three or four times at least more efficient
link |
than trucks in terms of greenhouse gas emissions.
link |
And on that lane, it varies a lot depending on demand,
link |
but maybe rail has a 20% advantage in cost, maybe 25%,
link |
but it's a couple of days slower.
link |
So now you cut the cost of that truck,
link |
transportation per mile by 30%.
link |
Now it's cheaper than rail and it gets the stuff there
link |
five days faster than rail.
link |
How many millions of containers are gonna leave LA
link |
and Long Beach on self driving trucks and go to Chicago?
link |
And it might look very much like a train
link |
if we go with the platooning solution,
link |
where you have these rows of like,
link |
imagine like rows of like 10, like dozens of trucks
link |
or like hundreds of trucks, like some absurd situation.
link |
Just going from LA to Chicago, just this train,
link |
but taking up a highway.
link |
I mean, this is probably a good place
link |
to talk about various scenarios.
link |
Well, before we get there,
link |
can I just make one interesting observation
link |
that I made as a driver?
link |
When you're in a truck, you're up higher.
link |
So you can see further and you can see the traffic patterns
link |
and cars move in packs.
link |
I'm sure there's academic research on this, right?
link |
But they move in packs.
link |
They kind of bunch up behind a slower car
link |
and then a bunch of them break free.
link |
And this is sort of on almost free flowing highways.
link |
They kind of move in packs
link |
and you can kind of see them in the truck.
link |
So, rather than platoons,
link |
we might have like hives of trucks, right?
link |
So you have like 20 trucks moving
link |
in some coordinated fashion, right?
link |
And then maybe the self driving cars are,
link |
cause people don't like to be around them
link |
or whatever it is, right?
link |
You might have a pod of 20 self driving cars
link |
sort of moving in a packet behind them.
link |
This is what, if the aliens came down
link |
or we're just observing cars,
link |
which is one of the sort of prevalent characteristics
link |
of human civilization is there seems to be these cars
link |
like moving around that would do this kind of analysis
link |
of like, huh, what's the interesting clustering
link |
of situations here,
link |
especially with autonomous vehicles, I like this.
link |
Okay, so what technologically speaking
link |
do you see are the different scenarios
link |
of increasing automation in trucks?
link |
What are some ideas that you think about?
link |
For the most part, I have no influence
link |
on sort of what these ideas were.
link |
So what the project was that I did was I said,
link |
technology is created by people, they solve for X
link |
and they have some conception of what they wanna do.
link |
And that's where we should start in sort of thinking
link |
about what the impacts might be.
link |
So I went and I talked to everybody I could find
link |
who was thinking about developing a self driving truck.
link |
And the question was essentially,
link |
what are you trying to build?
link |
Like, what do you envision this thing doing?
link |
It turned out that that for a lot of them
link |
was an afterthought.
link |
They knew the sort of technological capabilities
link |
that a self driving vehicle would have.
link |
And those were the problems that they were tackling.
link |
They were engineers and computer scientists and...
link |
Oh, robotics people, I love you so much.
link |
This is, I could talk forever about this,
link |
but yes, there's a technology problem,
link |
let's focus on that and we'll figure out
link |
the actual impact on society,
link |
how it's actually going to be applied,
link |
how it's actually going to be integrated
link |
from a policy and from a human perspective,
link |
from a business perspective later.
link |
First, let's solve the technology problem.
link |
That's not how life works, friends, but okay, I'm sorry.
link |
Yeah, yeah, so I mean,
link |
I'm sure you know the division of labor
link |
in these companies, right?
link |
There's sort of a business development side,
link |
you know, and then there's the engineering side, right?
link |
And the engineers are like, oh my God,
link |
what are these business development people, you know,
link |
why are they involved in this process?
link |
So I ended up sort of coming up with a few different ideas
link |
that people seem to be batting around
link |
and then really tried to zero in
link |
on a layman's understanding of the limitations, right?
link |
And it turns out that's really obvious and quite simple.
link |
Highway driving's a lot simpler, right?
link |
So, you know, the plan is simplify the problem, right?
link |
And focus on highways because city driving
link |
is so much more complicated.
link |
So from that, I came up with basically six scenarios,
link |
actually I came up with five
link |
that the developers were talking about.
link |
And then one that I thought was a good idea
link |
that I had read about, I think in like 2013 or 2014,
link |
which was actually something
link |
that the US military was looking at.
link |
I actually first heard about the idea
link |
of this kind of automation, at least in sketched out form
link |
in like 2011, I guess it was with Peloton,
link |
which was this sort of early technology entrant
link |
into the trucking industry,
link |
which was working on platooning trucks.
link |
And all they were doing was, you know,
link |
a cooperative adaptive cruise control
link |
as they came to call it.
link |
And we ended up on a panel together.
link |
And it's kind of interesting because I was on that panel
link |
because I was thinking about how we got the best return
link |
on investment for fuel efficient technologies.
link |
And if it's cool, I'll sort of set this up
link |
because it does, it comes into sort of the story
link |
of some of these scenarios.
link |
So when I studied the drivers,
link |
you had this like complete difference in the driving tasks,
link |
like we were talking about before
link |
with long haul and city, right?
link |
And you're not paid in the city,
link |
you've got congestion, the turns are tight.
link |
There's lots of, you know, pedestrians, you know,
link |
all the things that self driving trucks don't like,
link |
truckers don't like, right?
link |
And they're not paid, there's lots of waiting time.
link |
And then in the highway, they get to cruise,
link |
they're getting paid, they have control,
link |
they go at their own pace,
link |
they're making money, they're happy.
link |
Well, it turned out, I guess it was around 2010,
link |
this is still when we were thinking
link |
about regenerative braking, you know,
link |
and hybrid trucks being sort of like the solution.
link |
The problems with them sort of,
link |
and the advantages, you know,
link |
also split on what I was thinking of
link |
as kind of the rural urban divide at that time, right?
link |
So, like you got the regenerative braking, right?
link |
You can make the truck lighter,
link |
you can keep it local, right?
link |
You don't get any benefit from that, you know,
link |
hybrid electric on the rural highway,
link |
you want aerodynamics, right?
link |
There, you want low rolling resistance tires
link |
and these super aerodynamic sleek trucks, right?
link |
Where we know with off the shelf technology today,
link |
we could double the fuel economy,
link |
more than double the fuel economy of the typical truck
link |
in that highway segment,
link |
if we segmented the duty cycle, right?
link |
And so in the urban environment,
link |
you want a clean burning truck,
link |
so you're not giving kids asthma,
link |
you want it lighter,
link |
so it's not destroying those less strong pavements, right?
link |
You're not, you can make tighter turns,
link |
you don't need a sleeper cab,
link |
because the driver, you know,
link |
hopefully is getting home at night, right?
link |
In the long haul, you want that super aerodynamic stuff.
link |
Now that doesn't get you anything in the city,
link |
and in fact, it causes all kinds of problems,
link |
because you turn too tight,
link |
you crunch up all the aerodynamics
link |
that connect the tractor and the trailer.
link |
So the idea that I had was like, okay,
link |
what if we deliberately segmented it?
link |
Like, what if we created these droplets outside cities,
link |
where, you know, a local city driver who's paid by the hour
link |
kind of runs these trailers out once they're loaded,
link |
you know, doesn't sit there and wait while it's being loaded,
link |
they drop off a trailer, they go pick up one that's loaded,
link |
they run it out, when it's loaded, they call them,
link |
and they just run them out there and stage them.
link |
It's like an Uber driver, but for truckloads.
link |
Yeah, and we have like intermodal.
link |
We have like, basically this would be the equivalent
link |
of like rail to truck intermodal, right?
link |
So you put it on the rail, and then, you know,
link |
a trucker picks it up and delivers it, right?
link |
So instead of having the rail,
link |
you'd have these super aerodynamic, hopefully platoons,
link |
or what at the time was called long combination vehicles,
link |
which is basically two trailers connected together, right?
link |
Because this is like a huge productivity gain, right?
link |
And then instead of that driver like me,
link |
I would pick up something in upstate New York,
link |
drive to Michigan, drive to Alabama, you know,
link |
drive to Wisconsin, drive to Florida, you know,
link |
I'd get home every two weeks.
link |
If I'm just running that, you know, that double trailer,
link |
I might be able to go back and forth
link |
from Chicago to Detroit, right?
link |
Take two trailers there, pick up two trailers going back,
link |
right, and be home every night.
link |
Now, the problem with that at the time,
link |
or one of them was, you know, bridge weights.
link |
So you can't, not all bridges can handle
link |
that much weight on them.
link |
They can't handle these doubles, right?
link |
And some places can, some places can't.
link |
And so this platooning idea was happening at the same time,
link |
and we ended up on the same panel,
link |
and the founders were like,
link |
hey, so what's it like to follow
link |
really close behind another truck?
link |
Which was kind of the stage that they were at was like,
link |
you know, what's that experience gonna be like?
link |
And I was like, truckers aren't gonna like it, you know?
link |
I mean, that's just like the cardinal rule
link |
is following distance.
link |
Like that's the one you really shouldn't violate, right?
link |
And when you're out on the road,
link |
like you have that trucker like right on your ass,
link |
you know, people remember that.
link |
They don't remember the 99.9% of truckers
link |
that are not on their ass, you know?
link |
Like they're very careful about that.
link |
But when the trucks are really close together,
link |
there's benefits in terms of aerodynamics.
link |
So that's the idea.
link |
So like if you want to get some benefits of a platoon,
link |
you want them to be close together,
link |
but you're saying that's very uncomfortable for truckers.
link |
Yeah, so I mean, I think that ended up at the,
link |
I mean, Peloton I think is sort of winding down
link |
their work on this.
link |
And I think that ended up being still an open question.
link |
Like, and I had a chance to interview a couple drivers
link |
who at least one, maybe two of which
link |
had actually driven in their platoons.
link |
And I got completely different experiences.
link |
Some of them were like, it's really cool.
link |
You know, I'm like in communication with that other driver.
link |
You know, I can see on a screen what's out,
link |
you know, the front of his truck.
link |
And then some were like, it's too close.
link |
And it might be one of those things that's just,
link |
you know, it takes an adjustment to sort of get there.
link |
So you get the aerodynamic advantage,
link |
which, you know, saves fuel.
link |
There's some problems though, right?
link |
So, you know, you're getting that aerodynamic advantage
link |
because there's a low pressure system
link |
in front of that following truck.
link |
But the engine is designed with higher pressure
link |
feeding that engine, right?
link |
So there are sort of adjustments that you need to make
link |
and, you know, still the benefits are there.
link |
That's one scenario.
link |
And that's just the automation
link |
of that acceleration and braking.
link |
Starsky, which, you know,
link |
probably a lot of your listeners heard about,
link |
was working on another scenario,
link |
which was, you know, to solve that local problem
link |
was gonna do teleoperation, right?
link |
Sort of remote piloting.
link |
I had the chance to, you know,
link |
sort of watch them do that.
link |
You know, they drove a truck in Florida
link |
from San Francisco in one of their offices.
link |
That was really interesting.
link |
In case it's not clear,
link |
teleoperation means you're controlling the truck remotely,
link |
like it's a video game.
link |
So you've gotten the chance to witness it.
link |
Does it actually work?
link |
Yeah, I mean, so it's a...
link |
What are the pros and cons?
link |
You know, one of the problems with doing research like this
link |
with all these Silicon Valley folks is the NDAs.
link |
So, you know, I don't know what I'm able to say
link |
about sort of watching it,
link |
but obviously they're public statements
link |
about sort of what the challenges are, right?
link |
And it's about the latency
link |
and the ability to sort of in real time.
link |
There's challenges there.
link |
Let me say one thing.
link |
So I'm talking to the...
link |
You know, I've talked to the Waymo CTO.
link |
I'm in conversations with them.
link |
I'm talking to the head of trucking, Boris Softman,
link |
in next month, actually.
link |
I'm a huge fan of his because he was,
link |
I think the founder of Anki,
link |
which is a toy robotics company.
link |
So I love human robot interaction.
link |
And he created one of the most effective
link |
and beautiful toy robots.
link |
Anyway, I keep complaining to them on email privately
link |
that there's way too much marketing in these conversations
link |
and not enough showing off both the challenge
link |
and the beauty of the engineering efforts.
link |
And that seems to be the case
link |
for a lot of these Silicon Valley tech companies.
link |
They put up this, you're talking about NDAs.
link |
For some reason, rightfully or wrongfully,
link |
because there's been so much hype
link |
and so much money being made,
link |
they don't see the upside in being transparent
link |
and educating the public about how difficult the problem is.
link |
It's much more effective for them to say,
link |
we have everything solved.
link |
This will change everything.
link |
This will change society as we know it.
link |
And just kind of wave their hands
link |
as opposed to exploring together
link |
like these different scenarios.
link |
What are the pros and cons?
link |
Why is it really difficult?
link |
You know, what are the gray areas
link |
of where it works and doesn't?
link |
What's the role of the human in this picture
link |
of the both sort of the operators
link |
and the other humans on the road?
link |
All of that, which are fascinating human problems,
link |
fascinating engineering problems
link |
that I wish we could have a conversation about
link |
as opposed to always feeling like it's just marketing talk.
link |
Because a lot of what we're talking about now,
link |
even you with having private conversations under NDA,
link |
you still don't have the full picture of everything,
link |
of how difficult this problem is.
link |
One of the big questions I've had,
link |
still have is how difficult is driving?
link |
I disagree with Elon Musk and Jim Keller on this point.
link |
I have a sense that driving is really difficult.
link |
You know, the task of driving, just broadly.
link |
This is like philosophy talk.
link |
How much intelligence is required to drive a car?
link |
So from like a Jim Keller,
link |
who used to be the head of autopilot,
link |
the idea is that it's just a collision avoidance problem.
link |
It's like billiard balls.
link |
It's like you have to convert the drive.
link |
You have to do some basic perception,
link |
a computer vision to convert driving into a game of pool.
link |
And then you just have to get everything into a pocket.
link |
To me, there just seems to be some game theoretic dance
link |
combined with the fact that people's life is at stake.
link |
And then when people die at the hands of a robot,
link |
the reaction is going to be much more complicated.
link |
So all of that, but that's still an open question.
link |
And the cool thing is all of these companies
link |
are struggling with this question
link |
of how difficult is it to solve this problem sufficiently
link |
such that we can build a business on top of it
link |
and have a product that's going to make
link |
a huge amount of money
link |
and compete with the manually driven vehicles.
link |
And so their teleoperation from Starsky's
link |
is really interesting idea.
link |
How much can, I mean,
link |
there's a few autonomous vehicle companies
link |
that tried to integrate teleoperation in the picture.
link |
Can we reduce some of the costs
link |
while still having reliability,
link |
like catch when the vehicle fails
link |
by having teleoperation?
link |
It's an open question.
link |
So that's for you scenario number two
link |
is to use teleoperation as part of the picture.
link |
Yeah, let me follow up on that question
link |
of how hard driving is,
link |
because this becomes a big question for researchers
link |
who are thinking about labor market impacts,
link |
because we start from a perspective
link |
of what's hard or easy for humans.
link |
And so if you were to look at truck driving prior to a lot,
link |
I mean, there's been a lot of thinking and debate
link |
in academic research circles
link |
around sort of how you estimate labor impacts,
link |
what these models look like.
link |
And a lot of it is about how automatable is a job,
link |
object recognition, really easy for people, right?
link |
Really hard for computers.
link |
And so there's a whole bunch of things
link |
that truck drivers do that we see as super easy
link |
and as it would have been characterized 10 years ago,
link |
routine, and it's not for a computer, right?
link |
It turns out to be something that we do naturally
link |
that is sort of cutting edge, right?
link |
So on the teleoperation question,
link |
I think this is a more interesting one
link |
than people would like to sort of let on, I think, publicly.
link |
There are gonna be problems, right?
link |
And this is one of the complexities
link |
of sort of putting these things out in the world.
link |
And if you see the real world of trucking,
link |
you realize, wow, it's rough.
link |
There are dirt lots, there's gravel,
link |
there's salt and ice and cold weather,
link |
and there's equipment that just gets left out
link |
in the middle of nowhere,
link |
and the brakes don't get maintained,
link |
and somebody was supposed to service something
link |
and they didn't, you know?
link |
And so you imagine, okay, we've got this vehicle
link |
that can drive itself,
link |
which is gonna require a whole lot of sensors
link |
to tell it that the doors are still closed
link |
and the trailer's still hooked up
link |
and each of the tires has adequate pressure,
link |
and any number of, probably hundreds of sensors
link |
that are gonna be sort of relaying information.
link |
And one of them, after 500,000 miles or whatever,
link |
Now, do we have some fleet of technicians
link |
sort of continually cruising the highways
link |
and sort of servicing these things as they do what?
link |
Pull themselves off to the side of the road
link |
and say, I've got a sensor fault, I'm pulling over,
link |
or maybe there's some level of safety critical faults
link |
or whatever it might be.
link |
So that suggests that there might be a role
link |
for teleoperation even with self driving.
link |
And when I push people on it in the conversations,
link |
they all are like, yeah, we kind of have that
link |
on the bottom of the list,
link |
figure out how to rescue truck, you know?
link |
I guess on the to do list, right?
link |
After solving the self driving question is like,
link |
yeah, what do we do with the problems, right?
link |
I mean, no, we could imagine like, all right,
link |
we have some protocol that the truck is not,
link |
realizes the system says not safe for operation,
link |
Good, you have a crash, but now you got a truck stranded
link |
on the side of the road.
link |
You're gonna send out somebody to like calibrate things
link |
and check out what's going on,
link |
or that sounds like expensive labor,
link |
it sounds like downtime, it sounds like the kind of things
link |
that shippers don't like to happen to their freight,
link |
you know, in a just in time world.
link |
And so wouldn't it be great if you could just sort of,
link |
you know, loop your way into the controls of that truck
link |
and say, all right, we've got a sensor out,
link |
says that the tire is bad,
link |
but I can see visually from the camera, looks fine,
link |
I'm gonna drive it to our next depot,
link |
you know, maybe the next rider or Penske location, right?
link |
Sort of all these service locations around
link |
and have a technician take a look at it.
link |
So teleoperation often gets this, you know,
link |
so dismissive, you know, commentary from other folks
link |
working on other scenarios.
link |
But I think it's potentially more relevant
link |
than we hear publicly.
link |
But it's a hard problem.
link |
And, you know, for me, I've gotten a chance
link |
to interact with people that take on hard problems
link |
and solve them and they're rare.
link |
What Tesla has done with their data engine.
link |
So I thought autonomous driving cannot be solved
link |
without collecting a huge amount of data
link |
and organizing it well,
link |
not just collecting, but organizing it.
link |
And exactly what Tesla is doing now
link |
is what I thought it would be,
link |
like I couldn't see car companies doing that,
link |
And now that they're doing that, it's like, oh, okay.
link |
So it's possible to take on this huge effort seriously.
link |
To me, teleoperation is another huge effort like that.
link |
It's like taking seriously what happens when it fails.
link |
What's the, in the case of Waymo for the consumer,
link |
like ride sharing, what's the customer experience like?
link |
There's a bunch of videos online now
link |
where people are like the car fails and it pulls off
link |
to the side and you call like customer service
link |
and you're basically sitting there for a long time
link |
and there's confusion.
link |
And then there's a rescue that comes
link |
and they start to drive.
link |
I mean, just the whole experience is a mess
link |
that has a ripple effect to how you trust
link |
in the entirety of the experience.
link |
But like actually taking on the problem
link |
of that failure case and revolutionizing that experience,
link |
both for trucking and for ride sharing,
link |
that's an amazing opportunity there
link |
because that feels like it would change everything.
link |
If you can reliably know when the failures happen,
link |
which they will, you have a clear plan
link |
that doesn't significantly affect the efficiency
link |
of the whole process, that could be the game changer.
link |
And if teleoperation is part of that,
link |
it could be just like you're saying,
link |
it could be teleoperation or it could be like a fleet
link |
of rescuers that can come in, which is a similar idea.
link |
But teleoperation, obviously that allows you
link |
to just have a network of monitors
link |
of people monitoring this giant fleet of trucks
link |
and taking over when needed.
link |
And it's a beautiful vision of the future
link |
where there's millions of robots
link |
and only thousands of humans monitoring
link |
those millions of robots.
link |
That seems like a perfect dance
link |
of allowing humans to do what they do best
link |
and allowing robots to do what they do best.
link |
Yeah, yeah, so I mean, I think there are,
link |
and we just applied for an NSF we didn't get,
link |
anybody's watching, but with some folks from Wisconsin
link |
who do teleoperation, right?
link |
And some of this is used for like rovers
link |
and I mean, really high stakes, difficult problems.
link |
But one of the things we wanted to study
link |
were these mines, these Rio Tinto mines in Australia
link |
where they remotely pilot the trucks.
link |
And there's some autonomy, I guess,
link |
but it's overseen by a remote operator
link |
and it's near Perth and it's quite remote
link |
and they retrained the truck drivers
link |
to be the remote operators, right?
link |
There's autonomy in the port of Rotterdam
link |
and places like that where there are jobs there.
link |
And so I think, and maybe we'll get to this later,
link |
but there's a real policy question
link |
about sort of who's gonna lose and what we do about it
link |
and whether or not there are opportunities there
link |
that maybe we need to put our thumb on the scale
link |
a little bit to make sure that there's some give back
link |
to the community that's taking the hit.
link |
So for instance, if there were teleoperation centers,
link |
maybe they go in these communities
link |
that we disproportionately source truck drivers from today.
link |
Now, I mean, what does that mean?
link |
It may not be the cheapest place to do it
link |
if they don't have great connectivity
link |
and it may not be where the upper level managers wanna be
link |
and places like that, issues like that, right?
link |
So I do think it's an interesting question,
link |
both from sort of a practical scenario situation
link |
of how it's gonna work, but also from a policy perspective.
link |
So there's platoons, there's teleoperation,
link |
and this is taking care of some of the highway driving
link |
that we're talking about.
link |
Is there other ideas like,
link |
is there other ideas, scenarios
link |
that you have for autonomous trucks?
link |
Yeah, so I mean, the most obvious one actually
link |
is just facility to facility, right?
link |
The sort of, it can't go everywhere,
link |
but a lot of logistics facilities
link |
are very close to interstates
link |
and they're on big commercial roads
link |
without bikes and parked cars and all that stuff.
link |
And some of the jobs that I think are really first
link |
on the chopping block are these LTL,
link |
that less than truckload, what's called line haul, right?
link |
So these are the drivers who go from terminal to terminal
link |
with those full trailers.
link |
And those facilities are often located strategically
link |
to avoid congestion, right?
link |
And to be in big industrial facilities.
link |
So you could imagine that being the first place
link |
you see a Waymo self driving truck rollout
link |
might be sort of direct facility to facility
link |
for UPS or FedEx or less than truckload care.
link |
And the idea there is fully driverless,
link |
so potentially not even a driver in the truck,
link |
it's just going from facility to facility empty,
link |
Yeah, and those, because that labor is expensive,
link |
they don't keep those drivers out overnight,
link |
those drivers do a run back and forth typically,
link |
or in a team go back and forth in one day.
link |
So from the people you've spoken with so far,
link |
what's your sense?
link |
How far are we away from, which scenario is closest
link |
and how far away are we from that scenario
link |
of autonomy being a big part of our trucking fleet?
link |
Most folks are focused on another scenario,
link |
which is the exit to exit, right?
link |
Which looks like that urban truck ports thing
link |
that I laid out earlier.
link |
So you have a human driven truck
link |
that comes out to a drop lot,
link |
it meets up with an autonomous truck,
link |
that truck then drives it on the interstate to another lot,
link |
and then a human driver picks it up.
link |
There are a couple variations maybe on that.
link |
So, or let me just run through the last two scenarios.
link |
The other thing you could do, right,
link |
is to say, all right, I've got a truck that can drive itself,
link |
and I refer to this one as autopilot,
link |
but you have a human drive it out to the interstate,
link |
but rather than have that transaction
link |
where the human driven truck detaches the trailer
link |
and it gets coupled up to a self driving truck,
link |
they just, that human driver just hops on the interstate
link |
with that truck and goes in back and goes off duty
link |
while the truck drives itself.
link |
And so you have a self driving truck
link |
that's not driverless, right?
link |
And just to clarify,
link |
because Tesla uses the term autopilot instead of airplanes,
link |
and so everybody uses the word autopilot,
link |
we're referring to essentially full autonomy,
link |
but because it's exit to exit,
link |
the truck driver is onboard the truck,
link |
but they're sleeping in the back or whatever.
link |
Yeah, and this gets to the really weedy policy questions,
link |
So basically for the Department of Transportation,
link |
for you to be off duty for safety reasons,
link |
you have to be completely relieved of all responsibility.
link |
So that truck has to not encounter a construction site
link |
or inclement weather or whatever it might be,
link |
and call to you and say, hey, you know,
link |
or I mean, obviously, right,
link |
we're imagining connected vehicles as well, right?
link |
So you're in a self driving truck,
link |
you're in the back and trucks 20 miles ahead
link |
experience some problem, right?
link |
That may require teleoperation or whatever it is, right?
link |
And it signals to your truck,
link |
hey, you know, tell the driver 20 miles ahead,
link |
he's got to hop in the seat.
link |
That would mean that they're on duty
link |
according to the way that the current rules are written,
link |
they have some responsibility.
link |
And part of that is, you know,
link |
we need them to get rest, right?
link |
They need to have uninterrupted sleep.
link |
So that's what I call autopilot.
link |
The final scenario is one that I thought was actually
link |
the one scenario that was good for labor, you know,
link |
which I proposed is I was like, well, here's an idea,
link |
you know, that would be like, actually good for workers.
link |
And just another brief aside here.
link |
The history of trucking over the last, you know, 40 years,
link |
there's been a lot of technological change.
link |
So when I learned to drive the truck,
link |
I had to learn to manually shift it like I was describing,
link |
you had to read these fairly complicated, you know,
link |
big sets of laminated maps to figure out
link |
where the truck can go and where it couldn't,
link |
which is a big deal, you know,
link |
I mean, you take these trucks on the wrong road
link |
and you're destroying a bridge
link |
or you're doing a can opener,
link |
which is where you try to drive it under a bridge too low.
link |
You've probably seen that on YouTube,
link |
if not, you know, check it out, you know, truck can opener.
link |
You know, there's some bridges that are famous for it,
link |
right, and there's one I think called the can opener
link |
and you can find on YouTube.
link |
And, you know, you had to log those hours like manually
link |
and sort of do the math and plan your work routine.
link |
And I would do this every day.
link |
I'd say like, okay, I'm gonna get up at five.
link |
I've got to think about Buffalo and there's traffic there.
link |
So I wanna be through Buffalo by 6.30, you know,
link |
and then that'll put me, you know, in Cleveland at,
link |
you know, 9.30, which means I'll miss that rush hour, right,
link |
which is gonna put me in Chicago, you know,
link |
and so you do this and now today, you know,
link |
15 years later, truck drivers don't have to do any of that.
link |
You know, you don't have to shift the truck,
link |
you don't have to map, you know,
link |
you can figure out the least congested route
link |
to go on and your hours of service are recorded
link |
or a good portion of them are reported automatically.
link |
All of that has been a substantial de skilling
link |
that has, you know, put downward pressure on wages
link |
and allowed companies to kind of speed up, monitor
link |
and direct, I mean, the key technology, you know,
link |
that I did work under is satellite linked computers.
link |
So before you could kind of go out and plan your own work
link |
and the boss really couldn't see what you were doing
link |
and push you and say, you know, you've been on break
link |
for 10 hours, why aren't you moving?
link |
You know, and you might tell them, you know,
link |
cause I'm tired, you know, like I didn't sleep well,
link |
I've got to get a couple more hours, you know,
link |
they're only gonna accept that so many times
link |
or at least some of those dispatchers are.
link |
So all this technology has made the job sort of, you know,
link |
de skilled the job, you know, hurt drivers
link |
in the labor market, made the work worse.
link |
So I think the burden it's really on the technologists
link |
who are like, oh, this will make truck driver jobs better
link |
and sort of envision ways that it would.
link |
It's like, the burden's really, a proof is really on you
link |
to sort of really clearly lay out what that
link |
is gonna look like because it's 30 or 40 years of history
link |
suggests that that technology into labor markets
link |
where workers are really weak and cheap is what wins
link |
that new technology doesn't help workers
link |
or raise their wages.
link |
So it lowers the bar of entry in terms of skill.
link |
So that's really, that's tough.
link |
That's tough to know what to do with because yeah,
link |
from a technology perspective, you wanna make the life
link |
of the people doing the job today easier.
link |
Is that what you want?
link |
No, but that like, when you think about like what exactly,
link |
because the reality is you will make their life
link |
potentially a little bit easier,
link |
but that will allow the companies to then hire people
link |
that are less skilled, get those people
link |
that were previously working there fired or lower wages.
link |
And so the result of this easier is a lower quality of life.
link |
That's dark actually.
link |
I know, I'm sorry.
link |
But you were saying that was for you initially the hopeful.
link |
Oh no, so I'll get to that.
link |
But one more thing, cause this is not stopping.
link |
And this is another interesting question
link |
about the sort of automation.
link |
And I think Uber is an interesting example here
link |
where it's like, okay, if we had self driving trucks
link |
or self driving cars, we could automate
link |
what used to be taxi service.
link |
There's a whole bunch of stuff
link |
that's already been automated, like the dispatching.
link |
So the dispatchers are already out of work in rideshare
link |
and the payment is already automated.
link |
So you have to automate steps like this.
link |
So you have to have that initial link to dispatch the truck.
link |
You have to have the automated mapping.
link |
So we're sort of done all this incremental automation
link |
that could make the truck completely driverless.
link |
There's some important things happening right now
link |
with the remaining good jobs.
link |
So what you're really paying for
link |
when you get a good truck driver is, like I said,
link |
you get those kind of local skills
link |
of like backing and congested traffic.
link |
Those, it's really impressive to watch
link |
and there's some value on it certainly,
link |
but it's relatively low value
link |
in the actual driving technique, right?
link |
So you bump something backing into the dock,
link |
it might be a couple of thousand dollars
link |
because you ruin a canopy or something over a dock
link |
or tear up a trailer.
link |
What you really want,
link |
those highly skilled conscientious drivers,
link |
and that's really what it is.
link |
And that's what computers are really good at
link |
is about being conscientious, right?
link |
In the sense of like, they pay attention continually, right?
link |
And how I was describing those long haul segments
link |
where the driver just keeps out of the situations
link |
that could become problematic
link |
and just, they don't look at their phone.
link |
I mean, they take the job seriously and they're safe
link |
and you can give somebody a skills test, right?
link |
As a CDL examiner, you could take them out and say,
link |
all right, I need you to go around these cones
link |
and drive safely through this school zone.
link |
But what really proves that you're a safe driver
link |
is two years without an accident, right?
link |
Because that means that day after day,
link |
hour after hour, mile after mile,
link |
you did the right thing, right?
link |
And not when it was like, oh, some situations emerging,
link |
but just consistently over time
link |
kept yourself out of accident situations.
link |
And you can see this with drivers who are a million
link |
or 2 million safe miles.
link |
The value of those drivers for Walmart
link |
is they don't run over minivans.
link |
The company I worked for,
link |
they ran over minivans on a regular basis.
link |
So when I was trained, they said, we kill 20 people a year.
link |
We send someone to the funeral,
link |
there's a big check involved, don't be that.
link |
We don't wanna go to your funeral
link |
and you don't wanna be the person who caused that funeral.
link |
Okay, so they just write that off.
link |
Okay, that's just part of the business model.
link |
Now, forward collision avoidance
link |
can basically eliminate the vast majority
link |
of those accidents.
link |
That's what the value of a really expensive
link |
conscientious driver is based on.
link |
They don't run over minivans.
link |
So as soon as you have that forward collision avoidance,
link |
what's gonna happen to the wages of those drivers?
link |
By way of a therapy session, help me understand,
link |
is a collision avoidance,
link |
automated collision avoidance systems,
link |
are they good or bad for society?
link |
Yeah, I mean, this is, they're good.
link |
Right. They're good.
link |
But what do we do about the pain of a workforce
link |
in the short term because their wages are gonna go down
link |
because the job starts requiring less and less skill?
link |
Is there a hopeful message here
link |
where other jobs are created?
link |
So I'm a sociologist, right?
link |
So I'm gonna think about what's the structure behind that
link |
that creates that pain, right?
link |
And it's ownership, right?
link |
We don't call it capitalism for nothing.
link |
What capitalists do is they figure out cheaper,
link |
more efficient ways to do stuff.
link |
And they use technology to do that oftentimes, right?
link |
This is the remarkable history of the last couple centuries
link |
and all the productivity gains is,
link |
people who were in a competitive market saying,
link |
if I have to do it, right?
link |
I don't have a choice.
link |
Cause like my competitor over there is gonna eat my lunch
link |
if I'm not on my game.
link |
I don't have a choice.
link |
I've got to invest in this technology
link |
to make it more efficient, to make it cheaper.
link |
And what do you look for?
link |
You look for oftentimes, you look for labor costs, right?
link |
You look for high value labor.
link |
If I can take a hundred and,
link |
a lot of these truck drivers make good money,
link |
a hundred thousand dollars, good benefits,
link |
vacation, retirement.
link |
If I can replace them with a $35,000 worker
link |
when I'm competing with maybe a low wage retail employer
link |
rather than some other more expensive employers
link |
for skilled blue collar workers, I'm gonna do that.
link |
And that's just, that's what we do.
link |
And so I think those are the bigger questions
link |
around this technology, right?
link |
Is like, are workers gonna get screwed by this?
link |
Like, yeah, most likely.
link |
Like that's what we do.
link |
So one of the things you say is,
link |
I mean, first of all, the numbers of workers
link |
that will feel this pain is not perhaps as large
link |
as the journalists kind of articulate,
link |
but nevertheless, the pain is real.
link |
And I guess my question here is,
link |
do you have an optimistic vision
link |
about the transformative effects
link |
of autonomous trucks on society?
link |
Like if you look 20 years from now
link |
and perhaps see maybe 30 years from now,
link |
perhaps see these autonomous trucks
link |
doing the various parts of the scenarios you listed.
link |
And there's just hundreds of thousands of them,
link |
just like veins, like blood flowing through veins
link |
on the interstate system.
link |
What kind of world do you see that's a better world
link |
than today that involves such trucks?
link |
Yeah, can I defend myself first?
link |
Because I'm reading the comments right now
link |
of people, of the economists who are telling me.
link |
Another commenter, dear PhD in economics.
link |
Yes, yes, dear PhD in economics,
link |
I know that higher skilled jobs
link |
are created by technological advancement, right?
link |
I mean, there are big questions about how many of them,
link |
right, so the idea that we would create
link |
more expensive labor positions, right,
link |
with a new technology, right?
link |
You better check your business plan
link |
if your idea is to take a bunch of low wage labor
link |
and replace it with the same amount of high wage labor,
link |
right, so there's a question about how many of those jobs.
link |
And there's the really important social
link |
and political question of are they the same people, right?
link |
And do they live in the same places?
link |
And I think that kind of geography
link |
is a huge issue here with the impacts, right?
link |
Lots of rural workers.
link |
Interesting politically, lots of red state workers, right?
link |
Lots of blue state, maybe union folks
link |
who are gonna try to slow autonomy
link |
and lots of red state representatives in the house maybe
link |
who wanna stand up for their trucker constituents.
link |
So just to defend myself.
link |
Yeah, and to elaborate, I think economics as a field
link |
is not good at measuring the landscape
link |
of human pain and suffering.
link |
So sometimes you can forget in the numbers
link |
that it's real lives that are at stake.
link |
That's what I suppose sociology is better at doing.
link |
So we try sometimes, sometimes.
link |
Well, the problem with, I mean,
link |
I'm somebody who loves psychology and psychiatry
link |
and a little bit, I guess, of sociology.
link |
I realize how little, how tragically flawed the field is,
link |
not because of lack of trying,
link |
but just how difficult the problems are.
link |
To do really thorough studies
link |
that understand the fundamentals of human behavior
link |
and this, yes, landscape of human suffering,
link |
it's almost an impossible task without the data.
link |
And we currently don't, not everybody's richly integrated
link |
to where they're fully connected
link |
and all their information is being recorded
link |
for sociologists to study.
link |
So you have to make a lot of inferences.
link |
You have to talk to people.
link |
You have to do the interviews as you're doing.
link |
And through that really difficult work,
link |
try to understand, hear the music
link |
that nobody else is hearing,
link |
the music of what people are feeling,
link |
their hopes, their dreams, and the crushing of their dreams
link |
due to some kind of economic forces.
link |
Yeah, I mean, we've just lived that
link |
for four and a half years of probably elites,
link |
let me just go out on a limb and say,
link |
not understanding the sort of emotional
link |
and psychological currents
link |
of a large portion of the population, right?
link |
And just being stunned by it and confused, right?
link |
It wasn't confusing for me after having talked to truckers.
link |
Again, trucking is a job of last resort.
link |
These are people who've already lost
link |
that manufacturing job oftentimes,
link |
already lost that construction job to just aging, right?
link |
So what can we do, right?
link |
What's sort of the positive vision?
link |
Because like, we've got tons of highway deaths.
link |
We've got, and just the big picture is,
link |
and this is the opportunity, I guess, for investors,
link |
it's a hugely inefficient system.
link |
So we buy this truck,
link |
there's this low wage worker in it oftentimes.
link |
And again, I'm setting aside those really good
link |
line haul jobs in LTL, those are a different case.
link |
That low wage worker is driving a truck that they might,
link |
the wheels might roll seven to eight hours a day.
link |
That's what the truck is designed to do
link |
and that's what makes the money for the company.
link |
In other seven, eight hours a day,
link |
the driver's doing other kinds of work
link |
that is not driving.
link |
And then the rest of the day,
link |
they're basically living out of the truck.
link |
You really can't find a more inefficient use of an asset
link |
Now, a big part of that is we pay for the roads
link |
and we pay for the rest areas and all this other stuff.
link |
So the way that I work and the way that I think
link |
about these problems is I try to find analogies, right?
link |
Sort of labor processes and things that make economic sense
link |
that seem in the same area of the economy,
link |
but have some different characteristics for workers, right?
link |
And sort of try to figure out
link |
why does the economics work there, right?
link |
And so if you look at those really good jobs,
link |
the most likely way that you as a passenger car driver
link |
would know that it's one of those drivers
link |
is that there are multiple trailers, right?
link |
So you see these, like maybe it's three small trailers,
link |
maybe it's two sort of medium sized trailers.
link |
Some places you might even see
link |
two really big trailers together.
link |
You do that because labor is expensive, right?
link |
And it's highly skilled.
link |
And so you use it efficiently and you say, all right,
link |
rather than having you haul that little trailer
link |
out of the ports, that sort of half size container,
link |
we're gonna wait till we get three
link |
or we're gonna coordinate the movement
link |
so that they're three ready.
link |
You go do what truckers call make a set,
link |
put them together, right, and you go.
link |
That's a massive productivity gain, right?
link |
Because you're hauling two, three times as much freight.
link |
So the positive scenario that I threw out in 2018
link |
was why not have a human driven truck
link |
with a self driving truck that follows it, right?
link |
Just a drone unit.
link |
And to me, this seemed as a non computer scientist,
link |
a non computer scientist, a sociologist, right?
link |
This made a lot of sense because when I got done talking
link |
to the computer scientists and the engineers,
link |
they were like, well, it's like object recognition,
link |
decision making algorithm, all this stuff.
link |
It's like, all right, so why don't you leave
link |
the human brain in the lead vehicle, right?
link |
You got all that processing and then all that following.
link |
Now, again, this is sort of me being a lay person.
link |
I said, why don't, then that following truck, right,
link |
makes direction from the front.
link |
It uses the rear of the trailer as a reference point.
link |
It maintains the lane.
link |
You've got cooperative adaptive cruise control
link |
and that you double the productivity of that driver.
link |
You solve that problem that I hated
link |
in my urban truck ports thing about the bridge weight.
link |
Cause when you get to the bridges,
link |
the two trucks can just spread out just enough
link |
to make the bridge weight, right?
link |
And you can just program that in
link |
and they're 50 feet further apart,
link |
100 feet further apart.
link |
So interesting sort of, I think, story about this
link |
that leads to kind of, I think, the policy questions.
link |
In, I guess, 2017, Jack Reed and Susan Collins
link |
and requested from the Senate,
link |
the Senate requested research on what the impacts
link |
of self driving trucks would be.
link |
And the first stage of that was for the GAO
link |
to do a report, sort of looking at the lay of the land,
link |
talking to some experts.
link |
And I was working on my 2018 report,
link |
help contribute to that GAO report.
link |
And I had the six scenarios, right?
link |
I'm like, okay, here's what Starsky's doing.
link |
Here's what Embark and Uber are doing.
link |
Here's what Waymo might be doing.
link |
Nobody really knows, right?
link |
Here's what Peloton's doing.
link |
Here's the autopilot scenario.
link |
And then here's this one that I think
link |
actually could be good for drivers.
link |
So now you've got that driver who's got
link |
two times the freight.
link |
Their decisions are more important.
link |
They're managing a more complex system, right?
link |
They're probably gonna have to have
link |
some global understanding of how to,
link |
the environments in which it can operate safely, right?
link |
Now we're talking upskilling, right?
link |
And so the GAO sort of writes up these different scenarios
link |
and the idea is that it's gonna prepare
link |
for this Department of Transportation,
link |
Department of Labor set of processes
link |
to engage stakeholders and sort of get industry perspectives
link |
and then do a study on the labor impacts.
link |
So that DOT, DOL process starts to happen
link |
and I get to the workshop and a friend was sitting
link |
at the table next to me and he holds up the scenarios
link |
that they're gonna have us discuss at this workshop.
link |
And he's like, hey, these look really familiar, right?
link |
They were the scenarios from the report,
link |
but there were only five instead of six.
link |
The sixth scenario, which was the upskilling labor,
link |
good for workers scenario, wasn't discussed.
link |
So to clarify that the integral piece of technology
link |
there is platooning.
link |
Yeah, I mean, in a sense it's platooning,
link |
but, and in fairness, right, as I pitched that idea
link |
or sort of ran that idea by the computer scientists
link |
and engineers and product managers that I would talk to,
link |
they would say, we thought about that,
link |
but that following truck, it's not that simple.
link |
That thing, basically we had to engineer that
link |
to be capable of independent self driving,
link |
because what if there was a cut in
link |
or any number of scenarios in which it lost
link |
that connection to the lead truck for whatever reason.
link |
Now, I mean, I don't know.
link |
Boo hoo, platooning is hard.
link |
There's edge cases.
link |
I guarantee the number of edge cases in platooning
link |
is orders of magnitude lower than the number of edge cases
link |
in the general solo full self drive.
link |
You do not need to solve the full self driving problem.
link |
I mean, if you're talking about
link |
probability of dangerous events,
link |
it just seems with platooning,
link |
then like you can deal with cut ins.
link |
Yeah, so this is beyond,
link |
this is one of the challenge obviously of being a researcher
link |
who doesn't really have any background
link |
in the technology, right?
link |
So I can dream this up.
link |
I don't, you know, I have no idea if it's feasible.
link |
Well, let me speak, you spoke to the PhDs in economics.
link |
Let me speak to the PhDs in computer science.
link |
If you think platooning is as hard
link |
as the full self driving problem,
link |
we need to talk, because I think that's ridiculous.
link |
I think platooning, and in fact,
link |
I think platooning is an interesting idea
link |
for ride sharing as well,
link |
for the general autonomous driving problem,
link |
not just trucking, but obviously trucking
link |
is the big, big benefit,
link |
because the number of A to B points in trucking
link |
is much, much lower than the general ride sharing problem.
link |
But anyway, I think that's a great idea,
link |
but you're saying it was removed.
link |
Yeah, and so you can go, you know,
link |
and listeners could go to these reports.
link |
They're publicly available.
link |
And they explain why in the footnote.
link |
And they note that there was this other scenario
link |
suggested by at least me,
link |
and I can't remember if they said someone else did too.
link |
But they said, you know, we didn't include it
link |
because no developers were working on it.
link |
that was the approach that I took in my research, right?
link |
Which was to go to the developers and say,
link |
what's your vision, right?
link |
What are you trying to develop?
link |
That's what I was trying to do.
link |
And maybe, you know,
link |
and then I tried to think outside the box at the end
link |
by adding that one, right?
link |
Like, here's one that I have, you know,
link |
people aren't talking about that could be cool.
link |
Now, again, it had been proposed in like 2014
link |
for like fuel convoys.
link |
So you could just have like one super armored lead fuel
link |
In a, you know, bringing fuel to forward operating bases
link |
And then you wouldn't need, you know, the super heavy,
link |
you know, you wouldn't have to protect the human life
link |
in the following truck.
link |
So that's interesting.
link |
You're saying like, when you talk to Waymo,
link |
when you talk to these kinds of companies,
link |
they weren't at least openly saying they're working on this.
link |
So then it doesn't make sense to include in the list.
link |
And so, but here's the thing, right?
link |
This is the Department of Transportation, right?
link |
And the Department of Labor.
link |
Maybe they could consider some scenarios.
link |
Like maybe we could say, you know, this, we,
link |
this technology has got a lot of potential.
link |
Here's what we'd like it to do.
link |
You know, we'd like it to reduce highway deaths,
link |
help us fight climate change, reduce congestion,
link |
you know, all these other, other things.
link |
But that's not how our policy conversation
link |
around technology is happening.
link |
We're not, and people don't think that we should.
link |
And I think that's the fundamental shift
link |
that we need to have, right?
link |
I've been involved with this a little bit like NHTSA and DOT.
link |
The approach they took is saying,
link |
we don't know what the heck we're doing.
link |
So we're going to just let the innovators do their thing
link |
and not regulate it for a while, just to see.
link |
You don't, you think that's,
link |
you think DOT should provide ideas themselves.
link |
Well, so this is the, this is the great trick
link |
in policy of private actors,
link |
is you get narrow mandates for government agencies, right?
link |
So, you know, the safety case will be handled
link |
by organizations whose mandate is safety.
link |
So the Federal Motor Carrier Safety Administration,
link |
who is, you know, going to be a key player,
link |
I argue in an article that I wrote, you know,
link |
they're going to be a key player in actually determining
link |
which scenario is most profitable
link |
by setting the rules for truck drivers.
link |
Their mandate is safety, right?
link |
Now they have lots of good people there who want,
link |
you know, who care about truck drivers
link |
and who wish truck drivers jobs were better,
link |
but they don't have the authority to say,
link |
hey, we're going to write this rule
link |
because it's good for truck drivers, right?
link |
And so when you, you know, we need to say,
link |
you know, as a society, we need to not restrict technology,
link |
not stand in the way of things.
link |
We need to harness it towards the goals that matter, right?
link |
Not whatever comes out the end of the pipeline
link |
because it's the easiest thing to develop
link |
or whatever is most profitable for the first actor
link |
or whatever, but, you know, and we do,
link |
the thing is we do that, right?
link |
I mean, like when we sent people to the moon,
link |
you know, we did that,
link |
and there were tremendous benefits
link |
that followed from it, right?
link |
And we do this all the time in, you know,
link |
trying to cure cancer or whatever it is, right?
link |
I mean, we can do this, right?
link |
Now the interesting sort of epilogue to that story is,
link |
you know, six months or so, I don't know how long it was,
link |
after those meetings in which that sixth scenario
link |
was not considered, a company called Locomation,
link |
you know, ends up using that,
link |
essentially that basic scenario with a slight variation.
link |
So they leave the human driver in both trucks
link |
and then that following driver goes off duty.
link |
And then, you know, I've been trying to think
link |
of what the term is, they kind of,
link |
I think of it as like slingshotting,
link |
they sort of, when one runs out of hours,
link |
you know, the one who's off duty goes in front and,
link |
you know, and so, you know, if only they had been,
link |
you know, around six months earlier,
link |
that might've been considered by the OT,
link |
but it just says, you know, who has the authority
link |
to propose what these visions of the future are?
link |
Well, some of it is also just the company stepping up
link |
and just doing it, screw the authority,
link |
and showing that it's possible,
link |
and then the authority follows.
link |
So that's why I really love innovators in the space.
link |
The criticism I have, the very sort of real,
link |
I don't know, harsh criticism I have
link |
towards autonomous vehicle companies in the space
link |
is they've gotten culturally,
link |
they've, it's become acceptable somehow
link |
to do demos and videos,
link |
as opposed to the old school American way
link |
of solving problems.
link |
There's a culture in Silicon Valley
link |
where you're talking to VCs
link |
that have lost that kind of love of solving problems.
link |
They kind of like envision,
link |
if the story you told me in your PowerPoint presentation
link |
is true, how many trillions of dollars
link |
might I be able to make?
link |
There's something lost in that conversation
link |
where you're not really taking on like the problem
link |
in a real way, so these autonomous vehicle companies
link |
realize we don't need to,
link |
we just need to make nice PowerPoint presentations
link |
and not actually deliver products
link |
that like everybody looks outside and says,
link |
holy shit, this is life changing.
link |
This is where I have to give props to Waymo
link |
is they put driverless cars on the road
link |
and like forget PowerPoint slide presentations,
link |
actual cars on the road.
link |
Then you can criticize like,
link |
is that actually going to work?