Autonomy means a system can make its own decisions—like how to move and navigate—without someone constantly steering it. In self-driving discussions, it’s the big idea behind getting from research to real use.
Ekno Robotics is the company being discussed in the MIT interview. The point is to learn what they’re building and how it relates to real autonomous systems.
This is an interview tied to MIT and mobility, meaning transportation technology. It’s often about how self-driving ideas move from labs into real-world use.
Robotaxies are self-driving cars that you can call like a taxi. Instead of a human driver, the car handles the driving.
Concept
DARBA
This likely refers to DARPA, a U.S. research agency that funds big technology projects. In self-driving, DARPA has historically supported early breakthroughs in robotics and autonomy.
Chapter 11 is a legal process where a company tries to reorganize instead of shutting down right away. For railroads, that can mean big changes that affect service and jobs.
Conrail was a company created to take over struggling railroad lines in the Northeast. It helped stabilize rail service after several other railroads went bankrupt.
Waymo is a company that builds self-driving technology and runs robotaxi rides for the public. When they talk about ridership, they mean how many paid trips their cars are doing.
Driverless cars are cars that can drive themselves using technology and software. The point here is that if they work well, they can provide real transportation benefits to people.
Paid rides mean real customers are paying for the trip, not just testing. That usually suggests the service is working well enough to be offered commercially.
Here, productivity means how many trips each self-driving car is able to do in a day. More trips per car generally means the service is running efficiently.
Sometimes a car has to drive to a different spot before it can pick someone up. That “empty” driving doesn’t carry passengers, so it’s important for the system’s efficiency.
Parking is basically leaving a car sitting until you need it. The speaker is saying that’s what people do with their own cars, while shared/automated systems try to avoid waiting by moving vehicles around.
Uber and Lyft are apps that connect riders with drivers. The comparison here is that they can keep rides flowing because there’s a real driver ready to pick someone up.
V2X means cars can “talk” to other cars and to things around the road, like signals or infrastructure. That communication can help with safety and smoother traffic.
Connected vehicles are cars that can share information with other cars or with road systems. The hope is that traffic can run smoother and safer because everyone is “in the loop.”
A “commodity” is something where buyers see no big difference between brands. The discussion is saying self-driving shouldn’t become that—companies want it to feel unique.
Nissan is the car company being discussed. The point is that their self-driving strategy is supposed to be unique, not just a generic feature everyone has.
General Motors is the automaker being discussed. The idea is that they tested a lot of cars, but then changed course—likely to improve results or focus their testing.
Hands-free driving means the car can help steer for you, especially on highways. You still have to pay attention because the system only works in certain situations.
Subaru is the automaker being discussed, and they’re known for safety-focused driver-assist features. The episode suggests Subaru has been improving these systems for years.
EyeSight is Subaru’s safety system that uses cameras to watch what’s happening ahead. It can help the car slow down or avoid crashes when it detects danger.
Even if the vehicle drives itself, a person may still be responsible if something goes wrong. That’s why the driver might still have to watch and be ready to take over.
They’re talking about using the truck more efficiently—getting more hours of real work out of it. Automation could help the truck spend more time moving and less time waiting.
OSHA is a U.S. agency that focuses on workplace safety. The speaker is asking whether OSHA would need to review safety concerns if truck driving becomes more automated.
Modern AI is the kind of computer “learning” used in today’s apps and systems. Here, they’re recommending a book about AI, but then they explain it’s really focused on self-driving cars.
They’re saying the reason ride-sharing makes sense is basically money and efficiency. If more people share one trip, it can cost less and use less energy than everyone traveling separately.
Driverless vehicles are cars that can drive themselves. The host is saying the book is about the math that helps make that possible, not just general AI.
LIVE
Welcome back to the smart driving cars podcast. Thank you for spending time with us. I'm Fred
Fishkin, along with the faculty chair of autonomous vehicle engineering at Princeton
University, Alan Kornhouser. Hi, Alan. Hey, good morning, Fred. And happy Easter, I guess. Yeah.
For Jewish, we wish each other happy Easter short. Yeah. Why not? Hey, we're seeing all these great
views from where we were almost 60 years ago. Again, thank you. No problem. And, you know,
somebody suggests it's one world and we should get along. I mean, really, we didn't know that before.
We should get along. Thank you. Absolutely. Well, Alan, in the latest edition of smart driving cars,
you lead off with an MIT mobility interview with the CEO of Ekno Robotics, Carl Enyama,
says from DARBA to Robotaxies to warehouses, the full arc of autonomy. This is one you really recommend.
Well, yes, I put it up there on the top, actually, you know, put together this issue primarily because
of this, because it is a it is a really good one. I mean, it's been a great series and Jane
kudos to you for putting on such a great series. But this was this was really a good one. I'm,
you know, unfortunately, I haven't had time to attend half of them or I attended many but or
attended on Zoom. But this one, I just thought was really good and really serious and largely
fundamental. And and I encourage everybody to to watch it. I agree with at least most of them. Of
course, like, you know, you told me I'm not going to agree to everything. But really, it's, it's,
it's a good one. And thank you, Carl and Jin, why you really good.
And you also have a remembrance high up in the issue about the dinky from 50 years ago.
Wow. I mean, I, you know, I, I've forgotten. I hadn't really forgotten. But it didn't come to
mind on March 31. I can't believe it didn't come to mind because it seems as if it was just
just yesterday that 50 years ago, we welcomed the coming of Conrail, which of course doesn't exist
again anymore. I mean, it didn't even last the 50 years. I mean, less than only about half of them
actually, and maybe even less than half as to the construct to basically realign the railroads,
the seven bankrupt railroads of the Northeast into the memorably titled plan called the final
system plan. I mean, my goodness. Can you imagine somebody titling something, the final system plan?
Yeah, it's the final system plan, which basically took what was a total mess of a mobility situation
with respect to the railroads in the Northeast or really almost throughout the United States,
but certainly in the Northeast. And basically had the Congress say, geez, we have to do something
about this because this one of one of the very important mobility systems in the Northeast is
about to go bankrupt, not the good kind of there is a good kind of bankruptcy.
Chapter 11, the chapter six, which means it's not just it's not just Wall Street or investors that
get hurt. It's everybody because then you sell the assets, you tear up the tracks, you turn them
into bikeways. I guess there's some good that comes out of it. But they were in desperate
situations and the important date was April 1, 1976, when Conrail was formed. It was formed
not only for the freight railroads, freight part of the railroads, but what was left of the freight
part of the railroads, but also the passenger part. And of course, out of all that transition from
Penn Central and the other six bankrupt railroads to Conrail was the transition that took place
on April 1. It wasn't a it was not an April Fool's joke.
And Henry Posner III, Barrow Equinologist, I should say, who was a junior at the time at Princeton
like March 29, runs into my office. That's Cornish. Do you know what happens on a few days from now?
Conrail is born. Oh, great, Henry. Thank you. I guess I knew. Of course, I knew that since I'd
spent at least the previous year plus dealing with it. Oh, and you know what? The last train to be
dispatched to the Penn Central will be the Dinky at 1145 or whatever the time was. And the first
train dispatched by Conrail will be the Dinky coming back from the junction at 12068 or whatever
the time. Oh, Henry. Wow, I'm so thrilled. We have to have a party, Henry said. Well, you mean a party?
I'm all into parties. Sounds great. Henry, you bring the troops and I'll bring the champagne.
The rest is history. Probably the highest ridership of a Dinky ride,
certainly in the last 50 years, and maybe even in the previous 50 years, except for, you know,
Princeton Yale football game or something like that. Anyway, it was a memorable evening.
With most memorable evenings, there's always a drawback. One of the other photographs taken
in that was one that showed amidst the crowd toasting the coming of Conrail
and the demise of the Penn Central. We, of course, had to have the conductors join in the
team photo. And of course, they're holding up a glass also. But really, they were only holding
the glass. They did not, because as a railroad employee, I mean, one thing you didn't do was
drink on the job. They did not taste anything. But two days later, I could maybe the first
call made by a Conrail lawyer was, I'm a lawyer. To those guys really,
were they drinking? We need, no, no, no, they weren't. It was just for the photo.
They're the best guys in the world. They wouldn't think of it anyway. So
they didn't get in trouble when we had the photo and we celebrated the coming and
Conrail was created and Conrail turned out to be successful in any measure of success. And
the freight railroads in the Northeast are still doing pretty darn well. And freight does do well,
even though trucks move a lot of stuff. And if you really look at the success of the US railroad
industry, it is with freight, which basically had its birth 50 years ago, essentially today.
We celebrated it thanks to Henry Posner III, ferro equinologist for which he's been a student
of Iron Horse all these years. And we've mentioned him in podcasts in the last few years as well.
Oh yeah. I mean, he's still in an amazing career.
Talk about stirring up memories, Alan. Artemis II is carrying four astronauts around the moon and
back. You've got a link. And for both of us, you more than me, what memories this conjures?
Well, yeah, of course, it was part of the run up. It was,
it was. And, you know, we did that 60 years ago, almost 60 years ago. And,
and yes, it was, it was a big one simply because of the image that was the photograph that was
taken of Earthrise over the moon, you know, in December 1968. And, you know, they're trying to
put out some photographs and so on today. And they're very nice and so on and great. And it's
true that it's different. And, you know, when they pass around the back of the moon and they'll
don't get one that's similar to the one. But I told the client, I mean, just mentioned to the
class on Wednesday, you know, we did that 60 years ago, almost 60 years ago with slide rolls and
and an IBM 360 mainframe. Now with whatever. I mean, it should almost be easy. And it's not,
of course, but my goodness, it's a good thing. Everything's going well. I mean, in some sense,
if we progressed, it should. And we have a lot to do. There's a lot of discussion now about,
you know, what about science and science and science. And we can do science a lot better today
than we could do 60 years ago. And so we are doing a lot more science. Turns out we haven't done too
much about our own exploration. And yes, you know, do we really need to be there? I certainly
would like to be there. I guess I'm not going to make it, Fred. Whatever, yeah. Anyway, others will.
Well, exciting nonetheless. I made sure I told my grandchildren, Hey, be watching this.
Yeah, yeah, it's worth watching. And I think it's, it's worth watching the next ones. And it's,
it's, it's not, it's really nice to see the leadership that exists. And NASA today, at least,
you know, from my perspective, my goodness, if we had Jared, Isaac's ones, and any other places,
we would be, can you imagine how good we'd be?
Never mind. We don't get, I, we won't, we won't, I shan't be political.
What we're trying to do with driverless cars is I don't think political. It's just a value to,
to society. Absolutely. From MSN, Waymo has doubled its weekly ridership in under a year.
Now it's up to like half a million paid rides a week.
Yeah. I'm wondering why they're using a half a million per week, as opposed to a hundred
thousand per day. I don't know what the per week, why the per week measure sounds better than the
per day. I prefer per day version, but because the conduct, I don't know, maybe I'd like to think
in days, as opposed to weeks. But anyway, it is impressive. It is impressive. And it, you know,
it indicates that they're probably doing, you know, 25 rides or so per vehicle or per day.
And, you know, it's, that's pretty good productivity, especially when you, you know,
basically everybody you give a ride to is by themselves. And the question is,
you know, what can happen if you can get that average vehicle occupancy
above one? Because it's, it's not above one, because there's empty vehicle repositioning. I
doubt the things, but I don't know. Unless they, I don't think they just drop off a person and
pick up a person and then drop off a person and then pick up a person.
And the real value proposition of the driver list is empty vehicle reposition is to go from the
last person you gave off, gave a ride to or per sons to the next person or per sons,
you're going to give a ride to. And that's, that's the fundamental problem of mobility,
whether it's, whether it's coal, goods or people, where they're coming from.
It's not where they're going to. I mean, that's why you need it. And when you get there,
unfortunately, the darn thing is not as symmetric as you would like.
They're in somebody already there wanting to go somewhere else.
So either you sit around and wait, which is what people do with their own cars.
And therefore that's why you have parking.
It's the people you use that want to get back home.
So they use it to have the mobility sit around and wait for them on an advanced top concept.
But that's the value proposition of having an automation is to then say, okay,
pooh, nobody here wants to go, but maybe over there, there is. But when we go over there and pick
them up, which is really what Uber and Lyft do, except they have a person in there that needs
to be paid for that. And of course, in the numbers, it looks pretty good because there's
somebody in there doesn't look empty. They claim they have average vehicle occupancies greater than one.
Even though one of those entities in the vehicle is nothing but an algorithm.
I mean, it is a person. We shouldn't call person algorithm. I don't mean to be derogatory.
You can call him Al, right?
Right.
And so when people talk about or government talks about average vehicle occupancy,
they're counting the driver in there.
Buses end up looking good. I mean, wouldn't it be terrible if like Princeton University's
free bus system would have an average vehicle occupancy of less than one?
I mean, actually, it is. I hope it's only when I see it, there's nobody in there besides the driver,
but whatever. Next topic, we've got to get off.
That's not what you want to see. From simply Wall Street, Alan,
this might be a head scratcher. V2X Trial in Bordeaux, what's connected mobility story and focus?
I mean, I guess in Europe, they're still talking about connected vehicles with all this.
Sure, I'm glad that in the United States, we've moved on from that.
Anyway, and it talks about the demo ending and what to do next.
When a demo ends and there's nothing on the table to go on it, I guess, has that demo been successful?
Anyway, trying to think that somebody is going,
Mr. Trump is going to be out there controlling traffic with it because as information has to,
because they're connected, you can organize everything for greatest efficiency in our roadway
because it's going to be orchestrated from upon high. I'm glad I'm not going to live to see that.
Financial times as a report headlined, self-driving will not make Nissan a commodity, says CEO.
I mean, it's going to make. I think General Motors is not testing 200 vehicles. I guess
they're back to going to make two. They were making it with crews, but apparently
that wasn't good enough. I don't know. It's very nice that they're going. It's very nice.
They're going to differentiate themselves with respect to that.
Although one would like to do all this so that it is a commodity, so that everybody,
commodities are things that everybody can have as opposed to luxuries that only a few,
but anyway, and the way to stand down is to, I guess it's to have peacock feathers and
really stand out. So you put a chapeau on top of your car that everybody can see you coming.
I don't know. I guess I wasn't in a good mood when I saw that or something.
Whatever. I'll throw it out there for people. I think it's nice to see that somebody in the
auto industry wants to do that and the way they're doing it is with a chapeau.
Another automaker making some headlines, the Subaru from AutoBlog, the 2026 Subaru Outback
gains hands-free driving for the first time. This is similar to what I think Ford and GM have,
because it operates on interstate highways, at least some of them, many of them,
and so it's a step that Subaru is taking now. Subaru claims that they're better too,
so it's not only, it's even better than theirs. Their eyesight we've talked about in the past
was fantastic. I mean, for years I would show their commercial that they made in like 2014,
which basically promoted eyesight with this, you know, crash dummies and a crash test.
And it's just, to me, it's a classic. And at the end, it's basically an eyesight keeps it from
crashing and saves the dummies. And Subaru comes in there and says, basically, the best crash is a
no crash. I don't think that literally what they, essentially that's what they said, and it's beautiful.
And in fact, you know, they've been in all the tests that IIHS did early on with
automated emergency systems. Subaru is always out on top, if I had that working, and here they are
with this, and in the model they're offering it, it's standard equipment. And so I really praise
Subaru on this, and it's not just because Subaru's North American headquarters are in New Jersey,
but it's also because Subaru's North American headquarters are in New Jersey, they make such
good decisions about where they should locate, they've got to be good, right? We can argue against
that one, but in all seriousness, it's, they've been out there in the front of this. Mercedes
also, since I, you know, in 2014, I couldn't wait to go buy one. That was the first guy on my block.
But this is really a step up and committed and committed for the right reasons committed,
because it does really, at least they're convinced. Of course, I guess I believe them,
simply because the technologies of these things is getting to a point is that, you know,
they can do the driving function better than I can. They see better than I can, they react faster
than I can. I can't compete. So therefore, why not drive for me? Do I really have to sit there and
drive and do this lane keeping thing or trying to keep this thing between two white lines?
And in fact, I've argued that another one of the things that I'll be putting in there is basically
some comments about the Tesla truck
and the technology they're putting in there. But, but if you, if you, and the,
and the,
We're trying about the semi.
Yeah, the, the semi and Jay Leno, they do an interview with Jay Leno and Jay Leno talks about
how difficult the job of a truck driver really is. And it's not only difficult, it's dangerous.
And years back, I would assemble some data that basically was showing it more dangerous than
being a coal miner. It is a tough job, especially for long haulers. You're out there 24 hours to get
10 hours of productivities. Well, I said that you have some rest to be able to be productive
the following 10 hours. And so I guess you sit around, you sit around at a truck stop
or somewhere along a road.
But to me, the beauty of driverless on a long haul truck is that we should be able to make it so
that the life of the truck driver who's already dedicating 24 hours of her day to a truck
can actually let that truck and that cargo be productive more than 10.
Because if the truck's not moving and the cargo's not moving, or even if the empty repositioning
isn't moving to get to the next load, that's a lack of productivity and enormously valuable assets.
Even though it is, it is
just japeroned by a person 24 hours. Now they only get paid for 10, but they're obligated for 24.
They can't go home and be with their
their spouse, their kids, their parents, go to a ball game,
have to be with it. Or, you know, I guess if you're at some truck stop, there might be a
is is on is actually on duty 24 hours sitting there with that asset gets paid for 10
hours of service or whatever the hours of service are.
Why? Because most of that time is going to be spent trying to keep that truck between two white lines.
So she doesn't die.
Those folks need to be, they need to be paid. And I guess those that do it are paid.
But they really paid for 24? No, they're paid for 10.
Now you put automation, driverless automation on there
so that the truck and the cargo can be productive 22 hours a day
with the person there. Now you have 22 hours of productivity to pay for that
10 hours that you've been paying that trucker.
That doesn't mean they don't have to sit there and keep the darn thing between two white lines
because technology is doing that.
Technology is getting to a point where probably doing it at least as good if not better.
He could do it.
So all of a sudden you create a nice atmosphere in the trunk.
Have that person do something else during 10 hours spread over 24?
Oh and that truck moves now 22 hours during the day.
That asset churn and productive.
The entity that's responsible for that asset has a lot more money in their pocket
to pay this person.
Seems like a win-win and you haven't changed
the responsibility that the individual has to the truck which is still 24 hours.
You've changed what they do over that 24 hours.
Some of it is sleep. Why not?
And if something gets you need to be pulled over and you still need to have those little
triangles put out there to make sure nobody crashes into you like everybody would everybody.
Some people would like you to have instead of flashing lights.
You can person can be hey get up put out the thing they can do that.
They can do all the things that might need to be done under those situations.
They might be able to talk to a policeman if they're or keep it from running over a fire hose
or move it out of harm's way.
And I actually allow that truck to be productive now instead of 10 hours a day.
Not only the truck the car go in the back. That's inventory.
Anybody deals with that asset. I mean you know inventory that's sitting around.
I mean you don't need to be an orfy major at Princeton to realize who.
I mean anyway so the opportunity of driverless in trucks is not to remove the driver.
It's allow the driver to be productive. Allow the asset to be productive 22 hours and repeat it
many times. It's not talk about you know these postings that we see people writing these great
papers. It's a shame they're missing the obvious value proposition for this.
Now of course you know government is going to have to change the hour service rules.
But those hours or service rules are to make sure that the driver is able to keep the truck
between two white lines. That's not the job anymore. You know when you bring in a steam shovel
to dig a trench you don't need the gals out there with shovels shoveling.
You let the steam you let the machine do that.
They get to do you know things that end up being safer more productive.
Allow them to earn more better feed their family not take as many drugs.
Why isn't that the value proposition that everybody's talking about here?
Sorry. Well the focus has been on economics and the Wall Street Journal a couple of weeks ago
had a story about how much the truckers like being in this vehicle. So and as you were saying
that's that's a that's yeah right well if hey for office workers don't they like being on the
office worker with with a nice office isn't OSHA supposed to go in there and look at
what the surrounding ours for workers has OSHA gone in and looked at you know what truck drivers
are made to do made to keep a vehicle between two white lines when technology exists to do that
as well as they do it. And okay I just say as well because they are very good but likely better.
Terrific. Alan you've got a link to a book in the latest newsletter. Why machines learn the
elegant math behind modern AI by Anil Anantaswamy highly recommended by someone you know. Yeah I mean
just you know maybe you all want to wonder why you you know
John Hopfield I mean John Hopfield was reading it so I said who John's reading this I better
how am I so out of it I mean it's almost a year and a half late although I guess I hopefully I know
a lot of the stuff that's in there but it is and it's very well done it and and the really
and the arithmetic or math that's in it is explained very well and I just think it's a
it's a good read. And I understand you have a bit of a ride-sharing road trip coming up
Alan warming up. Yeah yeah we're we're taking our
design instead of flying we're going to road trip it and and do ride-sharing road trip so instead of
buying two airplane tickets will be two people in one car. Wow I think it's more energy efficient
and you know hopefully the Subaru system like that looks pretty on well for systems like that
they do then hey
why not we'll see what happens especially when it's ride-sharing simple economics right
I think so although you know airplanes have can you airplanes well yes
I guess some people don't fly commercial
and there is ride-sharing a lot of ride-sharing in commercial these days and in fact it's so
ride-sharing it's gotta be there isn't a seat available. A quick reminder to everyone
that you can find us these days on the transportation channel there's an app for iOS and Android
we're on Apple TV Fire TV Roku on the web as well so you can find us there you can also find us
of course at smart driving car dot com and my tech reports are at textination dot com
thank you for taking the time to watch or listen again have a wonderful holiday weekend
Easter and Passover stay safe. Yes and another thing I should say about about the book is that
it's not the math behind modern A to I it's really the math behind driverless vehicles
without that these we wouldn't be where we are
About this episode
Alan Kornhouser and Fred Fishkin kick off with a Princeton/industry-style roundup that blends autonomy news with big-picture mobility history. They highlight an MIT mobility interview with Ekno Robotics CEO Carl Enyama, tracing autonomy from DARPA to robotaxis to warehouses. The discussion then pivots to Waymo’s rapid ridership growth, why vehicle occupancy math can be misleading, and the core challenge of repositioning. They also cover connected-vehicle V2X trials, Subaru’s hands-free Outback update, and a long, economics-focused argument that driverless trucking should boost driver productivity—not remove drivers. They close with an AI book recommendation and a planned ride-sharing road trip.
An MIT interview on the Full Arc of Autonomy, Artemis II, progress at Waymo, Subaru Outback goes hands free and a warm reception for the Tesla Semi. All that and more on episode 410 of Smart Driving Cars with Princeton's Alain Kornhauser and co-host Fred Fishkin. Tune in and subscribe!
