Preloader image

The Medium Rules Episode 6: The Three Zeros – Self-Driving Cars with Oliver Mitchell of Autonomy Ventures

On this episode of The Medium Rules, host Alan Baldachin is joined in the HBA Podcast Studio by Oliver Mitchell, Founding Partner at Autonomy Ventures, to discuss all things automation. Alan asks “The Robot Rabbi” to predict the timeline in which fully autonomous vehicles will roam our streets, whether batteries are to blame for the seemingly “slow rise” of autonomous vehicles, what autonomy means for smart cities, and much more. You might be surprised by his answers! TO LISTEN TO THE PODCAST ON YOUR FAVORITE PLATFORM:

iTunes – https://apple.co/2J8Of9Q
Google Play – http://bit.ly/TheMediumRulesGooglePlay

LIKE, SHARE, AND SUBSCRIBE:

If you enjoy the podcast, be sure to like and subscribe to our YouTube Channel, and subscribe on iTunes to stay informed about long-term trends in media and technology.

INDUSTRY UPDATES AND INSIGHTS:

Follow @HBAMediaTech on Twitter and Instagram

Transcript:

Alan BALDACHIN: From the HBA Podcast studio in New York City, welcome to The Medium Rules. I’m Alan Baldachin.

[Intro music]

Joining me today in the HBA Podcast studio is Oliver Mitchell, the founding partner of Autonomy Ventures, a venture capital firm focused on autonomous and robotic technologies.

In addition, under his nom de guerre the “Robot Rabbi,” Oliver is a syndicated writer on all things autonomous and robotics, as well as a host of local robotics-focused tech events in New York City. Oliver is also the chair of the UJA New York Federation Tech Committee, of which I happen to be a member.

Oliver is in studio today to give us the lay of the land with respect to self-driving cars and the future of the self-driving vehicle industry. In short, we’re going to attempt to answer the $64,000 question: when will we have autonomy? As I think we’ll learn, this is a more complex question than people may realize.

So with that, let’s get started. Oliver, welcome to The Medium Rules. Thanks very much for coming in. I look forward to the conversation.

Oliver MITCHELL: Thanks so much, Alan. I did not know today I’m going to be a soothsayer and look into my crystal ball to figure out when autonomous cars are going to hit the street, but maybe I can put together some data points and together we could see what that would be.

BALDACHIN: Perfect. Sounds great. This is a concept that I find really compelling.

Mary Barra, the CEO of General Motors and I guess the only female CEO of a Fortune 100 company, and certainly of an OEM, famously talked about the “Three Zeroes” with respect to autonomous and electric: zero crashes, zero congestion, and zero emissions. Obviously the emissions is not just autonomous, but electric, which we can touch on.

But putting aside emissions for the moment, what is the dream of full autonomy?

MITCHELL: I think she captures it very well. Clearly autonomous cars is a big additive benefit to safety.

There are tens of thousands of fatalities on highways today, many of those caused by truck drivers with sleep apnea or just overworking conditions, where they’ll fall asleep on the wheel. They jackknife on the highway, which causes a huge backup both from a congestion standpoint and from a fatality standpoint.

Distracted driving is a big problem. In fact, today we have more safety benefits and more safety software inside the automobile – like the Mobileye sensor system, ABS brakes – but we have more crashes today because of the phone.

So it’s really important when thinking about autonomous cars, the really huge consumer and societal benefit from a safety standpoint.

Number two, emissions is huge. When you think about a car, it is the second most valuable asset that most American families have, but it is at one-fifth capacity most of the time. 90% of the time, it’s parked on the street. It’s not being utilized, it’s running under capacity, and it’s very expensive.

Autonomous cars can actually alleviate most people from that burden because it could drive someone to work and then it could go pick somebody else up. Rather than owning a huge asset, they would only own timeshares when they’re going to use it. I think that’s a huge aspect to autonomy as well.

Zero congestion. We live in New York City; we have the biggest congestion, I think in our lives, in this city. We have an older city. They’re talking about congestion pricing for the second time within the past decade. It looks like this time it’s going to past.

A huge aspect to that is the growth of Uber. They just put in legislation to curb that growth of ride-hail, of taxis, of people bringing personal cars into Manhattan, of trucks.

The idea of autonomy is you might not have to live in an urban center. You can live further away. You can take a nap on the way home and work on the way in to the office. It could be a huge benefit, when you think about that, in alleviating congestion.

BALDACHIN: Let me jump in there on congestion. I think the zero crashes – and we’ll get a little bit into some of the recent fatalities that have made the news – there’s a lot of misunderstanding and maybe misapprehension with respect to the safety/crashes/fatalities issue.

But just on congestion, again, I think that’s something that most people wouldn’t intuitively realize in the sense that you might think autonomous leads to more cars on the road. How can that reduce congestion? What is that story on congestion with autonomous? Why do we think there’ll be less traffic tie-ups? I guess accidents, but what else is going on there?

MITCHELL: I think we have to take a step back. We’re contemporaries. We grew up in a time period when owning a car was a moment of independence in our lives. When we got that license, we were finally independent from our parents. We could go where we want, we could say what we want.

BALDACHIN: Yeah, especially if you lived in the suburbs.

MITCHELL: Especially. The idea of the cars in the 1970s, the muscle cars, getting a Mustang, working on cars, collecting cars, that was a huge thing in our lives.

But my kids are in no hurry to get a license. I think going out 20 years from now, that’s going to be even more so. Every kid has independence. As long as they’ve got a mobile phone, they can get a ride-hail. They can get an Uber, they can get a Juno. They can get a yellow cab on a ride-hail. They don’t need to be wedded to the car.

So I think when thinking about congestion, there has to be some generational concerns. The idea of Zipcar, the idea of ride-hails, the idea of ridesharing, and the idea of autonomous cars go hand-in-hand.

Adam Jones, a Morgan Stanley analyst, put it really well. If you took ridesharing and you took autonomous cars, they really meet. They’re not two parallel lines; they’re two perpendicular lines, and they’re going down the same path. Really, what will happen with autonomous cars according to Adam Jonas – and I subscribe to it – is that people will own cars in a very different way.

I just leased my new Mini Cooper, and it’s just for me. It’s sitting in a garage on 59th and 11th most of the time. Why should that be? Why should I be paying monthly on that? Maybe I should only be paying when I’m going to use it.

Mary’s words ring really true. The car companies are thinking about what fractional ownership looks like. From an insurance standpoint and a liability standpoint, what does it look like? What could it mean that people with disabilities could have the independence that you and I had at 17?

Or when you think about the Baby Boomers – the population of the United States and globally, in the U.S. by 2030 we’re going to have 70 million people over the age of 65. As you know, geriatric senior citizens do not make the best drivers. Autonomous cars gives them an immense amount of liberty and independence and dignity.

So you have to put it in the whole societal confines – but that doesn’t answer your question of congestion. I think there’ll be less cars on the road, but it requires a different mindset in thinking about cars not just from the ownership, but thinking about where cars go and where people go and where bicycles go. This is a much bigger question that civil planners today are grappling with for the first time ever.

BALDACHIN: Staying on congestion for a second, even if you just think of zero crashes, less rubbernecking, zero lanes being blocked off, cars being able to have follow distance much closer one to the other – right there, you can see the difference.

Before we leave Mary’s Three Zeroes with respect to emissions, let’s just touch on electric in terms of emissions. It’s not directly self-driving, but as you said, it’s part of that broader ecosystem.

What is holding back electric? Is it battery technology? Is that the principal driver here as to why we’re not seeing electric cars all over the road today? I’m interested to hear your answer, and then I think we’ll go through where we are on autonomous, and that ties in as well.

MITCHELL: I don’t know if there’s one single answer. Clearly, government has a role to play. Clearly, the lobbyists for oil companies have a huge say in Washington, potentially in this administration. We had a former Secretary of State from the largest oil company in the world.

BALDACHIN: He didn’t last long. [laughs]

MITCHELL: Right. He didn’t like the salad. [laughs]

BALDACHIN: He didn’t like the dining room situation.

MITCHELL: But I think that with electric, there’s a number of things. I think every car company has a goal of switching over to electric. Zero emissions is not novel to GM. Ford is out there with the Bolt. Mercedes announced that they’re going to have a competing car to the Tesla on the luxury side. Every single car company has an electric mission, and I believe by 2021 they want to start converting over.

BALDACHIN: Rolling them out, yeah.

MITCHELL: When thinking about lithium technology, there are a lot of shortcomings. Lithium ion batteries are somewhat unstable. Liquid electrodes, if you poke it, it will explode.

There are people like Ionic Materials in Cambridge that have worked on solid state lithium that you could actually poke a hole through or you could throw it and it will not explode. You saw with many of the Tesla crashes, they were exacerbated by the reaction to liquid lithium ion explosions, where the entire front end was blown up.

When you think about a Tesla today – we’re in New York City; we don’t have any Supercharging stations, but I know even with a Supercharging station, you have anxiety about running low on charge. To charge at a Supercharging station still takes 40 minutes, and that’s with the best charging. To fill up your tank with gas takes maybe 5-10 minutes. Clearly from a consumer standpoint, that requires a different mindset.

I think that autonomous vehicles and electric vehicles go hand-in-hand. Autonomous vehicles use an immense amount of processing power. Today, the Google minivan, the Chrysler Pacifica is a little bit of a hybrid, but still is burning a lot of fossil fuels to run that huge processing power, that deep learning system.

When you run deep learning systems, it’s burning a lot of energy. In order to do that, it needs fuel. It’s really inefficient, from a fossil fuel standpoint, how much it burns. In order to make it fuel efficient, you have to switch to electric.

We’re not really there yet on the autonomous car side on getting the deep learning systems to be energy-efficient, and also cost-efficient. You mentioned Lunewave, which is a radar system.

BALDACHIN: I hadn’t, but we will come to that.

MITCHELL: You mentioned it to me pre-[show].

BALDACHIN: Yeah. Just to jump in, that is an Arizona-based tech transfer company out of the University of Arizona that just announced this morning a $5 million seed round.

MITCHELL: One of the things of its value proposition is it’s a very cost-effective radar system, and it professes to be even better than existing systems out in the marketplace like LIDAR and other radar systems. So cost-effectiveness is really important. That’s why I bring it up; sorry to jump the gun there.

BALDACHIN: No, not at all. Go ahead and finish your thought.

MITCHELL: On the Google car, which is being tested in Phoenix, Arizona, the components inside just to do the deep learning cost upwards of $300,000. So unless you’re uber-rich, you’re not going to be able to afford that. Even on a fractional standpoint.

The idea of the car companies, right now they’re testing out these systems from a quality control standpoint, but they’re also looking at how they can bring this down from a cost-effective standpoint so it could mean margin and mean value to the consumer.

BALDACHIN: Let’s segue from there into a little bit of education. In 2014, the Society of Automotive Engineers defined five self-driving levels, with “0” being no self-driving features and “5” being completely autonomous, meaning no steering wheel, no gas, no brake pedal, no horn. Maybe they’ll leave the horn in. [laughs]

Can you walk us through those five levels, and maybe put us on the map as to where each level sits in terms of what people can relate to and what we’re reading about and hearing about?

MITCHELL: As a Manhattanite, I hope the horn is the first thing they take out. [laughs]

BALDACHIN: There’d be no more movies about Manhattan. [laughs]

MITCHELL: Exactly. When you think about Level 1 to Level 5 – we were talking about 1970s muscle cars. That’s a Level 1, with a carburetor, burning lots of oil and gas. When you think about cars that you buy off the lot today with some ABS technology, those are Level 2.

When you think about a Tesla running on autopilot or an Audi, which is coming out, running on the NVIDIA autopilot system, that’s a Level 3. It will drive hands-free and pedal-free with a human driver behind the wheel, with their hands close by where they would have to take control of it.

BALDACHIN: So-called safety driver?

MITCHELL: Yes and no. In the safety driver scenario, those are ones where they’re being tested out. They’re called a safety driver. The Google car and the Uber car is more supposed to be a forward-thinking Level 5.

This would be actually the owner of the car – you’re commuting from Scarsdale, and potentially on the Saw Mill Expressway there’s a lane dedicated just for hands-free driving. Then you would activate the AI and you could take your hands off the wheel. You would go bumper-to-bumper, safely, and everyone would be safe there.

BALDACHIN: I can look at my phone. I’m off the clock.

MITCHELL: Right. It would require you still to pay attention, but it would be in a zone.

But it’s really important – let me just take a step back, Alan. There’s about 15 public tests on the road of autonomous cars of different levels. (We’ll go back to the Level 1 to Level 5.) There’s a lot of excitement out there. Google is actually running tests without any safety driver, and we can talk about that more in a second.

But these are in geo-fenced-in areas. These are in areas where they have shielded traffic and pedestrians and such away, and where the car has already mapped out everything going on there. Of course there’ll be instances where there will be unknowns, but it’s as close as possible to a perfect scenario. They’re in places where the climate is sunny most of the time, there’s not a lot of weather to contend with, and they’re geo-fenced in.

That’s really important when you think about autonomous cars. Even coming to New York, when you think about ridesharing, I think there’ll be a lot more autonomous buses in geo-fenced-in areas within New York City, going up and down Sixth Avenue, than there will be consumer cars.

Thinking about that from a Level 1 to Level 5, it’s a simplification, but there’ll be offshoots of that. If you think about it, New York City and many cities across the world have been running autonomous trains to airports for decades. Those are geo-fenced in. They’re on their own track; they don’t contend with Metro-North trains or human trains or traffic and stuff.

Taking that to the automobile and to municipal shuttling services and also trucking, that seems really foreseeable if you can create that geo-fenced-in area.

So today we have a number of competing technologies at Level 2+ – I’ll just call it that today to – Level 4. Let’s call it Level 3.

It’s a very dangerous scenario in the sense that when I’m driving, I’m focused on the road. Even if a driver swerves, I can react very quickly to avoid that.

If I take my hands off the wheel – in the case of the most famous crash, which happened a couple years ago, the Tesla crash in Florida – it was a fatal crash where the guy was watching a movie and a white truck was making a left-hand turn as the sun was shining down. To the AI, to the computer vision, it saw it as just horizon because it was a white truck, and before it was the sun.

BALDACHIN: With glare.

MITCHELL: And with glare. So the vision did not see anything wrong with that, and he was going 70 miles per hour. When the guy tried to take control of the wheel, it was just too late.

In fact, Gil Pratt a couple years ago at CES, said that’s the most dangerous scenario, that Level 2 to Level 3 driving, that autopilot scenario where people are taking their hands off the wheel. For them to reengage could mean two seconds of life to death.

So when thinking about the transition to full autonomy without a steering wheel, without pedals – which is much further out on the continuum – we have to think about maybe there are scenarios where that could be done safely.

The idea of lanes – when you go on the LIE, there’s HOV lanes. Potentially you could have hands-free lanes, and those could be fuel-efficient lanes, and those could be carpool lanes. They could be all the good stuff that you want to encourage the population to do, including autonomy.

When thinking about shuttling services, you take Via, which is a ridesharing service here in New York and in Chicago.

BALDACHIN: Countrywide, yeah.

MITCHELL: Potentially that could be autonomous, going down avenues.

BALDACHIN: Specific routes, meaning.

MITCHELL: Yes.

BALDACHIN: Let me jump in there and say this really gets into what I think a lot of people don’t understand. I use this example; a friend of mine maybe 6-7 years ago calls me from a Google self-driving car telling me, “I’m calling you, nobody’s touching the wheel. This is insane. We’re on Highway 101.” That was a while ago. So yes, there’s the technology, but what’s the phase-in look like?

I think you really started to answer that question, which is it’s not all at once. It’s sort of incremental. Benedict Evans wrote a very compelling blog post about phasing in where a vehicle might be L5 in one situation and L2 in another.

Maybe talk about that phase-in, because that gets to “when will we have autonomy?” to some extent.

MITCHELL: Yeah. I think in this current culture of instant gratification, whether on the mobile device or flipping a light switch, we tend to think of technology as flipping a light switch. We don’t really see it as the long view of the development.

It’s really accelerating very quickly. There’s Moore’s law, which has held true for 50+ years; with processing speed and cost, every 18 months it will double. It’s moved extremely quickly. If you think about the first DARPA Challenge, the Carnegie Mellon team went out to the desert and went seven miles, and they won the DARPA Autonomous Vehicle Challenge. Two years later, they went 300 miles.

BALDACHIN: It’s amazing.

MITCHELL: We have many of the ingredients of the technology to successfully drive a car autonomously.

I think back to a picture that I shared with you over breakfast that was in the IEEE Spectrum a couple years ago of Magnavox RCA engineers doing autonomous cars in 1965 on a track in Detroit. So this has been something that’s been in discussion for many decades, maybe even since the beginning of the automobile.

I think with that long view, let’s talk about where we’re at. We have a lot of technologies that are available, that are getting less expensive – like sensor-based technologies like laser, directional technologies like LIDAR, like radar and sonar, like computer vision technologies – and we have technology knowhow.

But at the same time there’s a societal aspect to change, which moves a lot slower. If you look at the adoption to technology traditionally, for society to adopt new technology, it’s been 12 years. We’re at an 8-year cycle, but at the same time, the acceleration of advanced computing technology is like a hockey stick. It’s almost surpassed it.

We saw that with Uber. Uber’s popularity outpaced regulatory…

BALDACHIN: Ability to adapt.

MITCHELL: Yeah. So we’re seeing them play catch-up.

We cannot talk about autonomous cars without talking about deep learning systems. Traditionally, a computer has been binary. It’s been programmed. When X happens, Y happens. But deep learning systems, neural networks, think like our brain. They have to input everything that it sees.

A computer vision technology or a LIDAR or sensor vision of multiple sensors that are now on the automobile, that goes to a deep learning system. It’s now gaining knowledge in the same way that a 17-year-old driver is gaining knowledge.

Google’s car, for these past 6 years, has been driving to gain knowledge. It could not be driving around Phoenix, Arizona without a safety driver without those earlier tests in San Francisco on 101.

It had a few accidents. Humans are not really good drivers, and it expected every human to drive well. When a bus went into its lane and didn’t slow down to go into the lane, it sideswiped the Google car. So it’s learning. Sometimes it might have to ramp up the aggressiveness of its driving.

I think these systems are learning right now, and we’re going through a period where people are getting really excited. They’re seeing lots of milestones, in the same way the space race was, but we have to remember that it’s dangerous, just as there were accidents along the way to landing a man on the moon. That movie First Man is going to be coming out about Neil Armstrong.

So I think with autonomous cars and autonomy, we’re going to see a lot of hiccups and bumps along the road (pun intended there), but at the same time we have to look at the long view. It’s not like turning on a light switch.

BALDACHIN: Would it be fair to say – let’s just talk about the U.S. for a moment – that because of driving conditions and four season weather in the Northeast versus beautiful weather in the Southwest, that that movement may go west to east and south to north?

In other words, what I’m grappling with is New York City is such a complicated, dense, pedestrian caught-up city in many ways – although obviously Uptown – but that’s such a different environment, let’s say, than Arizona. New city, big roads, few pedestrians. That’s part of the roadmap, presumably?

MITCHELL: Yeah, I think that’s correct. I think something that’s missing from that description is that in America we have 15,900+ municipalities that govern their local roads. And we have 50 states that give licenses. So getting everyone on the same page – and then we have the National Highway Safety Board that sets guidelines.

Even in the least regulatory environment – which in my opinion, maybe as a liberal Upper West Sider, is dangerous because regulations help guide innovators on what to develop towards and have confidence in what’s released – I think even with that, you’re going to have different use cases for different types of scenarios.

We’ve been talking about Arizona. I lived in Arizona; the roads are six-lane roads. They’re extremely wide, extremely well-marked. It’s sunny 360 days a year. It’s pretty easy. It’s on a grid system. You’re not talking about lower Manhattan, where some streets are still cobblestone.

BALDACHIN: Much less London, much less the Balkans, much less Eastern Europe.

MITCHELL: And you have weather, you have visibility, you have marking of lanes, you have pedestrians, you have bicyclists, you have baby carriages. You have a lot of unknowns there. So yeah, definitely it’s a lot easier [in Arizona].

But with that said, it was announced last year that GM’s Cruise Automation group is going to be starting testing in lower Manhattan.

BALDACHIN: That’s their group out of Tel Aviv, correct?

MITCHELL: No, it was normally out of San Francisco, and then they integrated it within Detroit. They paid a billion dollars for Cruise. It was one of the big exits out there in that field.

I would say that Governor Cuomo has a number of initiatives throughout the state. For example, center state, which is up in Syracuse/Rome, New York, has an accelerator for unmanned systems called Genius New York.

And they’re not any different than say in Pittsburgh. The mayor there was very aggressive in trying to attract Uber and share the talent – at least from the mayor’s and Uber’s perspective, not from the Carnegie Mellon standpoint – of Carnegie Mellon with Uber. Carnegie Mellon sees it that they poached 15 top professors with very high-paying salaries. [laughs]

Arizona tried to court Uber executives with a freewheeling, un-regulatory environment to bring its autonomous driving test of autonomous taxis to Arizona from San Francisco – which in many ways led to maybe negligence in them not having two safety drivers and just having one. Don’t know.

We’re right now in a very unknown period where there’s a lot of initiatives happening. We’re accomplishing a lot of data-gathering.

And we can’t just be America-centric. Singapore NuTonomy (which was a Cambridge, Massachusetts based company which was bought by Delphi, which is a Tier 1) did most of its testing –

BALDACHIN: Tier 1 OEM.

MITCHELL: Tier 1 parts supplier to an OEM. Delphi is one of the largest ones out there. They paid I think about $400 million for NuTonomy; they were a very young startup.

They did most of their testing not near the MIT campus, but in Singapore, because Singapore is very aggressive in trying to be the tech hub of the world for many things. For drones, for autonomous shipping, for cybersecurity, for IoT. That’s a whole separate discussion.

BALDACHIN: That is a whole separate discussion, which we can have.

MITCHELL: But in Europe you’re seeing a lot of interesting things happen, even in England and in France.

So I think in looking at the crystal ball, I can’t really give any dates. The industry has said 2021 is a magic year. That’s what many executives have said. But the one thing about making predictions is you can always push it out; I imagine that will happen.

I would say some of the early use cases will be going back to alleviating municipal bottleneck of mass transit, which we have in New York. If we didn’t have Via in New York, think about how much congestion there would be underground, even more so. It would’ve been the summer of Dante’s Inferno. Just hell.

BALDACHIN: Brutal summer in New York City this summer.

MITCHELL: Think about Citi Bike, how much that’s alleviated mass transit. In LA right now you have all these scooters out there, Bird and stuff. So you’ll see that.

I think a big area that we haven’t spoken about is long-haul shipping. Driving on the highway is a much easier premise.

BALDACHIN: Trucks, we’re talking about.

MITCHELL: Yeah, and shipping logistics. There’s a huge investment made by the big shippers out there. FedEx – I was just down in Memphis in the FedEx headquarters, seeing what’s happening at the warehouse down there.

BALDACHIN: That’s interesting.

MITCHELL: It’s very cool. It’s fairly automated, but they have a lot of human labor. Really, the human labor is just flipping the boxes so the robots can read the scans and weigh them.

BALDACHIN: Is that right?

MITCHELL: Yeah.

BALDACHIN: Wow.

MITCHELL: This is just a side note that I learned. When you fill out that FedEx package and you write down how much it weighs, and you’re always like “I’m just going to say 1 pound,” and it’s an envelope – the sorter actually weighs every package and scans it. That’s only to train you on what to expect for your billing.

In fact, that robot, that computer system that weighs everything – it’s a Swiss computer that’s been in effect for like 15 years – ties directly to their P&L. It goes right to the revenue line. It’s going right into the bank. It’s pretty magical.

So FedEx, Uber, as you know, UPS. When you think about Amazon, a huge amount of their –

BALDACHIN: Trucking distribution, delivery.

MITCHELL: And also aviation. And when you think about how retail has shifted, we’re ordering so much more on ecommerce. A great example is – this is going to be for all those Manhattanites who have doormen – doormen didn’t have to deal with too many boxes. Now they’re overwhelmed with boxes.

BALDACHIN: Funny you say that. Just side note, the idea of having a non-doorman building – I mean, I hate to sound like a privileged – by the way, I’m an Upper West Sider too. You mentioned Scarsdale; wouldn’t want anyone to get the wrong impression. I am also a New Yorker.

MITCHELL: Proud Upper West Sider.

BALDACHIN: Yes, very proud. But I would say the idea of [a doorman] – at this point, forget about security for your kids. It’s receiving packages and returning packages. That’s what you really need them for. Groceries, everything.

MITCHELL: So let’s think about that from even an autonomous vehicle standpoint. Let’s just expand autonomous cars to autonomous vehicles. When you order that package on Amazon, you click “Buy,” that message gets sent to a computer.

That computer connects to the Kiva systems that they bought in 2012 for $775 million – $75 million more than they paid for Zappos. Those are robots that are in the warehouse that bring the shelving to the picker, who’s going to put that in a box.

It used to be, before then, there was an Amazon employee on roller-skates that would go three football fields down to get that package. I don’t know how many accidents happened on the roller-skates.

BALDACHIN: I had a friend in high school who did that for Canadian Tire up in Canada in their warehouse, and they were pickers. Literally, hand-pickers. Their warehouses were massive, and they went around in trucks and by hand. Yeah.

MITCHELL: Amazon had rollerblades. You can imagine how many collisions happened there. [laughs]

The Kiva system not only brings the shelving to the picker, but reorganizes the entire warehouse based upon bestsellers. That’s pretty incredible.

Now that gets put into a box, it gets scanned, it gets put on a truck. That truck has to go to a distribution center. That truck can go autonomously to a distribution center. It goes on an airplane. So it goes to a distribution center, the distribution center scans that box, it goes on many conveyor belts, autonomously scanned, weighed, and then gets specially packed to be weighted so they don’t burn too much fuel inside the plane.

I talked about lithium batteries – this is very cool. The planes actually have robot arms now, the new 777s in FedEx, and they shared this with the industry. If there is a fire within their boxes of a lithium battery that catches on fire, it can autonomously break through and put out that fire.

They shared this technology with all the other shippers out there so the plane doesn’t explode in the air. Pretty good.

BALDACHIN: Good idea, yeah.

MITCHELL: Most of the time that plane is going on autopilot. It takes off from Memphis, it lands, it goes to a distribution center, and it needs to go back on the highway. There, you can have an autonomous truck.

When Tesla released its autonomous truck, it got a huge bump in its stock. Mercedes has an autonomous truck that’s been at the truck shows for a number of years. Uber bought – I have a mental block right now – Lior Ron’s company for about $680 million. I’ll think of the name of the company.

BALDACHIN: That’s okay, it’ll come back to you. Don’t worry about it.

MITCHELL: They actually put a halt on their autonomous trucks. They’re looking to partner with someone on that. It’s done deliveries for – it’s proved out on the highway. So that’s a real aspect.

Then you’ve seen examples of drones delivering. I don’t know how efficient that is. Maybe in rural areas that makes a lot of sense.

We’ve seen terrestrial robots deliver within urban areas. We’ve even seen Boston Dynamics have amazing bipedal humanoid robots do flips, or the robot dog. They can open up doors and they can actually deliver packages.

The idea of this ecosystem of autonomous vehicles is very true. When it will happen…

BALDACHIN: I think that’s a great run-through in the sense that thinking about being able to take a ride from here to my cousin’s house in Westchester or Philly, that’s a tiny aspect of it.

I’ve read that reality for consumer use cases and traffic and just getting around is like 2027, 2030, who knows. But that’s what I’ve seen. Whether that applies to New York or whether that’s West Coast, I guess we’ll see.

MITCHELL: I think we’ll see incremental steps. Really, today the car that you drive is a computer on wheels already. Eventually there’ll be a button for AI when you’re in traffic, and rather than you stopping and going, the AI will be able to do that. Or hands-free driving. The Mercedes has hands-free cars. Tesla obviously does the autopilot system.

You have to give some time for these municipalities to figure it out. One thing that the National Highway Safety Board regulates is the interstate highway system, so clearly on the interstates they could create a national policy of lanes for hands-free driving. Whether this administration or the next administration will, that’s to be seen.

On a side note, it’s pretty amazing – if you think about technology and you think about government, there is no Department of Technology. Technology is a subset of every department agency. Eventually you might have a cabinet post or technology czar that will have to work out all these issues.

BALDACHIN: Interconnection issues, yeah. One area which I don’t know that we’re going to have a chance to touch on, but I’m going to tee it up for the next time you come back in – which I hope you will, to have a conversation – is smart cities, Second-Order Effects. What are cities going to look like in a self-driving future?

That in and of itself is obviously an extremely rich area to plumb. It also engages the workforce displacement that is going to occur, not just with self-driving, but with robotics generally – which I think is going to be a great topic for us to address.

I’m teasing it because what I want to do is finish off with a perspective on investing, and also give you an opportunity to plug the UJA Tech Mission to Israel 2019. I know that you are a big fan of investing in deep technology centers around the world, including in Israel, which you have said is a land flowing with milk, honey, and AI and autonomous.

Talk about both those things. Talk about how you think about investing, and then maybe talk about the UJA Tech Mission, which is an annual trip to Israel to do an amazing deep dive into technology there and what’s going on in innovation – which I went on last year and I would highly recommend. But let me let you address both of those things.

MITCHELL: Yeah, I think a couple things. Maybe I can segue in just telling that Israel actually does have a smart city, Ashdod. The Port of Ashdod is a smart city.

When walking around whatever city you’re in, look up. Look at your lampposts. You’ll see a bunch of fishtail antennas. Those are sensors that are monitoring traffic right now – pedestrian traffic, automobile traffic, air quality. There’s a whole bunch of sensors that are being deployed, and there’s a number of companies that are working with municipalities. There’s a number of RPs out there, there’s a number of partnerships with the Department of Transportation on the federal level with the local municipalities, as well as other federal agencies on that.

Every year – and you were privileged to go on the last one – I lead a tech mission with the United Jewish Appeal as well as the Israeli Economic Mission here in New York. It’s a 4-day, intensive trip where we visit accelerators, incubators, academic labs, other venture capitalists, investors, really the leaders and influencers of this magic landscape.

Clearly it has an immense amount of historic connection to the Jewish people, but today, Israel outside of the Jewish community is just seen as another Silicon Valley. It’s actually called Silicon Wadi.

Talking about autonomous cars – in 3 years, we were meeting with Gil at GM. They said they went from about 12 engineers to 300+ engineers, and GM looks to Israel for the latest AI and autonomous technology as well as –

BALDACHIN: Not to mention cyber. We haven’t talked about cyber auto, but that’s a whole other topic.

MITCHELL: I know we’ve been hanging out a lot, because that was the next word that I was going to say.

BALDACHIN: [laughs] There you go.

MITCHELL: That’s really exciting. This year we’re going to be doing the mission again in conjunction with the Cybersecurity Conference, and we’re going to be meeting with some amazing companies as well as going up to the Technion, which is in Haifa, for the first time.

That is where Apple’s headquarters are. There’s a Google hub there, Intel’s hub is up there. The Technion is like the MIT of Israel. Some of the leading companies have come out of the Technion. It’s pretty incredible. They actually have robots that can detect smell up there. Actually, an Israeli Arab developed that technology.

BALDACHIN: Mind-blowing technology. Fantastic.

MITCHELL: I could maybe do one or two of these meetings on my own; I could not do six a day.

BALDACHIN: It’s a packed, packed itinerary. It’s incredibly high level.

MITCHELL: Extremely. Really what we do combining with the UJA is we have a social impact component as well as, with the Israeli Economic Mission, we have a business and investor component. We do a really great job of curating a great program and creating great context around it.

For example, on the UJA side we went to the basement of the latest WeWork and we saw a 3D printing lab where they’re printing prosthetics at inexpensive prices to enable kids around the world – doesn’t matter what religion, what race you are, around the world in India, in Africa – to be able to afford a prosthetic limb out of that 3D lab. That’s all backed by the UJA grant system.

Then we left there and we had dinner with Yair Shamir, one of the seed investors in Mobileye, which was sold to Intel for $15+ billion. We had the head of the Innovation Authority, which gives a number of grants out there. It was an amazing experience.

We went to the Capsula Lab last time, and Tel Aviv University Ventures. Tel Aviv University was voted this year as the third most active university for producing startups worldwide.

BALDACHIN: That’s incredible.

MITCHELL: It’s an amazing trip. It’s intensive, it’s a lot of fun. We have great people like you that go on it.

BALDACHIN: Thank you.

MITCHELL: I hope maybe this webcast/podcast brings in other amazing people. Any questions, feel free to contact Alan.

BALDACHIN: Absolutely. I was going to say on that note, in a future podcast coming up, we’re going to get Hadar Weiss here, who is at the UJA on staff, puts this together internally at UJA, and goes on the trip. We’re going to devote a whole podcast to talking about the UJA Tech Mission, because I think it’s that worthwhile, and the other things UJA does.

MITCHELL: I look forward to coming back to discuss the future of the work, because that’s something I’m very passionate about.

BALDACHIN: That sounds great. I think that’s a great place to stop, Oliver. Thanks so much for coming in. I look forward to having you back.

MITCHELL: My pleasure. Thank you, Alan.

[Outro music]

That’s a wrap on this episode of The Medium Rules with Alan Baldachin. For more information, go to our website at www.hballp.com. You can also follow us on Instagram, Twitter, and Facebook, and don’t forget to rate us on Apple Podcasts.