A Deep Dive Into Google's Self-Driving Car Project

It's easy to get so involved with the exciting and rapidly evolving technology of self-driving cars, and forget about the big picture. Where are we really, in terms of getting these things on the roads? How much will a self-driving car actually cost once they're on the market? And what the heck is a search engine doing pioneering automotive technology in the first place?

In this week's episode of Industry Focus: Tech, Sean O'Reilly and Dylan Lewis answer these questions and more about the autonomous car landscape today.

A full transcript follows the video.

This podcast was recorded on April 8, 2016. 

Sean O'Reilly: Why exactly is Google (GOOG -1.10%) (GOOGL -1.23%) making self-driving cars in the first place? All that and more on this tech edition of Industry Focus. 

Greetings, Fools! Sean O'Reilly here in Alexandria, Virginia at Fool headquarters. It is April 8th, 2016, and joining me, my partner in crime, the dynamic duo is back, Mr. Dylan Lewis.

Dylan Lewis: Yeah, great to be back. I had some fun doing the show without you.

O'Reilly: I think I speak for all of our listeners right now when I say, great job on the April Fool's podcast!

Lewis: I'm glad you enjoyed it.

O'Reilly: Can you give us a little behind-the-scenes as to what went into that or anything before we start?

Lewis: Yeah. April Fool's was Friday. We had some of the content and some of the tech -- if you haven't seen it, listeners...

O'Reilly: Spoiler alert: You really need to listen to it. (laughs) 

Lewis: Yeah, listen to the show, and if you just Google "Motley Fools Rush In," the page will come up. But in preparation for April Fool's Day, I think I was up until probably 2:00 a.m. at the office. I slept at the office getting the content ready. Ilan Moscovitz, who was on the show with me, and who is just so committed to the bit...

O'Reilly: And he's been doing it for 10 years.

Lewis: Yeah. He was up until 6:00 or 7:00 in the morning, getting everything together. Reuben Dalke and Ron Southwick, our tech guys, were also up until 7:00 in the morning or so. So, big ups to them for making it happen. It landed with a lot of people. We got some really awesome responses from listeners and readers. It's always a treat. I am happy to say we won't have to do it again for...

O'Reilly: Another year.

Lewis: ... another 12 months. But, if you have any good ideas for us, let us know, because we are always interested in getting some pitches on that.

O'Reilly: That's right, folks. Email us at [email protected] with your April Fool's ideas.

Lewis: Way to work in the email there.

O'Reilly: I tried, thank you! All right, so, Dylan, this is a self-driving car episode, but more specifically, it's actually the Google self-driving car episode. We've talked about this a lot before, it's been in the public nexus for years now. I've actually... gosh, I've probably known about the Google car's existence for... you see all these videos on the news all the time of these things scooting around Southern California. Why do you think Google first started doing this? They're a search engine. They dabble in things, but they really have committed to this.

Lewis: Yeah. This is classic Google chasing innovation. But, where all of this really started was kind of in the mid-2000s. So, the Defense Advance Research Project Agency, commonly known as DARPA, which is...

O'Reilly: Way to work in DARPA there. (laughs) 

Lewis: (laughs) Yeah, right? They're the most prominent resource organization for the United States Department of Defense. They held this grand challenge, and the idea was to create these long-distance competitions for driverless cars. And you had teams from some of the top engineering schools in the country, Stanford, Carnegie Mellon, a couple others competing. What ultimately happened was, a lot of the people who were involved in this competitions, in those university-led teams, wound up going over to Google and becoming part of these projects.

O'Reilly: Oh my gosh.

Lewis: Namely, Sebastian Thrun, who was the leader and foreman of the project for a while, and then Chris Urmson, who is now the director of the project.

O'Reilly: At the competitions that you're describing, do you think a guy with an eye patch came up and said, "I'd like to talk to you about the Driverless Cars Initiative?" (laughs) 

Lewis: Yeah, (laughs) "What can you do for me in about 10 years? Because I would really love to be able to drive around." Yeah, that's one of the beauties of driverless car technology, is it gives mobility to a lot of people who otherwise wouldn't have it.

O'Reilly: That was an Avengers joke, by the way.

Lewis: Oh, it was?

O'Reilly: It did not land with you at all.

Lewis: It went way over my head. I thought we were talking about getting people that might not be able to get licenses going. That's what I get for not being pop culturally aware.

O'Reilly: (groans) That's right! You know what, I just realized I made a terrible mistake, because I know you don't like superhero movies. Sorry.

Lewis: Yeah. It's a good thing I don't do the CG show.

O'Reilly: Yeah, very good thing. So, you talked a lot about this in your South by Southwest preview. And you actually profiled the guy over at Google who's doing this, Chris Urmson. Cool guy, I assume?

Lewis: He was a pretty cool guy, yeah.

O'Reilly: You went there, you heard him talk. What were some of the biggest takeaways, as a rehash? Because it's obviously pertinent to what we're talking about today.

Lewis: Yeah. We actually didn't have a chance to talk about the South by Southwest recap on the show. I did South-by sound bites, maybe three weeks ago.

O'Reilly: I like the pun there. (laughs) 

Lewis: A bunch of little interviews I did at South by Southwest, some were on virtual reality. But I thought it'd be fun to recap some of the stuff on the driverless car side, because that was obviously another major theme there. So, Urmson talked about how the biggest challenge -- you talk about consumer attitudes, you talk about the regulatory environment, these are things people often cite as huge road bumps for them that they need to get over. But for him, it's still on the technical side. So, that's the major issue, and that's where a lot of their resources are focused.

O'Reilly: So, he's not even thinking about that stuff?

Lewis: He's thinking about it, and we can get into some things that signal that the regulatory thing is definitely on his mind. But, I think in his eyes, he wants the product and what the cars are capable of to be perfect, that there's no possibility of something catastrophic happening.

O'Reilly: Approval will be a foregone conclusion.

Lewis: Exactly. So, more specifically, he talked about how these cars are basically having to handle this infinite number...

O'Reilly: Literally infinite.

Lewis: ... of situations. Yeah. So, you have a lot of machine learning going on, and you have cars learning from situations that other cars have run into. So, it's not just this, "I've been on the road for eight hours, I've learned eight hours' amount." It's, "I've learned from the entire fleet and what they're learning." But, that's still a challenge. People don't always act rationally, you have these irrational things happening on the road. I mean, we saw a screenshot of a bunch of people playing Frogger with the Google cars. People see the Google cars and they start acting out and doing weird things to them. So, that's just a whole 'nother thing they have to handle.

O'Reilly: Yeah. This is just me being a Bond villain, but did Google ever think of going to a town with like 10,000 people and being like, "Hey, if you outlaw regular cars, we'll give everybody a Google car," and then that would be a great test ground for the driverless cars, because as far as I can tell, the biggest problem with driverless cars is non-driverless cars. So, did anything like that come up at all?

Lewis: Well, what they've done with a lot of their test sites, this is before they wound up going into major metropolitan areas, is these abandoned naval bases and things like that.

O'Reilly: I like where this is going.

Lewis: So, they have these urban setups, city roads, and...

O'Reilly: They build fake towns?

Lewis: No. They took over the closed course areas. So, they'd have engineers walking around and having the cars interacting with them, and that was the training ground for these cars. And now, they've hit the point where they're able to roll out into these major metropolitan areas. First off, it was Mountain View and Austin, then I think...

O'Reilly: Kirkland, Washington?

Lewis: Within the last year, I think it was Kirkland, Washington. And now, recently, they've announced Phoenix, Arizona. So, one of the things you'll notice, and this is another one of the challenges they talked about was, those four cities, they all offer different advantages and disadvantages. Mountain View and Austin, it makes sense that those were the first cities that they launched in, because...

O'Reilly: That's where they are.

Lewis: Yeah. They're in Mountain View. And then, in Austin, in terms of engineering for a climate, relatively easy. You're not dealing with crazy road conditions, there's not going to be snow or a ton of rain, anything like that.

O'Reilly: Right, whereas in Seattle, you'll get a ton of rain.

Lewis: Exactly. You move to Kirkland, Washington, it's much more representative of what you'll see all four seasons. And with the most recent announcement of how they're going to be in Phoenix, they talked about how they want to see their components and all of their technology interacts with extreme heat and the idea of there being a lot of dust in the atmosphere, that kind of thing. So, they've been very tactical with the rollouts and the metropolitan areas they want to test in. And, I think that is, in a lot of ways, to stress test the technology, and put it in a position where they can succeed, but they can also reach the challenges of each relative climate.

O'Reilly: Got it. Before we move on Dylan, I wanted to point out to our listeners that the April Fool's Financial Literacy Month book giveaway is now over. Sorry. I think we're announcing the winners soon, or maybe they already did. The promotion at podcasts.fool.com, that is over, but at any point, if you're interested in more Foolish stock ideas, head over to focus.fool.com. For all of our listeners, there's a special offer for $129 for a full two-year subscription to The Motley Fool's Stock Advisor newsletter. I think we get that for free, so I read it anyways. I highly recommend it, folks.

So, cost of technology. We just got the release of -- well, it'll probably be driverless in a couple years, it's partially driverless now, but, the Tesla (TSLA -1.92%) Model 3. Big announcement last week, talking about costing $35,000-40,000. A Tesla Model S costs $80,000-90,000 upwards of a bit below six figures. How much does a Google car cost?

Lewis: I would have loved to get a specific number on this.

O'Reilly: They don't look expensive, do they?

Lewis: No. The idea here is -- and Urmson talked about this -- there's nothing really crazy, there's no unobtainium in what is powering these cars. The components are relatively cheap components. It's the act of research and development and all that that is adding cost to this project. And so, once it's ready for mainstream consumers, it should be priced in a relatively accessible manner. So, that's great news for you and me and the folks who would really benefit from it, like we talked about in the beginning of the show, maybe people who are a little bit older and don't feel comfortable driving, people who are sick, things like that. So, you have to be really encouraged by that. And I think that'll probably boost consumer demand for it.

O'Reilly: So, as I alluded to just now, a lot of the technology is already here, especially over at Tesla HQ. Dan Sparks had that one piece with the driverless capabilities. He owns a Tesla, and he verbally told me over our Slack instant messaging system, he actually said, "Yeah, I pretty much let it go 98% of the time."

Lewis: Yeah, that article was like, what, "I Rode 60-Something Miles in a Tesla Driverless Car?"

O'Reilly: Yeah. The very idea that it already came with the hardware, and just required the software, that's crazy. Anyways, it's already here, at least partially, with Tesla. When are we going to start seeing more from Google driverless cars?

Lewis: That's the other thing consumers are really honed in on. You see estimates, and some of them are like, "Three years!" Some of them are like, "20-30 years!" And part of that is, it depends on whether you're talking about initial introduction or mainstream adoption. But Urmson also talked about the idea of this being an incremental rollout. And much the same way that they have been strategic about their test cities, it might be possible that the technology comes out in climates that are more conducive to the technology, first.

O'Reilly: Is it going to be a little bit like how Google rolled out Fiber? Do you remember, they let a city win, and it was a big deal, when... who was it, who won? Oklahoma, Kansas City won. I remember, I was living in Grand Rapids, Michigan at the time, and there was talk of Grand Rapids winning, and it was like, "Oh, we have to win!" Do you think they'll do that sort of thing?

Lewis: I think it's more about the technology interacting with the climate. And, obviously, this is something that could be totally shaped by regulation as well. As individual states are willing to...

O'Reilly: They need to do Florida next for some hurricanes. (laughs) 

Lewis: Yeah. As individual states are willing to outline specific regulations, especially as it pertains to cars without steering wheels and brake pedals, which is what Google is developing, that'll push innovation and the rollout. But, I think, more than anything, has the technology been proven out in the type of conditions this area has.

O'Reilly: Got it. I noticed that they haven't really flirted too much with snow. I noticed, there was no Buffalo, New Yorks, Cleveland, Ohios, or Chicagos in there. (laughs) 

Lewis: I'm just waiting for them to come to the Northeast. I don't know if it'll ever happen.

O'Reilly: Game over, there's no way.

Lewis: And, I guess the cities in the Northeast present different challenges. The urban landscape, for a lot of East Coast cities, little bit tougher.

O'Reilly: Well, yeah, the streets are narrower, they were designs for horse and buggies, they're kind of...

Lewis: Yeah, there's a reason they haven't picked Boston yet.

O'Reilly: Yeah, there's a firm reason. "It's really hard to get a driverless car to drive on cobblestone, we don't know why!" So, everybody's question, obviously, is, what are Google's plans? Because, correct me if I'm wrong, Apple has been more forthcoming with their plans. They're like, "Yeah, we're totally going to sell a car, and it'll probably be coming out in 2019." What are Google's plans?

Lewis: Yeah. So, the party line for a long time with Google has been, "We don't want to make cars." Instead, they seem like they're focused on being the platform player partnering with the automotive industry incumbents, and leveraging the expertise that each of them have. So, Google will bring this excellent system and technology, the automotive companies will bring all of the manufacturing infrastructure and know-how there.

O'Reilly: Is this kind of like a Microsoft-PC relationship?

Lewis: Yeah, that's kind of a good way to think about. In the lead-up to CES, we thought we were going to get that announcement that we'd been waiting for. A couple different outlets had reported that Ford was going to announce their partnership with Google at the trade show, and details were kind of scarce, but we knew it would be non-exclusive. And, like I said, it was going to be, "Google has access to the tech side, Ford would be leveraging their manufacturing know-how." Kind of a win-win there. And I think while that didn't happen, that is probably still the way we're going to see this play out, just because Sergey Brin has said things in the past that have made it clear that they don't want to be making cars. Urmson said it again at South by Southwest.

Despite the fact that this deal fell through, you look at the advantages, and they're pretty clear. Ford gets access to this really great technology that they might struggle to develop in-house. Google doesn't have to spend billions of dollars setting up all the infrastructure to develop these cars and manufacture them. Because, they don't want to be setting up automotive plants. And, the systems side of business is much more in line with what Google does. And that's where their margin profile sits. They're not a hardware manufacturer, and they're not a car manufacturer.

O'Reilly: Right. I really wonder if what I posited is true, that's it's going to be a Microsoft-PC kind of relationship, and Google's just spreading the technology and the software. I wonder what that's worth. (laughs) 

Lewis: Yeah, it's tough to ballpark. But we'll see it play out.

O'Reilly: Cool. Thanks for your thoughts, Dylan. Have a good weekend! If you're a loyal listener and have questions or comments, we would love to hear from you, just email us at [email protected]. Again, that's [email protected]. As always, people on the program may have interests in the stocks they talk about, and The Motley Fool may have formal recommendations for or against those stocks, so don't buy or sell anything based solely on what you hear on this program. For Dylan Lewis, I am Sean O'Reilly, thanks for listening and Fool on!