1 00:00:04,280 --> 00:00:09,280 Speaker 1: From Bloomberg News and iHeartRadio. It's the big take. I'm 2 00:00:09,280 --> 00:00:21,079 Speaker 1: West Caasova today. Will driverless cars ever become a reality? 3 00:00:24,440 --> 00:00:28,080 Speaker 1: For years now, we've been hearing how driverless cars are 4 00:00:28,320 --> 00:00:31,920 Speaker 1: just around the corner. You hop in, maybe you take 5 00:00:31,920 --> 00:00:35,639 Speaker 1: a nap, and the car it's zips along in traffic 6 00:00:35,840 --> 00:00:40,239 Speaker 1: with no effort from you. Taxi service called Waymo is 7 00:00:40,240 --> 00:00:44,280 Speaker 1: coming to Los Angeles down the road. Level five autonomy 8 00:00:44,320 --> 00:00:47,680 Speaker 1: would mean fully driverless cars waiting on this car here 9 00:00:48,600 --> 00:00:54,320 Speaker 1: now it's clear and it's going, Oh my gosh. And 10 00:00:54,520 --> 00:00:59,880 Speaker 1: yet despite billion spent on prototypes and testing autonomous vehicle 11 00:01:00,240 --> 00:01:04,920 Speaker 1: on the open road, still appear years away. Why has 12 00:01:04,959 --> 00:01:08,880 Speaker 1: this been so elusive? Yeah, if you could just kind 13 00:01:08,920 --> 00:01:13,320 Speaker 1: of like get pedestrians to behave more reliably, and if 14 00:01:13,360 --> 00:01:17,240 Speaker 1: you could remove all the human driven cars from the road, 15 00:01:17,600 --> 00:01:20,319 Speaker 1: then you could have very safe, very effective driverless cars. 16 00:01:20,680 --> 00:01:25,120 Speaker 1: The problem is good luck. That's my Bloomberg colleague Max Chafkin. 17 00:01:25,560 --> 00:01:28,960 Speaker 1: He's been reporting on the push for driverless cars, and 18 00:01:29,080 --> 00:01:33,200 Speaker 1: he's here to talk about the persistent roadblocks still in 19 00:01:33,200 --> 00:01:37,360 Speaker 1: the way. First, though, in Insider's view of the problem, 20 00:01:38,240 --> 00:01:43,039 Speaker 1: Anthony Lewandowski pioneered the first driverless cars for Google and 21 00:01:43,080 --> 00:01:47,080 Speaker 1: then Uber, and he was an early evangelist for the technology, 22 00:01:47,680 --> 00:01:52,560 Speaker 1: but recently he's become skeptical that autonomous vehicles can actually 23 00:01:52,640 --> 00:01:56,200 Speaker 1: work in the real world, at least in the near future. 24 00:01:56,600 --> 00:01:59,480 Speaker 1: He's now taking his efforts in a different direction with 25 00:01:59,560 --> 00:02:05,480 Speaker 1: a new company called Pronto. Anthony, you were one of 26 00:02:05,600 --> 00:02:08,440 Speaker 1: the pioneers, I guess you could say, really the pioneer 27 00:02:08,800 --> 00:02:12,880 Speaker 1: of autonomous cars and the idea that cars were going 28 00:02:12,960 --> 00:02:15,240 Speaker 1: to be able to drive on the road by themselves 29 00:02:15,280 --> 00:02:20,040 Speaker 1: alongside cars being driven by people. When you first started 30 00:02:20,440 --> 00:02:23,160 Speaker 1: really looking into this and putting yourself into it, what 31 00:02:23,240 --> 00:02:25,480 Speaker 1: did you imagine? What was the world that you imagined 32 00:02:25,639 --> 00:02:28,120 Speaker 1: and how would it look to me? The world that 33 00:02:28,160 --> 00:02:32,680 Speaker 1: I imagine is one where the fatalities that folks are 34 00:02:32,960 --> 00:02:35,600 Speaker 1: suffering every year tens of thousands in the US, you know, 35 00:02:35,639 --> 00:02:39,680 Speaker 1: over a million worldwide, and the inconvenience you know is gone. 36 00:02:39,720 --> 00:02:42,840 Speaker 1: So I think of like the Jets sens where everything 37 00:02:42,840 --> 00:02:45,880 Speaker 1: that you want comes to you, and whenever you need 38 00:02:45,919 --> 00:02:48,560 Speaker 1: to go somewhere, there's this magic carpet that's just around 39 00:02:48,560 --> 00:02:51,280 Speaker 1: the corner. You hop in and you go there. There 40 00:02:51,360 --> 00:02:54,280 Speaker 1: wasn't a single moment where that just became true. But like, 41 00:02:54,960 --> 00:02:57,519 Speaker 1: the potential for this is huge, right, It's it's probably 42 00:02:57,560 --> 00:03:00,880 Speaker 1: bigger than the Internet. It's just a very complicated way 43 00:03:00,880 --> 00:03:03,600 Speaker 1: to get there. And you worked a long time. You 44 00:03:03,639 --> 00:03:07,280 Speaker 1: were at Google in their autonomous car program, which led 45 00:03:07,320 --> 00:03:09,400 Speaker 1: the way for a long time, spent a lot of money, 46 00:03:09,560 --> 00:03:13,200 Speaker 1: a lot of time working on it. And then ultimately, 47 00:03:13,240 --> 00:03:16,360 Speaker 1: if we flash forward till now, you've kind of lost 48 00:03:16,400 --> 00:03:19,520 Speaker 1: faith that autonomous cars in the way that you described 49 00:03:19,560 --> 00:03:22,320 Speaker 1: are going to happen. What change what made you realize 50 00:03:22,800 --> 00:03:25,560 Speaker 1: this wasn't going to work. If you were to take 51 00:03:25,600 --> 00:03:28,000 Speaker 1: a ride and a lot of demonstrations that are happening 52 00:03:28,000 --> 00:03:31,440 Speaker 1: by these companies, you would walk away from that and 53 00:03:31,480 --> 00:03:34,280 Speaker 1: be like, oh my goodness, this is like happening next week. 54 00:03:34,320 --> 00:03:37,280 Speaker 1: I didn't see this coming. I need to completely change 55 00:03:37,320 --> 00:03:39,360 Speaker 1: my business. I need to completely changed what's going on. 56 00:03:39,840 --> 00:03:42,840 Speaker 1: And so there's a little bit of euphoria that people 57 00:03:42,880 --> 00:03:46,680 Speaker 1: have when they experienced these really magical demonstrations. And the 58 00:03:46,720 --> 00:03:49,160 Speaker 1: reality of that is that you could be very good 59 00:03:49,200 --> 00:03:51,160 Speaker 1: on this one route in this one area, but if 60 00:03:51,160 --> 00:03:52,880 Speaker 1: you were to just move to a new city, it 61 00:03:53,040 --> 00:03:56,840 Speaker 1: just couldn't drive one hundred feet down the road outside 62 00:03:56,840 --> 00:03:59,360 Speaker 1: of your house. And so what we're doing now is 63 00:03:59,440 --> 00:04:03,880 Speaker 1: not learning to drive. We're really perfecting this demonstration that 64 00:04:03,960 --> 00:04:05,960 Speaker 1: we've been doing for the last ten years and making 65 00:04:06,000 --> 00:04:09,280 Speaker 1: that more and more usable as opposed to really knowing 66 00:04:09,280 --> 00:04:11,680 Speaker 1: how to drive. And it just works everywhere. Well, let's 67 00:04:11,680 --> 00:04:15,240 Speaker 1: explain a little bit why that is. Can you tell 68 00:04:15,320 --> 00:04:19,240 Speaker 1: us kind of just saying Layman's terms, how a driverless 69 00:04:19,240 --> 00:04:23,800 Speaker 1: car an autonomous vehicle actually works. There's sensors on there 70 00:04:24,120 --> 00:04:25,440 Speaker 1: that to kind of tell you where you are to 71 00:04:25,480 --> 00:04:27,760 Speaker 1: start off in the world. You think of this as 72 00:04:27,760 --> 00:04:29,920 Speaker 1: like a GPS. Then there's a sensors to tell you 73 00:04:30,000 --> 00:04:32,400 Speaker 1: what's going on around you. You can think of cameras 74 00:04:32,400 --> 00:04:36,360 Speaker 1: and lasers and radars and those two kind of two things. 75 00:04:36,400 --> 00:04:38,880 Speaker 1: One is they help you like refine where you are 76 00:04:38,880 --> 00:04:40,880 Speaker 1: on the map, because the GPS might tell you you're 77 00:04:40,960 --> 00:04:44,640 Speaker 1: at like Main Street and Second Avenue, and then the 78 00:04:44,720 --> 00:04:46,440 Speaker 1: camera and the laders around you're telling you're like, oh, 79 00:04:46,440 --> 00:04:49,120 Speaker 1: I'm endling two and I'm two cars back from the lights, 80 00:04:49,160 --> 00:04:50,720 Speaker 1: and the lights you know, read and so forth, so 81 00:04:50,760 --> 00:04:52,800 Speaker 1: you have a little bit more information there. Then you 82 00:04:52,800 --> 00:04:54,919 Speaker 1: have the kind of the computers on board which have 83 00:04:55,040 --> 00:04:58,160 Speaker 1: like many many CPUs, tons of GPUs, like a lot 84 00:04:58,160 --> 00:05:00,880 Speaker 1: of computing processing and wheels. There's a software component which 85 00:05:00,920 --> 00:05:02,800 Speaker 1: is really where all the magic, in my mind happens 86 00:05:02,800 --> 00:05:05,000 Speaker 1: and still needs to happen. And then there's a drive 87 00:05:05,000 --> 00:05:07,960 Speaker 1: by wirepiece. So after the sensors have been processed and 88 00:05:08,520 --> 00:05:10,920 Speaker 1: the destination has been entered, and the vehicles trying to 89 00:05:10,960 --> 00:05:13,920 Speaker 1: navigate there and needs to physically control the steering wheel 90 00:05:13,960 --> 00:05:16,880 Speaker 1: that break, the gas, the windshield wipers, the lights and 91 00:05:16,920 --> 00:05:19,120 Speaker 1: all this other stuff, and that's kind of the interface there. 92 00:05:19,800 --> 00:05:23,320 Speaker 1: You say, the software is where the magic happens. Has 93 00:05:23,440 --> 00:05:26,440 Speaker 1: that been the really difficult thing to get right? Absolutely? 94 00:05:26,480 --> 00:05:28,560 Speaker 1: So a lot of people like to romanticize and be like, 95 00:05:28,600 --> 00:05:31,240 Speaker 1: if only my camera could see further or my laser 96 00:05:31,240 --> 00:05:33,520 Speaker 1: could see further, then I can just solve the problem. 97 00:05:33,920 --> 00:05:36,240 Speaker 1: And in reality that's just not the case. Right even 98 00:05:36,279 --> 00:05:40,320 Speaker 1: almost it's like day one, in like ten years plus ago. 99 00:05:40,760 --> 00:05:43,160 Speaker 1: All of the places where you know the software made 100 00:05:43,200 --> 00:05:46,479 Speaker 1: a mistake, it was never because they couldn't see the thing. 101 00:05:47,320 --> 00:05:49,960 Speaker 1: It just couldn't interpret the sensor data that was coming 102 00:05:49,960 --> 00:05:53,000 Speaker 1: to it correctly. And so the software is where the 103 00:05:53,080 --> 00:05:56,040 Speaker 1: actual value in the self driving cars comes from. And 104 00:05:56,320 --> 00:06:02,120 Speaker 1: that's what's missing, And is that because a lot of 105 00:06:02,160 --> 00:06:05,919 Speaker 1: AI is based on lots and lots of scenarios that 106 00:06:05,960 --> 00:06:08,400 Speaker 1: a computer has to fall back on to say, this 107 00:06:08,440 --> 00:06:10,640 Speaker 1: looks like something else and in that case, I should 108 00:06:10,680 --> 00:06:13,560 Speaker 1: behave in this way. And yet the number of scenarios 109 00:06:13,560 --> 00:06:16,160 Speaker 1: are kind of end listen, they just aren't enough of 110 00:06:16,200 --> 00:06:19,320 Speaker 1: those examples for the computer to be able to reference. 111 00:06:19,839 --> 00:06:22,800 Speaker 1: It's obviously very complicated, but the way to think about 112 00:06:22,800 --> 00:06:26,040 Speaker 1: it is like there are more like possibilities for your 113 00:06:26,120 --> 00:06:29,120 Speaker 1: drive from home to work them. There are like atoms 114 00:06:29,120 --> 00:06:31,280 Speaker 1: in the universe, right there could be like this leaf 115 00:06:31,279 --> 00:06:32,880 Speaker 1: could have fallen, and this bird could have been here, 116 00:06:32,920 --> 00:06:34,200 Speaker 1: and the you know, the road could have been here 117 00:06:34,200 --> 00:06:37,480 Speaker 1: in the sun. There's just so many permutations, and so 118 00:06:37,520 --> 00:06:40,440 Speaker 1: the question is about generalization and how can you build 119 00:06:41,120 --> 00:06:45,920 Speaker 1: AI software that can really process these and categorize them 120 00:06:45,920 --> 00:06:49,000 Speaker 1: correctly and then handle them correctly. On that and we're 121 00:06:49,040 --> 00:06:51,880 Speaker 1: pretty good for better than a person at finding lanes 122 00:06:51,880 --> 00:06:56,360 Speaker 1: and where you can drive, detecting vehicles detecting people because 123 00:06:56,400 --> 00:06:59,680 Speaker 1: we've had so many instances of like these are cars 124 00:06:59,760 --> 00:07:02,200 Speaker 1: drive on the street and some person somewhere drew a 125 00:07:02,200 --> 00:07:04,039 Speaker 1: little box around it said like this is a car 126 00:07:04,480 --> 00:07:06,599 Speaker 1: and this is a person, and we have like tens 127 00:07:06,600 --> 00:07:08,559 Speaker 1: and tens of millions of these there, so we actually 128 00:07:08,560 --> 00:07:12,080 Speaker 1: have enough generalization there, but all the interactions are where 129 00:07:12,080 --> 00:07:14,400 Speaker 1: the difficulty is, and those are harder to do because 130 00:07:15,160 --> 00:07:17,320 Speaker 1: unlike when you just drive by and you see everything 131 00:07:17,320 --> 00:07:20,320 Speaker 1: around you, you're labeling everything, you're only doing a very 132 00:07:20,320 --> 00:07:23,560 Speaker 1: few interactions per drive. And you know, we wouldn't want 133 00:07:23,600 --> 00:07:25,360 Speaker 1: to have the cars just trying and crashing all the 134 00:07:25,360 --> 00:07:26,880 Speaker 1: time and be like, oh no, that didn't work. Oh 135 00:07:26,960 --> 00:07:28,600 Speaker 1: let's try it again. Let's try it again, right, So 136 00:07:28,920 --> 00:07:30,400 Speaker 1: it's a lot more tricky to do that, and some 137 00:07:30,480 --> 00:07:32,840 Speaker 1: people have simulators of course to do it. But that's 138 00:07:32,880 --> 00:07:36,960 Speaker 1: where the limitations are falling short from being just pressed 139 00:07:37,000 --> 00:07:39,080 Speaker 1: the button and let the computer figure it out. Can 140 00:07:39,120 --> 00:07:41,560 Speaker 1: I ask you about one real world scenario that I 141 00:07:41,640 --> 00:07:47,200 Speaker 1: found fascinating is that some autonomous vehicles have difficulty just 142 00:07:47,440 --> 00:07:52,280 Speaker 1: making a left turn. Yes, so there's an industry lingo 143 00:07:52,360 --> 00:07:57,040 Speaker 1: they call UPL unprotected left turn. And the issue there is, 144 00:07:57,320 --> 00:08:00,440 Speaker 1: you know, the software is running, isn't comfortable estimating knowing 145 00:08:00,880 --> 00:08:03,960 Speaker 1: whether or not the uncoming traffic we'll see them and 146 00:08:04,000 --> 00:08:07,040 Speaker 1: stop for them, or will they wait for it to 147 00:08:07,080 --> 00:08:10,040 Speaker 1: be super clear so it's not possible for them to 148 00:08:10,120 --> 00:08:12,200 Speaker 1: get in an accident, and then they'll make the left turn. 149 00:08:12,200 --> 00:08:15,160 Speaker 1: And so sometimes as a human you're like, all right, 150 00:08:15,240 --> 00:08:17,360 Speaker 1: we're in rush hour jam here, I need to like 151 00:08:17,400 --> 00:08:20,239 Speaker 1: assert myself if that person were to just keep driving 152 00:08:20,320 --> 00:08:23,760 Speaker 1: straight even though they have the right away, because we're 153 00:08:23,800 --> 00:08:25,520 Speaker 1: in a traffic condition and I'm just jamming in, they're 154 00:08:25,520 --> 00:08:27,000 Speaker 1: going to slow down a little bit and let me go. 155 00:08:27,880 --> 00:08:30,440 Speaker 1: But if you're behind one of these Automas vehicles, it's 156 00:08:30,480 --> 00:08:32,439 Speaker 1: not able to do that kind of interaction, and so 157 00:08:32,440 --> 00:08:34,160 Speaker 1: I'm just gonna wait for it to be open enough 158 00:08:34,200 --> 00:08:35,800 Speaker 1: for you to go. And if you're the person behind that, 159 00:08:35,920 --> 00:08:38,520 Speaker 1: you're like honking the horn. You're like, come on, let's go. 160 00:08:38,640 --> 00:08:40,280 Speaker 1: I can do this, you know, And so it just 161 00:08:40,320 --> 00:08:43,640 Speaker 1: gets frustrating on that level. And while it's frustrating for 162 00:08:43,679 --> 00:08:47,000 Speaker 1: that person, it's just like, have patience for the Automas vehicle. 163 00:08:47,040 --> 00:08:48,720 Speaker 1: They're trying really hard. They just don't know, you know, 164 00:08:48,760 --> 00:08:50,880 Speaker 1: they wanted to be super safe, right, so they're kind 165 00:08:50,880 --> 00:08:54,040 Speaker 1: of more on the train or airplane model where they 166 00:08:54,120 --> 00:08:56,719 Speaker 1: really just want to drive where they know it's going 167 00:08:56,720 --> 00:08:59,000 Speaker 1: to be safe. And the way we drive is we're 168 00:08:59,120 --> 00:09:03,520 Speaker 1: constantly flirting with like comfortable areas of like is this 169 00:09:03,559 --> 00:09:05,160 Speaker 1: person going to go or not? And if you start 170 00:09:05,200 --> 00:09:07,640 Speaker 1: going you see that person not starting in, you're gonna 171 00:09:07,840 --> 00:09:09,560 Speaker 1: maybe hit the brakes and stop and let them go 172 00:09:09,640 --> 00:09:11,160 Speaker 1: and then be like wow, that was close. Or you're 173 00:09:11,160 --> 00:09:14,120 Speaker 1: gonna jam on the gas and squeeze it out. That's 174 00:09:14,120 --> 00:09:16,560 Speaker 1: just that's something that you're gonna want the machine to 175 00:09:16,600 --> 00:09:19,480 Speaker 1: do right, And so it's hard to get that level 176 00:09:19,480 --> 00:09:24,319 Speaker 1: of fidelity. I think you've said something like autonomous cars 177 00:09:24,360 --> 00:09:28,000 Speaker 1: are always five years away. There five years away since 178 00:09:28,240 --> 00:09:31,120 Speaker 1: I think two thousand hand at least, you know, and 179 00:09:31,160 --> 00:09:34,880 Speaker 1: then probably even earlier than that. This is not like, 180 00:09:35,000 --> 00:09:38,000 Speaker 1: oh it's impossible. This is not an impossible task. It's 181 00:09:38,000 --> 00:09:41,439 Speaker 1: just a very very difficult task. And my point is 182 00:09:41,440 --> 00:09:43,719 Speaker 1: that we just don't know how close we are, right 183 00:09:43,720 --> 00:09:49,359 Speaker 1: and so we're we're showing amazing feats of achievements and demonstrations, 184 00:09:50,040 --> 00:09:51,720 Speaker 1: but we don't know. We can't say like, oh, I'm 185 00:09:51,800 --> 00:09:55,440 Speaker 1: ninety seven point two percent done, Anthony. Other companies are 186 00:09:55,480 --> 00:09:58,920 Speaker 1: still investing in autonomous cars that can drive on the 187 00:09:58,960 --> 00:10:03,080 Speaker 1: open road, and your skepticism about that is a bit 188 00:10:03,120 --> 00:10:06,840 Speaker 1: controversial because you're a prominent person in the industry and 189 00:10:06,920 --> 00:10:11,679 Speaker 1: you still have skin in the game, but also because 190 00:10:12,120 --> 00:10:15,640 Speaker 1: you started at Google, you went to Uber, and then 191 00:10:15,800 --> 00:10:19,560 Speaker 1: Google said you took proprietary information with you, and you 192 00:10:19,640 --> 00:10:23,760 Speaker 1: pleaded guilty to one count of stealing trade secrets, and 193 00:10:23,760 --> 00:10:27,520 Speaker 1: you were spared an eighteen month prison sentence when President 194 00:10:27,640 --> 00:10:32,200 Speaker 1: Trump pardoned you. When you look back on all of that, 195 00:10:32,400 --> 00:10:35,960 Speaker 1: what do you take away from it that informs what 196 00:10:36,040 --> 00:10:38,520 Speaker 1: you're doing now? That's a really good question. I mean, 197 00:10:38,520 --> 00:10:40,240 Speaker 1: it kind of took me under scores a little bit 198 00:10:40,240 --> 00:10:42,400 Speaker 1: that the importance of the work they were doing. I mean, 199 00:10:42,400 --> 00:10:45,640 Speaker 1: I'm really focused on building something that has value, and 200 00:10:45,800 --> 00:10:49,199 Speaker 1: to have such big companies fighting over what ended up 201 00:10:49,280 --> 00:10:53,640 Speaker 1: to be nothing is sumbering. It means that people really 202 00:10:53,679 --> 00:10:57,880 Speaker 1: care about it. Right in two sixteen, many companies were 203 00:10:57,880 --> 00:11:01,120 Speaker 1: claiming that this was going to be tens of billions 204 00:11:01,120 --> 00:11:04,080 Speaker 1: of dollars a year of revenue next year, and then 205 00:11:04,080 --> 00:11:07,000 Speaker 1: even more and more, and we're it's what like almost 206 00:11:07,160 --> 00:11:08,880 Speaker 1: ten years maybe not to say it's it's like over 207 00:11:08,960 --> 00:11:11,880 Speaker 1: five years from now, and the revenue is like not 208 00:11:11,920 --> 00:11:14,160 Speaker 1: even one hundredth of a percent of the investment that 209 00:11:14,200 --> 00:11:17,440 Speaker 1: has gone you know, in the entire industry, and so 210 00:11:17,480 --> 00:11:19,680 Speaker 1: that's obviously going to change over time, but it's just 211 00:11:19,679 --> 00:11:23,319 Speaker 1: gonna take a lot longer. Everybody's excited about the future 212 00:11:23,320 --> 00:11:25,560 Speaker 1: and they wanted to happen, it's just very, very difficult. 213 00:11:25,880 --> 00:11:27,679 Speaker 1: And so what I kind of took from from all 214 00:11:27,720 --> 00:11:29,960 Speaker 1: of that is really like focusing on what I'm good at, 215 00:11:30,400 --> 00:11:34,600 Speaker 1: just product market fit and building something that's useful today 216 00:11:34,640 --> 00:11:37,320 Speaker 1: with the technology that we have on hand. Now our 217 00:11:37,400 --> 00:11:50,079 Speaker 1: conversation continues after the break, So now you've decided to 218 00:11:50,120 --> 00:11:53,920 Speaker 1: take this in a completely different direction with your company, Pronto, 219 00:11:54,240 --> 00:11:58,320 Speaker 1: Can you describe what you're doing now? So Pronto right 220 00:11:58,320 --> 00:12:03,439 Speaker 1: now does off road autonomy and so we do construction 221 00:12:03,640 --> 00:12:07,640 Speaker 1: and mining truck automation, and so we move material from 222 00:12:07,640 --> 00:12:10,520 Speaker 1: point A to point B, and then we go back 223 00:12:10,559 --> 00:12:12,360 Speaker 1: to point A, pick up more material, and come back 224 00:12:12,400 --> 00:12:14,839 Speaker 1: to point B and dump in and do that all 225 00:12:14,920 --> 00:12:19,560 Speaker 1: day long. So Pronta, we're building autonomous vehicles for driving 226 00:12:19,600 --> 00:12:23,640 Speaker 1: off road on a closed course. Unlike on the open 227 00:12:23,720 --> 00:12:28,120 Speaker 1: road where you can drive anywhere and anybody can drive 228 00:12:28,280 --> 00:12:31,640 Speaker 1: on your course private property, it's closed course, so you 229 00:12:31,640 --> 00:12:34,800 Speaker 1: can control who comes in and who gets to drive 230 00:12:34,800 --> 00:12:37,800 Speaker 1: in that area. So how many customers do you have? 231 00:12:37,840 --> 00:12:41,040 Speaker 1: How many of these closed courses for different kinds of 232 00:12:41,040 --> 00:12:44,880 Speaker 1: industries are you now operating? Yeah, so we're almost at 233 00:12:44,920 --> 00:12:49,040 Speaker 1: a dozen customers now and the exact number of sites 234 00:12:49,040 --> 00:12:52,160 Speaker 1: we're not disclosing. But it's great to see that when 235 00:12:52,160 --> 00:12:54,679 Speaker 1: you have a product market fit, that there's a demand 236 00:12:54,760 --> 00:12:58,439 Speaker 1: for more of the same thing that you're making. So 237 00:12:58,480 --> 00:13:01,040 Speaker 1: this is a business that operates on its own now. 238 00:13:01,080 --> 00:13:03,959 Speaker 1: But do you see this as a means to an end, 239 00:13:04,040 --> 00:13:06,920 Speaker 1: like do you see this as one step along the 240 00:13:06,960 --> 00:13:09,680 Speaker 1: way to achieving the ultimate goal of autonomous cars on 241 00:13:09,679 --> 00:13:13,280 Speaker 1: the open road? Our objective is to automate everything that 242 00:13:13,320 --> 00:13:16,200 Speaker 1: has wheels, But we're starting with the off road mining 243 00:13:16,240 --> 00:13:18,800 Speaker 1: market because that's something that we can do today, and 244 00:13:18,840 --> 00:13:21,480 Speaker 1: that's something that we can deliver to customers, and you know, 245 00:13:21,520 --> 00:13:24,640 Speaker 1: there's more demand there than we can handle for a 246 00:13:24,720 --> 00:13:27,560 Speaker 1: company our site, So that's a great fit to start there. 247 00:13:28,000 --> 00:13:30,840 Speaker 1: But ultimately I think that the technology will scale and 248 00:13:30,880 --> 00:13:35,360 Speaker 1: grow into other applications. Do you think that we'll ever 249 00:13:35,520 --> 00:13:38,520 Speaker 1: have a time when we can have purely autonomous cars 250 00:13:38,520 --> 00:13:41,240 Speaker 1: on the open road or do you think that the 251 00:13:41,360 --> 00:13:45,400 Speaker 1: human factor, the ability to make those decisions based on 252 00:13:45,520 --> 00:13:49,720 Speaker 1: feel and experience, will always be required in order for 253 00:13:50,080 --> 00:13:54,040 Speaker 1: cars to operate safely. Yeah, that's a great question. A 254 00:13:54,040 --> 00:13:57,040 Speaker 1: lot of people ask that over the years, and what 255 00:13:57,160 --> 00:14:00,520 Speaker 1: we've seen is that you could think, like a stop signing, 256 00:14:00,520 --> 00:14:01,800 Speaker 1: you look at the other person and when they're going 257 00:14:01,880 --> 00:14:04,000 Speaker 1: to go or when you're gonna go, And so we 258 00:14:04,080 --> 00:14:06,199 Speaker 1: initially thought that there was going to be a need 259 00:14:06,240 --> 00:14:09,160 Speaker 1: for some sort of human interaction, but it turns out 260 00:14:09,200 --> 00:14:10,880 Speaker 1: that even the ones that have the most kind of 261 00:14:11,480 --> 00:14:14,439 Speaker 1: are you going, am I going? And then hesitating, those 262 00:14:14,880 --> 00:14:17,920 Speaker 1: don't seem to be true. So we really are just 263 00:14:18,000 --> 00:14:21,560 Speaker 1: like misinterpreting the signals and not so much needing to 264 00:14:21,640 --> 00:14:25,800 Speaker 1: have that humanness be added to the system to be 265 00:14:25,840 --> 00:14:28,240 Speaker 1: able to drive. So I do think there will be 266 00:14:29,000 --> 00:14:31,840 Speaker 1: fully driverless with other vehicles that are different by people 267 00:14:32,240 --> 00:14:35,440 Speaker 1: happening at some point in time. It's just not with 268 00:14:35,480 --> 00:14:37,840 Speaker 1: the technology that we haven't available today. We started this 269 00:14:37,880 --> 00:14:41,080 Speaker 1: conversation with you saying you think we're gonna get there eventually, 270 00:14:41,520 --> 00:14:44,600 Speaker 1: and then talking about how like it's always five years 271 00:14:44,640 --> 00:14:46,960 Speaker 1: away forever. What do you see when you look down 272 00:14:47,000 --> 00:14:49,320 Speaker 1: the road, like do you think that we are going 273 00:14:49,360 --> 00:14:53,960 Speaker 1: to have autonomous cars in normal life? I don't know. 274 00:14:54,040 --> 00:14:58,200 Speaker 1: Ten years, twenty years from now, yeah, I mean the 275 00:14:58,280 --> 00:15:00,000 Speaker 1: answer is like, I don't know, and I don't think 276 00:15:00,120 --> 00:15:04,240 Speaker 1: anybody knows, right, And so is it five years? Probably not? 277 00:15:05,120 --> 00:15:07,920 Speaker 1: Can it happen in ten? It probably can it? We 278 00:15:08,200 --> 00:15:11,840 Speaker 1: might take more than that, probably too. We just don't know. 279 00:15:11,960 --> 00:15:15,440 Speaker 1: There's there's the fundamental breakthroughs that are needed. Obviously AI 280 00:15:15,600 --> 00:15:19,080 Speaker 1: is making fantastic strides so that the speed at which 281 00:15:19,080 --> 00:15:23,400 Speaker 1: things could happen is almost impossible to predict. But we've 282 00:15:23,400 --> 00:15:26,240 Speaker 1: been doing this for ten plus years. There's thousands of 283 00:15:26,280 --> 00:15:30,640 Speaker 1: engineers working on this, and we're still somewhere really far away. 284 00:15:30,640 --> 00:15:33,440 Speaker 1: It's a very complicated thing to the point. It's also 285 00:15:33,800 --> 00:15:37,320 Speaker 1: the consequences of doing it prematurely is very high, right, 286 00:15:37,360 --> 00:15:40,000 Speaker 1: And so you can't just go and throw the app 287 00:15:40,040 --> 00:15:42,640 Speaker 1: out into the world and see if people work, and 288 00:15:42,640 --> 00:15:44,120 Speaker 1: then you fix the bugs and you keep fixing the 289 00:15:44,160 --> 00:15:46,600 Speaker 1: bugs for the next ten years as you're perfecting it. 290 00:15:47,480 --> 00:15:51,120 Speaker 1: When this doesn't work, it has like real life consequences 291 00:15:51,160 --> 00:15:54,400 Speaker 1: for the people involved. And so the speed is not 292 00:15:54,680 --> 00:15:58,000 Speaker 1: just because it's a difficult problem. It's also because everybody's 293 00:15:58,040 --> 00:16:01,800 Speaker 1: taking a lot of caution and care safety and building 294 00:16:01,880 --> 00:16:04,840 Speaker 1: something that works and isn't going to cause him to 295 00:16:04,840 --> 00:16:07,640 Speaker 1: do or extra harm just because it's out in the world. 296 00:16:08,680 --> 00:16:12,120 Speaker 1: Anthony Lewandowski, thanks for talking with me. Thank you as 297 00:16:12,120 --> 00:16:17,960 Speaker 1: a pleasure, Thanks for having on now has promised. Max 298 00:16:18,120 --> 00:16:21,800 Speaker 1: Chafkin is here. He's been reporting the ups and downs 299 00:16:21,840 --> 00:16:25,040 Speaker 1: of driverless cars, including in a big story for Bloomberg 300 00:16:25,080 --> 00:16:30,320 Speaker 1: Business Week that digs into why advances have been so slow. Max, 301 00:16:30,360 --> 00:16:33,280 Speaker 1: I was just talking to Anthony Lewandowski is one of 302 00:16:33,280 --> 00:16:36,080 Speaker 1: the pioneers of autonomous cars and really saw a bright 303 00:16:36,160 --> 00:16:38,640 Speaker 1: future for them, but is kind of reconsidered and now 304 00:16:38,760 --> 00:16:41,400 Speaker 1: is doing it in a much more limited form with 305 00:16:41,480 --> 00:16:45,040 Speaker 1: his new company. How right is he? So it's important 306 00:16:45,040 --> 00:16:48,840 Speaker 1: to say that Anthony Lewandowski is one of the key pioneers, 307 00:16:48,880 --> 00:16:52,640 Speaker 1: if not you know, the pioneer in terms of commercializing 308 00:16:52,840 --> 00:16:56,680 Speaker 1: self driving cars. He is really the guy who created 309 00:16:56,960 --> 00:17:00,000 Speaker 1: the Google self driving car, the very first Google self 310 00:17:00,080 --> 00:17:02,280 Speaker 1: driving car, which became waymow and which kind of kicked 311 00:17:02,280 --> 00:17:06,840 Speaker 1: off this entire industry. He's also probably the most controversial 312 00:17:06,920 --> 00:17:11,200 Speaker 1: person in this industry because, you know, after leaving Google 313 00:17:11,280 --> 00:17:14,760 Speaker 1: and going to uber. Google accused him of trade secrets theft, 314 00:17:14,840 --> 00:17:17,399 Speaker 1: and he was eventually charged criminally. So this is a 315 00:17:17,440 --> 00:17:20,919 Speaker 1: guy who has sort of pissed off all the major 316 00:17:21,000 --> 00:17:25,120 Speaker 1: players in the industry. Now, my sense, from having reported 317 00:17:25,119 --> 00:17:27,400 Speaker 1: the story and spend time, you know, really on both 318 00:17:27,400 --> 00:17:30,199 Speaker 1: sides of the story over the last few years, is 319 00:17:30,240 --> 00:17:33,240 Speaker 1: that he's worth listening to and that the things that 320 00:17:33,240 --> 00:17:35,679 Speaker 1: he's saying, the criticisms that he's making of the industry 321 00:17:36,000 --> 00:17:38,400 Speaker 1: are being borne out by the progress of the industry, 322 00:17:38,440 --> 00:17:42,520 Speaker 1: which we're seeing basically, after many years really of promises 323 00:17:42,560 --> 00:17:45,680 Speaker 1: about self driving cars, that there hasn't been a whole 324 00:17:45,760 --> 00:17:48,520 Speaker 1: lot of progress. And that's what Anthony Lewandowski is pointing to. 325 00:17:48,880 --> 00:17:51,240 Speaker 1: One of the things you've written about in covering this 326 00:17:51,600 --> 00:17:54,720 Speaker 1: is how important mapping is and the billions of dollars 327 00:17:54,720 --> 00:17:57,760 Speaker 1: that have been put into try to develop maps that 328 00:17:58,000 --> 00:18:02,479 Speaker 1: cars can reliably use. Yeah, exactly, and that is how 329 00:18:02,720 --> 00:18:05,880 Speaker 1: whole driverless car things started. Actually it's hard to remember, 330 00:18:05,960 --> 00:18:09,280 Speaker 1: but early on Google Maps started this program called street 331 00:18:09,320 --> 00:18:12,520 Speaker 1: View where they were driving cars around taking pictures of 332 00:18:12,600 --> 00:18:15,720 Speaker 1: houses and of the roads. And that same team in 333 00:18:15,800 --> 00:18:19,000 Speaker 1: which Lewandowski was on and helped develop technology for ended 334 00:18:19,040 --> 00:18:23,200 Speaker 1: up building the early driverless cars that Google launched around 335 00:18:23,240 --> 00:18:27,440 Speaker 1: twenty ten, just before twenty ten, and with mapping, the 336 00:18:27,480 --> 00:18:29,399 Speaker 1: idea here is you want to know every corner of 337 00:18:29,440 --> 00:18:31,920 Speaker 1: the road so that the computer can make decisions. And 338 00:18:31,960 --> 00:18:34,920 Speaker 1: then there's the sort of question of human behavior right 339 00:18:34,920 --> 00:18:37,640 Speaker 1: where even if you know all the things of the road, 340 00:18:37,680 --> 00:18:39,800 Speaker 1: you'd actually don't know what other cars are going to do. 341 00:18:40,040 --> 00:18:43,080 Speaker 1: And that's kind of where Lewandowski's new thing sort of 342 00:18:43,160 --> 00:18:45,520 Speaker 1: tries to solve the problem where essentially, because they're on 343 00:18:45,560 --> 00:18:49,240 Speaker 1: an industrial site, because it's like a closed course, people 344 00:18:49,240 --> 00:18:51,480 Speaker 1: are not allowed to walk in front of the machines 345 00:18:52,040 --> 00:18:54,520 Speaker 1: or really even be there at all. All you really 346 00:18:54,560 --> 00:18:56,800 Speaker 1: need to do is create a really good map and 347 00:18:56,920 --> 00:18:59,639 Speaker 1: essentially program the cars, and it turns them into something 348 00:19:00,200 --> 00:19:03,680 Speaker 1: sort of more akin to like industrial robots rather than 349 00:19:04,160 --> 00:19:07,800 Speaker 1: these like fully featured robo drivers that can go anywhere, 350 00:19:07,880 --> 00:19:12,040 Speaker 1: can react to anything, or what have you. It's part 351 00:19:12,040 --> 00:19:15,360 Speaker 1: of the tension here between having driverless cars doing their 352 00:19:15,400 --> 00:19:17,840 Speaker 1: thing next to people who are driving cars and the 353 00:19:17,880 --> 00:19:19,959 Speaker 1: two don't mix that if you had a system that 354 00:19:20,080 --> 00:19:24,160 Speaker 1: was entirely driverless cars, that you could create a system 355 00:19:24,200 --> 00:19:28,159 Speaker 1: where you would take out some of the other variables exactly. 356 00:19:28,240 --> 00:19:30,680 Speaker 1: And I think we've actually for the last few years 357 00:19:30,680 --> 00:19:33,280 Speaker 1: that's been one of the kind of lines that you 358 00:19:33,359 --> 00:19:36,800 Speaker 1: hear from the people who are most enthusiastic about driverless cars, 359 00:19:36,800 --> 00:19:38,960 Speaker 1: which is that, Yeah, if you could just kind of 360 00:19:39,000 --> 00:19:43,159 Speaker 1: like get pedestrians to behave more reliably, and if you 361 00:19:43,160 --> 00:19:46,920 Speaker 1: could remove all the human driven cars from the road, 362 00:19:47,280 --> 00:19:50,000 Speaker 1: then you could have very safe, very effective driverless cars. 363 00:19:50,200 --> 00:19:54,359 Speaker 1: The problem is, good luck, good luck removing pedestrians from 364 00:19:54,400 --> 00:19:56,160 Speaker 1: our streets. I don't think any of us really want 365 00:19:56,200 --> 00:19:58,040 Speaker 1: to do that. I'm not sure that's society a city 366 00:19:58,119 --> 00:20:01,080 Speaker 1: that I want to live in. And good luck getting 367 00:20:01,119 --> 00:20:04,240 Speaker 1: rid of human drivers, because the problem with the auto industry, 368 00:20:04,320 --> 00:20:05,960 Speaker 1: right is that people tend to keep their cars for 369 00:20:06,000 --> 00:20:09,120 Speaker 1: a really long time, and there isn't like some magical 370 00:20:09,160 --> 00:20:12,080 Speaker 1: way you can swap out all the vehicles. If you 371 00:20:12,119 --> 00:20:14,840 Speaker 1: could do that, then yeah, we'd have trains basically, and 372 00:20:15,119 --> 00:20:17,280 Speaker 1: it would be great, and they'd be super safe and 373 00:20:17,720 --> 00:20:20,960 Speaker 1: reliable and probably a lot of the kind of crazy 374 00:20:21,119 --> 00:20:24,200 Speaker 1: futuristic things that the urban planning types that get excited 375 00:20:24,200 --> 00:20:26,399 Speaker 1: about this stuff could happen. Right, You could have fewer 376 00:20:26,440 --> 00:20:28,720 Speaker 1: parking spaces, you could have the lower levels of car 377 00:20:28,760 --> 00:20:31,159 Speaker 1: ownership and so on. It's just this is something that 378 00:20:31,280 --> 00:20:34,440 Speaker 1: is going to happen over kind of like a generational scale, 379 00:20:34,640 --> 00:20:37,159 Speaker 1: rather than something that's going to happen in years or 380 00:20:37,200 --> 00:20:48,479 Speaker 1: even decades. We'll be right back, Max. I think a 381 00:20:48,480 --> 00:20:50,520 Speaker 1: lot of the population might not like the idea of 382 00:20:50,520 --> 00:20:52,880 Speaker 1: a robot driving them around. They kind of want to 383 00:20:52,920 --> 00:20:57,040 Speaker 1: be in charge of their cars. Yeah. Absolutely, although I 384 00:20:57,080 --> 00:20:59,920 Speaker 1: think what the driverless car boosters would say is like, oh, 385 00:21:00,080 --> 00:21:02,440 Speaker 1: once you get used to not driving or being able 386 00:21:02,440 --> 00:21:03,919 Speaker 1: to take a nap in your car, then maybe you'll 387 00:21:03,960 --> 00:21:07,520 Speaker 1: feel differently. That said, we do have it's important to say. Right. 388 00:21:07,560 --> 00:21:11,520 Speaker 1: They're two key companies, Weymo and cruise Waymo's the Google 389 00:21:11,560 --> 00:21:14,239 Speaker 1: one cruises the GM one, and they are operating to 390 00:21:14,280 --> 00:21:18,360 Speaker 1: some extent, largely in San Francisco and Arizona. There's been 391 00:21:18,520 --> 00:21:21,159 Speaker 1: chatter about other cities kind of in the Sun Belt, 392 00:21:21,520 --> 00:21:24,960 Speaker 1: and so we do have some reports from basically customers 393 00:21:25,000 --> 00:21:27,359 Speaker 1: riding in these vehicles. And the thing is they go 394 00:21:27,560 --> 00:21:30,720 Speaker 1: a lot more slowly than a human driven car, partly 395 00:21:30,720 --> 00:21:33,199 Speaker 1: because they're following the speed limit, but partably because these 396 00:21:33,240 --> 00:21:35,359 Speaker 1: companies kind of like limit the speeds so that they're 397 00:21:35,359 --> 00:21:38,160 Speaker 1: able to react more properly, and so you have people 398 00:21:38,200 --> 00:21:40,880 Speaker 1: who are kind of riding them, but not riding them 399 00:21:41,320 --> 00:21:43,960 Speaker 1: like as an everyday thing. It's more like a fun thing. 400 00:21:44,000 --> 00:21:46,840 Speaker 1: And as one driverless car engineer described it to me, 401 00:21:46,880 --> 00:21:48,679 Speaker 1: it's kind of like an amusement park ride at the 402 00:21:48,720 --> 00:21:50,440 Speaker 1: moment right where it's like, hey, this will be a 403 00:21:50,520 --> 00:21:52,639 Speaker 1: kind of a fun thing. I'll take a robot to 404 00:21:52,640 --> 00:21:54,320 Speaker 1: work and it's gonna take twice as long and it's 405 00:21:54,359 --> 00:21:56,720 Speaker 1: gonna maybe stop a few times, but gee, that's gonna 406 00:21:56,720 --> 00:21:59,959 Speaker 1: be fun. And it is possible that that will get 407 00:21:59,800 --> 00:22:02,000 Speaker 1: better and that more and more people will do it, 408 00:22:02,040 --> 00:22:03,840 Speaker 1: And that's kind of the dream, at least if you're 409 00:22:03,840 --> 00:22:06,240 Speaker 1: talking about the companies that are still pursuing this aggressively. 410 00:22:07,760 --> 00:22:09,840 Speaker 1: So as you look forward, you said it's going to 411 00:22:09,920 --> 00:22:13,400 Speaker 1: take a while for this to happen. Lewandowski thinks too 412 00:22:13,440 --> 00:22:15,760 Speaker 1: that it's going to get there eventually, but it's just 413 00:22:15,840 --> 00:22:19,359 Speaker 1: going to take time. What are the big problems that 414 00:22:19,440 --> 00:22:21,600 Speaker 1: need to be solved in are some of those solutions 415 00:22:21,640 --> 00:22:25,160 Speaker 1: actually in the works? The problem and this is why 416 00:22:25,320 --> 00:22:28,679 Speaker 1: when you ask people like Lewandowski, like when is it 417 00:22:28,720 --> 00:22:31,199 Speaker 1: going to happen. They have a hard time giving you 418 00:22:31,240 --> 00:22:34,679 Speaker 1: a definitive answer is because the things that have to happen, 419 00:22:34,840 --> 00:22:39,280 Speaker 1: we don't really know what they are. Machine learning as 420 00:22:39,280 --> 00:22:43,720 Speaker 1: it exists today is really good at memorization, right. It 421 00:22:43,760 --> 00:22:47,359 Speaker 1: has to see all of these scenarios in order to 422 00:22:47,359 --> 00:22:49,280 Speaker 1: be able to do them again. And that's why drive 423 00:22:49,359 --> 00:22:52,479 Speaker 1: becomes such a big problem because there's so many different 424 00:22:52,520 --> 00:22:55,680 Speaker 1: things and you have to drive just almost an infinite 425 00:22:55,760 --> 00:22:57,600 Speaker 1: number of miles to be able to show the computer 426 00:22:58,080 --> 00:23:00,560 Speaker 1: all of the scenarios it needs to in order for 427 00:23:00,600 --> 00:23:03,040 Speaker 1: it to sort of learn what it needs to do. 428 00:23:03,080 --> 00:23:05,760 Speaker 1: You talk to the critics, and what they say is 429 00:23:06,040 --> 00:23:08,639 Speaker 1: it's not that we need a little bit more work 430 00:23:08,720 --> 00:23:10,479 Speaker 1: or something or a little bit more money. You need 431 00:23:10,520 --> 00:23:14,680 Speaker 1: a fundamental breakthrough in AI in order to make that happen. 432 00:23:14,720 --> 00:23:17,400 Speaker 1: And that's a breakthrough that we haven't had yet. One 433 00:23:17,440 --> 00:23:20,560 Speaker 1: person who's getting into it is he Elon Musk. He 434 00:23:20,560 --> 00:23:22,400 Speaker 1: had done it with Tesla. Now he's kind of making 435 00:23:22,440 --> 00:23:25,080 Speaker 1: a different sort of bet on driverless cars too. I mean, 436 00:23:25,320 --> 00:23:27,480 Speaker 1: one thing that happens when Alon Musk gets into any 437 00:23:27,480 --> 00:23:30,520 Speaker 1: spaces he generates a lot of attention for Alon Musk, 438 00:23:30,600 --> 00:23:34,360 Speaker 1: but then he tends to also advanced technologies a lot 439 00:23:34,359 --> 00:23:36,760 Speaker 1: of the time. Is there something that he's doing that's 440 00:23:36,800 --> 00:23:39,360 Speaker 1: different that would maybe accelerate some of the things you're 441 00:23:39,359 --> 00:23:41,920 Speaker 1: talking about. Oh so, there are two things that Musk 442 00:23:42,040 --> 00:23:45,560 Speaker 1: is doing that make Tesla especially interesting and make a 443 00:23:45,600 --> 00:23:47,560 Speaker 1: lot of people look at it and think, well, if 444 00:23:47,560 --> 00:23:49,600 Speaker 1: anybody's going to get there, you know, maybe it's gonna 445 00:23:49,600 --> 00:23:53,920 Speaker 1: be Tesla. So one of those things is that Tesla 446 00:23:54,080 --> 00:23:56,679 Speaker 1: has a lot of cars on the road collecting data. Right, 447 00:23:56,720 --> 00:23:59,760 Speaker 1: They have this gigantic fleet. They have these sort of 448 00:23:59,840 --> 00:24:04,600 Speaker 1: semi autonomous technologies. The cars are hooked up to cameras 449 00:24:04,640 --> 00:24:08,920 Speaker 1: and training Tesla's machine learning algorithms to be better at driving. 450 00:24:09,080 --> 00:24:10,760 Speaker 1: You know, I mentioned that getting a lot of miles 451 00:24:10,840 --> 00:24:13,840 Speaker 1: is really important, so Tesla has this kind of built 452 00:24:13,840 --> 00:24:17,200 Speaker 1: in advantage. The other thing that Tesla has is elon 453 00:24:17,280 --> 00:24:21,120 Speaker 1: Musk you may be aware of, this doesn't give a 454 00:24:21,200 --> 00:24:24,680 Speaker 1: you know what about rules and regulations, and so one 455 00:24:24,680 --> 00:24:28,400 Speaker 1: of the things that has kind of slowed slowed things down, 456 00:24:28,480 --> 00:24:31,280 Speaker 1: right is regulators. And Musk just sort of went ahead 457 00:24:31,480 --> 00:24:34,160 Speaker 1: and launched this thing that he called full self driving 458 00:24:34,480 --> 00:24:37,320 Speaker 1: without essentially you know, on his own, and he's sort 459 00:24:37,320 --> 00:24:39,800 Speaker 1: of in this kind of protracted fight with various parts 460 00:24:39,840 --> 00:24:43,000 Speaker 1: of government over various things, one of which has to 461 00:24:43,080 --> 00:24:47,440 Speaker 1: do with these semi autonomous software programs that Tesla has launched, 462 00:24:47,520 --> 00:24:49,600 Speaker 1: and that is you know, has on the road, and 463 00:24:49,680 --> 00:24:52,000 Speaker 1: so you could debate the morals of it, but having 464 00:24:52,040 --> 00:24:55,040 Speaker 1: that fleet and having that kind of more aggressive approach 465 00:24:55,320 --> 00:24:58,360 Speaker 1: to regulation does give him an advantage. Now, I would 466 00:24:58,440 --> 00:25:00,280 Speaker 1: argue he's going to run into a lot of the 467 00:25:00,320 --> 00:25:03,199 Speaker 1: same problems that all these other companies are running into, 468 00:25:03,359 --> 00:25:05,240 Speaker 1: but maybe he does have a leg up in some ways. 469 00:25:06,680 --> 00:25:09,080 Speaker 1: You could see the federal government wanting to get a 470 00:25:09,080 --> 00:25:13,040 Speaker 1: hand in this, especially if it starts to become more common. 471 00:25:13,440 --> 00:25:15,639 Speaker 1: The government doesn't have such a great track record of 472 00:25:15,720 --> 00:25:18,960 Speaker 1: keeping up with technology. Do you think that government involvement 473 00:25:19,040 --> 00:25:22,080 Speaker 1: would actually help or hurt? Well, I mean, the government 474 00:25:22,160 --> 00:25:25,760 Speaker 1: is very involved in this. The original technology that kicked 475 00:25:25,760 --> 00:25:29,000 Speaker 1: this off was at a DARPA event, the DARPA Grand Challenge, 476 00:25:29,080 --> 00:25:31,280 Speaker 1: you know, which is the Defense Department thing. So like 477 00:25:31,320 --> 00:25:33,679 Speaker 1: a lot of the fundamental research has come out of 478 00:25:33,680 --> 00:25:36,359 Speaker 1: the government, and the government is maybe not at the 479 00:25:36,400 --> 00:25:39,800 Speaker 1: federal level, but certainly the state level regulating this stuff. 480 00:25:39,800 --> 00:25:44,040 Speaker 1: California in particular has a really elaborate regulatory regime, which 481 00:25:44,080 --> 00:25:46,199 Speaker 1: is like part of the reason I think that the 482 00:25:46,240 --> 00:25:50,119 Speaker 1: companies have largely been in California, and they're a handful 483 00:25:50,160 --> 00:25:53,040 Speaker 1: of other states as well, Arizona and Nevada. But yeah, 484 00:25:53,040 --> 00:25:55,160 Speaker 1: I mean the government is trying. There are a bunch 485 00:25:55,200 --> 00:25:58,040 Speaker 1: of questions, one of which is like what counts as 486 00:25:58,080 --> 00:26:00,800 Speaker 1: a self driving car? Like is Elon must allowed to 487 00:26:00,920 --> 00:26:04,439 Speaker 1: advertise this Tesla technology as full self driving if you 488 00:26:04,520 --> 00:26:07,040 Speaker 1: have to keep your hands on the wheel. And then 489 00:26:07,040 --> 00:26:10,800 Speaker 1: the other question is kind of like safety, like at 490 00:26:10,800 --> 00:26:14,440 Speaker 1: what point do we think these cars are safe enough? 491 00:26:14,480 --> 00:26:17,320 Speaker 1: Did they have to be equal to human driving? Do 492 00:26:17,400 --> 00:26:19,960 Speaker 1: they have to be better? And again how much better? 493 00:26:20,240 --> 00:26:23,520 Speaker 1: And those are questions that are not solved yet. No, 494 00:26:23,600 --> 00:26:25,199 Speaker 1: those are not answered, and that is going to be 495 00:26:25,280 --> 00:26:27,920 Speaker 1: yet another kind of impediment because we as a society 496 00:26:27,960 --> 00:26:30,840 Speaker 1: off to decide at what point do we turn the 497 00:26:30,960 --> 00:26:34,560 Speaker 1: keys over to the bots. Max Chaffking, thanks for coming 498 00:26:34,600 --> 00:26:38,160 Speaker 1: on the show, Thanks for having me, Thanks for listening 499 00:26:38,160 --> 00:26:40,320 Speaker 1: to us here at the Big Take. It's a daily 500 00:26:40,359 --> 00:26:43,720 Speaker 1: podcast from Bloomberg and iHeartRadio. For more shows from my 501 00:26:43,800 --> 00:26:47,840 Speaker 1: heart Radio, visit the iHeartRadio app Apple podcast or wherever 502 00:26:47,880 --> 00:26:50,800 Speaker 1: you listen, and we'd love to hear from you. Email 503 00:26:50,880 --> 00:26:54,719 Speaker 1: us questions or comments to Big Take at Bloomberg dot net. 504 00:26:55,440 --> 00:26:58,880 Speaker 1: The supervising producer of The Big Take is Vicky Bergolina. 505 00:26:58,960 --> 00:27:03,240 Speaker 1: Our senior producer is Catherine Fink. Federica Romagnello is our producer. 506 00:27:03,520 --> 00:27:08,879 Speaker 1: Our associate producer is Zenobsiddiki. Raphael Msily is our engineer. 507 00:27:09,280 --> 00:27:13,680 Speaker 1: Our original music was composed by Leo Sidrin. I'm West Kasova. 508 00:27:13,920 --> 00:27:16,280 Speaker 1: We'll be back tomorrow with another Big Take