1 00:00:15,604 --> 00:00:28,684 Speaker 1: Pushkin. Welcome back to Risky Business, a show about making 2 00:00:28,724 --> 00:00:31,444 Speaker 1: better decisions. I'm Maria Kanakova. 3 00:00:31,364 --> 00:00:34,284 Speaker 2: And I'm Nate Silver today on the show, Maria, have 4 00:00:34,324 --> 00:00:38,404 Speaker 2: you ever taken a way Moo or another autonomous vehicle? 5 00:00:38,804 --> 00:00:43,124 Speaker 1: I have, not, Nate, but I know, I know, I know, 6 00:00:44,004 --> 00:00:47,004 Speaker 1: I know that you have. We've talked about that before. 7 00:00:47,404 --> 00:00:51,284 Speaker 1: But yeah, we'll be talking more in depth about kind 8 00:00:51,284 --> 00:00:56,204 Speaker 1: of the whole autonomous vehicle future, the risks, the rewards, 9 00:00:56,244 --> 00:00:58,964 Speaker 1: what's good, what's bad, how we should think about it, 10 00:00:59,004 --> 00:01:01,204 Speaker 1: how we should evaluate it. I think there are lots 11 00:01:01,204 --> 00:01:04,084 Speaker 1: of questions here and lots of things to think about 12 00:01:04,244 --> 00:01:06,884 Speaker 1: as we kind of figure out the future of driving. 13 00:01:06,884 --> 00:01:08,724 Speaker 2: We got a little risk, we got a little psychology, 14 00:01:08,884 --> 00:01:10,204 Speaker 2: we got a little algorithm TLIC. 15 00:01:10,244 --> 00:01:13,884 Speaker 3: It's gonna be a good episode, Maria. 16 00:01:15,284 --> 00:01:18,644 Speaker 1: So, Nate, you told me about some amazing experiences that 17 00:01:18,684 --> 00:01:22,284 Speaker 1: you've had using waymos in San Francisco. 18 00:01:23,084 --> 00:01:24,924 Speaker 3: Yeah. No, I uh. 19 00:01:25,044 --> 00:01:28,564 Speaker 2: The first time I took a WEEIMO, I was actually 20 00:01:28,564 --> 00:01:32,324 Speaker 2: at a meeting with Manifold, which is a prediction market company, 21 00:01:33,004 --> 00:01:35,484 Speaker 2: just shooting the shit. I'm friendly with those guys, and 22 00:01:35,484 --> 00:01:38,804 Speaker 2: they're like, we're gonna hail you Awaimo back to your 23 00:01:38,964 --> 00:01:42,164 Speaker 2: hotel in San Francisco, and I'm like, okay, this seems 24 00:01:42,244 --> 00:01:44,684 Speaker 2: kind of stupid, right, I don't know, we're afraid, and 25 00:01:44,804 --> 00:01:47,284 Speaker 2: like it was fucking it felt like the fucking future. 26 00:01:47,364 --> 00:01:47,564 Speaker 3: Right. 27 00:01:47,804 --> 00:01:50,644 Speaker 2: You get in there, there's no driver, there's like space 28 00:01:50,684 --> 00:01:54,004 Speaker 2: age music playing, and like San Francisco is like not 29 00:01:54,764 --> 00:01:57,804 Speaker 2: a particularly easy city to navigate. There's a lot of traffic, 30 00:01:58,204 --> 00:02:02,404 Speaker 2: there's lots of hills. I wouldn't call SF the most. 31 00:02:02,084 --> 00:02:06,564 Speaker 3: Traffic obedient place. Probably better than like Boston or something. Right, Oh, 32 00:02:06,564 --> 00:02:07,164 Speaker 3: I don't think. 33 00:02:07,044 --> 00:02:09,364 Speaker 1: Anything is worse than Boston when it comes to. 34 00:02:10,684 --> 00:02:11,804 Speaker 3: You know, my dad had like it. 35 00:02:12,524 --> 00:02:15,004 Speaker 2: We lived briefly in Boston for a year, and like 36 00:02:15,404 --> 00:02:17,644 Speaker 2: he got a falsely accused and like being in a 37 00:02:17,644 --> 00:02:20,684 Speaker 2: car accident. There's a Brian Q Silver and he was 38 00:02:20,764 --> 00:02:24,084 Speaker 2: Brian D Silver. Anyway, we'll leave it aside family lore. 39 00:02:24,644 --> 00:02:27,684 Speaker 3: Brian Q Silver, Fuck you, Brian Q Silver. He fucking 40 00:02:27,764 --> 00:02:29,084 Speaker 3: my dad's credit reading. 41 00:02:29,164 --> 00:02:30,964 Speaker 1: Oh my god, No, that's I mean, this is a 42 00:02:31,084 --> 00:02:34,244 Speaker 1: fun like this is a hilarious side of family lore, 43 00:02:34,644 --> 00:02:37,324 Speaker 1: but like this is actually a consideration when it comes 44 00:02:37,364 --> 00:02:39,924 Speaker 1: to self driving cars versus human driving cars. 45 00:02:40,044 --> 00:02:43,244 Speaker 2: No, it felt like the fucking future, right, and like it, 46 00:02:43,444 --> 00:02:46,084 Speaker 2: like I would say, and by the way, there's also 47 00:02:46,164 --> 00:02:49,404 Speaker 2: like lots of let's be honest, especially a couple of 48 00:02:49,444 --> 00:02:53,084 Speaker 2: years ago, and this happened a lot of vagrancy, homeless people, 49 00:02:53,884 --> 00:02:56,964 Speaker 2: sketchy people in downtown San Francisco, right, So like you're 50 00:02:56,964 --> 00:03:00,964 Speaker 2: dealing with an unpredictable element that might not be programmed 51 00:03:00,964 --> 00:03:03,964 Speaker 2: into some baseline model of like how driving is supposed 52 00:03:04,004 --> 00:03:06,764 Speaker 2: to work. But now the WAIMA did like a very 53 00:03:06,884 --> 00:03:09,484 Speaker 2: very good job. It's very smooth with the acceleration and 54 00:03:09,964 --> 00:03:12,644 Speaker 2: the deceleration. It's a little bit a little bit nitty 55 00:03:13,084 --> 00:03:16,724 Speaker 2: about following traffic signals and things like that. I was 56 00:03:16,804 --> 00:03:19,044 Speaker 2: you would expect, but it did like lots of I 57 00:03:19,084 --> 00:03:23,124 Speaker 2: think fairly saying in rational things, right, like I was 58 00:03:23,124 --> 00:03:25,764 Speaker 2: staying at some slightly funky hotel and like it pulls 59 00:03:25,804 --> 00:03:27,284 Speaker 2: up on the opposite side of the hotel where it's 60 00:03:27,324 --> 00:03:29,924 Speaker 2: like safer to get out and not the officially recommended spot, 61 00:03:29,964 --> 00:03:31,564 Speaker 2: and like it just you know, it would be like 62 00:03:31,564 --> 00:03:35,924 Speaker 2: in the ninety seventh or ninety eighth percentile of Uber drivers, 63 00:03:36,204 --> 00:03:37,684 Speaker 2: I would say, you know, whether or. 64 00:03:37,684 --> 00:03:39,444 Speaker 3: Not you like the conversation, Yeah, go ahead. 65 00:03:39,844 --> 00:03:41,764 Speaker 1: I was gonna say, it's funny that you say they're nitty. 66 00:03:42,044 --> 00:03:46,004 Speaker 1: So I'm in Las Vegas right now, and Vegas is 67 00:03:46,204 --> 00:03:50,524 Speaker 1: rolling out a lot of autonomous vehicles, not just way Mos, 68 00:03:50,524 --> 00:03:54,004 Speaker 1: but what's the what's the other one called zekes There 69 00:03:54,004 --> 00:03:59,804 Speaker 1: we go zu zukes all right, So Vegas is rolling 70 00:03:59,804 --> 00:04:03,764 Speaker 1: out a lot of autonomous vehicles, including zekes here. So 71 00:04:03,884 --> 00:04:07,284 Speaker 1: when I came back for the NAPT the North American 72 00:04:07,324 --> 00:04:09,684 Speaker 1: Poker to Our last month, was the first time that 73 00:04:09,724 --> 00:04:12,644 Speaker 1: they had done really the full rollout already where you 74 00:04:12,684 --> 00:04:15,524 Speaker 1: can hail them, you know as Uber, Left, et cetera. Well, 75 00:04:15,564 --> 00:04:17,684 Speaker 1: actually I don't know if it's both Uber and Left. Anyway, 76 00:04:17,724 --> 00:04:20,524 Speaker 1: you can use ride sharing services with them, and they're 77 00:04:20,564 --> 00:04:23,884 Speaker 1: all over the streets. And I was trying to drive, 78 00:04:24,244 --> 00:04:27,164 Speaker 1: and I realized very early on, and this says something 79 00:04:27,164 --> 00:04:30,524 Speaker 1: about how close I like to you know, I sometimes 80 00:04:30,604 --> 00:04:32,484 Speaker 1: cut things a little bit close when it comes to 81 00:04:32,524 --> 00:04:36,084 Speaker 1: making it to places on time. I realized right away 82 00:04:36,124 --> 00:04:39,164 Speaker 1: that I did not want to be behind them. Basically, 83 00:04:39,164 --> 00:04:41,924 Speaker 1: if I saw them in my lane, I tried to 84 00:04:41,964 --> 00:04:44,284 Speaker 1: like get into a different lane because they would not 85 00:04:44,324 --> 00:04:46,924 Speaker 1: go a single mile above the speed limit. They would 86 00:04:47,084 --> 00:04:51,164 Speaker 1: stop like and sometimes get a little bit, you know, confused. 87 00:04:51,204 --> 00:04:53,684 Speaker 1: They were very nitty, very cautious drivers, and I was like, 88 00:04:53,804 --> 00:04:55,684 Speaker 1: come on, guys, you can go thirty seven in the 89 00:04:55,684 --> 00:04:56,564 Speaker 1: thirty five zone. 90 00:04:56,604 --> 00:04:57,684 Speaker 3: It's not going to kill you. 91 00:04:57,724 --> 00:05:00,604 Speaker 1: It's gonna be okay. But they wouldn't, and which was 92 00:05:00,684 --> 00:05:02,684 Speaker 1: really interesting. And I don't know if it's you know, 93 00:05:02,724 --> 00:05:05,404 Speaker 1: if it's this particular company, but they were definitely very 94 00:05:05,484 --> 00:05:07,724 Speaker 1: nitty drivers, which, by the way, is not necessarily a 95 00:05:07,764 --> 00:05:10,444 Speaker 1: bad thing, right, Like they were being they were being cautious, 96 00:05:10,444 --> 00:05:12,484 Speaker 1: they were being safe. But for me, I was like, okay, 97 00:05:12,884 --> 00:05:16,324 Speaker 1: I want to go slightly faster. Sorry, Las Vegas Police. 98 00:05:16,324 --> 00:05:18,444 Speaker 1: I never go more than a few miles above the 99 00:05:18,444 --> 00:05:22,044 Speaker 1: speed limit, I promise, But like I got stuck behind 100 00:05:22,044 --> 00:05:25,564 Speaker 1: them a few times, and like they wouldn't take some 101 00:05:25,684 --> 00:05:28,844 Speaker 1: turns at intersections which I think a human driver might 102 00:05:28,844 --> 00:05:31,484 Speaker 1: have taken. Probably it's a little bit safer, Like they 103 00:05:31,644 --> 00:05:34,964 Speaker 1: just really play it safe in those scenarios. But I 104 00:05:35,044 --> 00:05:36,404 Speaker 1: just want it to be like, Okay, you can go, 105 00:05:36,484 --> 00:05:39,084 Speaker 1: you have room, let's go, let's make that turn. 106 00:05:40,284 --> 00:05:42,644 Speaker 2: I mean, there are there are times when it's probably 107 00:05:43,564 --> 00:05:46,644 Speaker 2: I don't drive, actually I do bike sometimes, right, like 108 00:05:46,684 --> 00:05:48,804 Speaker 2: on the city bikes or whatever. Right, there are times 109 00:05:48,844 --> 00:05:54,804 Speaker 2: when violating traffic laws is probably safer. Yeah, you know 110 00:05:54,804 --> 00:05:57,804 Speaker 2: what I mean, like mile violations where or just on 111 00:05:57,924 --> 00:06:00,044 Speaker 2: the common sentence, so you're trying to yeah, I don't know. 112 00:06:00,964 --> 00:06:03,004 Speaker 3: I think that's really and that's more true in New 113 00:06:03,084 --> 00:06:05,444 Speaker 3: York than probably other locations. I would think. 114 00:06:05,844 --> 00:06:08,284 Speaker 1: Now, sometimes you do have to make these judgment calls 115 00:06:08,324 --> 00:06:11,644 Speaker 1: because sometimes like you will see something and in order 116 00:06:11,764 --> 00:06:14,524 Speaker 1: to be safer, like you will have to violate a 117 00:06:14,684 --> 00:06:17,764 Speaker 1: traffic rule. And that's I think one of these broader 118 00:06:18,084 --> 00:06:20,644 Speaker 1: you know, as we as we start talking about kind 119 00:06:20,684 --> 00:06:24,164 Speaker 1: of the broader risk reward and kind of how this 120 00:06:24,324 --> 00:06:28,444 Speaker 1: future looks, and you know what we think about autonomous vehicles, 121 00:06:28,764 --> 00:06:31,604 Speaker 1: I think that a lot of the questions do revolve around, 122 00:06:31,964 --> 00:06:34,884 Speaker 1: you know, how do you program the cars to actually 123 00:06:34,924 --> 00:06:38,324 Speaker 1: make these decisions, because you know, it's not It would 124 00:06:38,364 --> 00:06:43,964 Speaker 1: be one thing if we had this immediate transition right 125 00:06:44,164 --> 00:06:47,404 Speaker 1: where like you can just go like this snap of 126 00:06:47,444 --> 00:06:49,844 Speaker 1: the fingers and all of a sudden, all the cars 127 00:06:49,844 --> 00:06:53,084 Speaker 1: on the road are autonomous vehicles. That's very different from 128 00:06:53,124 --> 00:06:56,564 Speaker 1: having kind of multiple years of in between where you 129 00:06:56,644 --> 00:07:00,604 Speaker 1: have human drivers and autonomous vehicles and you know, more 130 00:07:00,644 --> 00:07:03,484 Speaker 1: human drivers than autonomous vehicles in all of these kind 131 00:07:03,524 --> 00:07:07,684 Speaker 1: of the transitional period, I think from a risk safety 132 00:07:08,164 --> 00:07:10,964 Speaker 1: standpoint and from making those types of trade offs is 133 00:07:11,044 --> 00:07:15,004 Speaker 1: probably going to be the most difficult to navigate because 134 00:07:15,764 --> 00:07:18,644 Speaker 1: you have to account not just for ideal driving and 135 00:07:18,724 --> 00:07:21,964 Speaker 1: what you're supposed to be doing, but what humans actually 136 00:07:22,004 --> 00:07:24,764 Speaker 1: do right, and the mistakes that humans make, and how 137 00:07:24,804 --> 00:07:29,124 Speaker 1: you share the road with humans, some of whom are 138 00:07:29,124 --> 00:07:31,964 Speaker 1: bad drivers, some of whom are impaired drivers, right, because 139 00:07:32,004 --> 00:07:35,564 Speaker 1: those risks don't suddenly magically go away. Now, it's kind 140 00:07:35,604 --> 00:07:37,364 Speaker 1: of you know, it's funny. We talk a lot about 141 00:07:37,404 --> 00:07:39,964 Speaker 1: poker and kind of ideal poker on this show. It's 142 00:07:40,044 --> 00:07:42,524 Speaker 1: kind of like, you know, if you have the GTO 143 00:07:42,684 --> 00:07:45,524 Speaker 1: cars right, but then you have all of these crazy 144 00:07:45,564 --> 00:07:47,844 Speaker 1: and different types of players on the road, and you 145 00:07:47,884 --> 00:07:50,004 Speaker 1: have to try to figure out like, sure, you have 146 00:07:50,044 --> 00:07:52,164 Speaker 1: to drive in a GTO way, but you also have 147 00:07:52,244 --> 00:07:55,644 Speaker 1: to make sure that you are optimizing for the fact 148 00:07:55,644 --> 00:07:58,124 Speaker 1: that you have, you know, these loose cannons and all 149 00:07:58,164 --> 00:08:00,924 Speaker 1: of these different elements that are still going to be 150 00:08:00,924 --> 00:08:03,604 Speaker 1: sharing the roads with you. And unless there's like this 151 00:08:03,804 --> 00:08:05,764 Speaker 1: huge thing where all of a sudden, the government's like, 152 00:08:05,964 --> 00:08:08,524 Speaker 1: we're going to pay you, you know, tons of money 153 00:08:08,524 --> 00:08:11,924 Speaker 1: to give up your car and use autonomous vehicles. Like 154 00:08:11,964 --> 00:08:13,964 Speaker 1: that transition, to me is going to be one of 155 00:08:13,964 --> 00:08:15,964 Speaker 1: the riskier parts of this whole equation. 156 00:08:17,764 --> 00:08:19,324 Speaker 3: Well, I assume that these. 157 00:08:20,684 --> 00:08:23,524 Speaker 2: Cars are programmed in like find if you call it, 158 00:08:23,604 --> 00:08:26,644 Speaker 2: like an exploitive strategy in poker, right, but as soon 159 00:08:27,044 --> 00:08:31,284 Speaker 2: they are at least doing some defensive driving, right. I 160 00:08:31,364 --> 00:08:34,684 Speaker 2: mean again, San Francisco not a place which is ideal 161 00:08:34,764 --> 00:08:38,684 Speaker 2: for traffic laws or anything else, really right, and they're 162 00:08:38,724 --> 00:08:40,164 Speaker 2: kind of training these real world environments. 163 00:08:40,164 --> 00:08:41,924 Speaker 3: I mean, look, I mean to back up a second, right, Like, 164 00:08:42,044 --> 00:08:43,124 Speaker 3: in general. 165 00:08:43,204 --> 00:08:47,724 Speaker 2: There's this critique that like AI driven systems are good 166 00:08:47,764 --> 00:08:53,044 Speaker 2: with simple manipulation tasks, so language, certain types of math, right, 167 00:08:53,124 --> 00:08:53,884 Speaker 2: certain types. 168 00:08:53,684 --> 00:08:54,844 Speaker 3: Of games when they're well trained. 169 00:08:54,844 --> 00:08:56,444 Speaker 2: I know we've talked in this program about how poker 170 00:08:56,804 --> 00:08:59,084 Speaker 2: if you're not training on poker specifically, might be an exception. 171 00:08:59,684 --> 00:09:02,884 Speaker 2: So it is kind of amazing that like that they're 172 00:09:02,924 --> 00:09:05,004 Speaker 2: as good as they are, right, But like, yeah, the 173 00:09:05,084 --> 00:09:09,364 Speaker 2: thing people are recognize is like the human drivers are 174 00:09:09,404 --> 00:09:11,644 Speaker 2: I have lots of problems, you know what I mean, 175 00:09:12,564 --> 00:09:14,524 Speaker 2: They're distracted I mean, I was just in where was 176 00:09:14,724 --> 00:09:17,804 Speaker 2: I was just in Florida, right, And like. 177 00:09:18,164 --> 00:09:22,284 Speaker 3: In Florida you can talk on your cell phone when 178 00:09:22,284 --> 00:09:23,284 Speaker 3: you're driving. 179 00:09:23,204 --> 00:09:26,124 Speaker 1: And like, oh my god, is that true night? 180 00:09:26,244 --> 00:09:28,524 Speaker 3: Oh my god, you do neither you nor there? 181 00:09:29,124 --> 00:09:31,884 Speaker 1: No, no, no, that's actually that's actually both here and there. 182 00:09:31,924 --> 00:09:34,044 Speaker 1: I don't remember if I've ever told our listeners, but 183 00:09:34,604 --> 00:09:36,924 Speaker 1: the one of the scariest experiences I've had in an 184 00:09:37,044 --> 00:09:39,604 Speaker 1: uber was in California. It was in LA. We were 185 00:09:39,644 --> 00:09:41,484 Speaker 1: stuck in traffic in the middle of you know, one 186 00:09:41,524 --> 00:09:45,044 Speaker 1: of these horrible LA freeways, and my driver was already 187 00:09:45,124 --> 00:09:48,524 Speaker 1: like exhibiting signs of strange behavior, and then he just 188 00:09:48,564 --> 00:09:52,804 Speaker 1: like stares intently at me in the rear view mirror 189 00:09:52,844 --> 00:09:55,444 Speaker 1: and says, do you believe in God? 190 00:09:56,164 --> 00:09:58,004 Speaker 3: That I was like, oh no, oh. 191 00:09:57,884 --> 00:10:01,124 Speaker 2: God, no, oh no. 192 00:10:01,564 --> 00:10:04,524 Speaker 1: Anyway, this is actually very Germane because that is one 193 00:10:04,564 --> 00:10:07,524 Speaker 1: of the things that I think self driving, the autonomous 194 00:10:07,564 --> 00:10:11,364 Speaker 1: vehicles definitely have a leg up on. You are not 195 00:10:11,524 --> 00:10:14,284 Speaker 1: going to have one of those do you believe in God? 196 00:10:14,444 --> 00:10:17,684 Speaker 1: Experiences that are going to make you want to immediately 197 00:10:17,684 --> 00:10:28,844 Speaker 1: get out of the car and we'll be back right 198 00:10:28,884 --> 00:10:42,164 Speaker 1: after this. So I think that one of to me, 199 00:10:42,444 --> 00:10:45,524 Speaker 1: one of the big good things about the future of 200 00:10:45,564 --> 00:10:49,764 Speaker 1: autonomous vehicles is the fact that you kind of sometimes 201 00:10:50,004 --> 00:10:53,684 Speaker 1: like this is evidence for when taking the human equation 202 00:10:53,764 --> 00:10:56,004 Speaker 1: out of it can be good. Right, You don't have 203 00:10:56,124 --> 00:10:59,964 Speaker 1: to like every time that I'm coming home, you know, 204 00:11:00,004 --> 00:11:02,564 Speaker 1: every time I'm going back to Brooklyn from JFK. When 205 00:11:02,604 --> 00:11:05,644 Speaker 1: I get back to New York, I end up taking 206 00:11:05,684 --> 00:11:09,164 Speaker 1: the taxi, but I get into arguments with the driver 207 00:11:09,524 --> 00:11:13,564 Speaker 1: every single time because inevitably they want to take the 208 00:11:13,604 --> 00:11:15,564 Speaker 1: Belt Parkway, which is never the best way to go. 209 00:11:15,604 --> 00:11:17,884 Speaker 1: You always get stuck in traffic, it's always horrible, but 210 00:11:17,924 --> 00:11:19,604 Speaker 1: they always want to do that, and we just get 211 00:11:19,644 --> 00:11:21,324 Speaker 1: into a back and forth. I'm like, no, I want 212 00:11:21,324 --> 00:11:22,844 Speaker 1: you to do this, and they're like, no, I'm gonna 213 00:11:22,884 --> 00:11:26,884 Speaker 1: do that. And with presumably with autonomous vehicles, you know, 214 00:11:26,924 --> 00:11:28,644 Speaker 1: a lot of this won't come up, right, they have 215 00:11:29,044 --> 00:11:33,164 Speaker 1: the roots and they will kind of they will, they will. 216 00:11:33,004 --> 00:11:38,444 Speaker 2: Do Yeah, I'm a little more sympathetic empirically, and I 217 00:11:38,484 --> 00:11:40,804 Speaker 2: don't drive, but my partner does and we drive decently 218 00:11:40,964 --> 00:11:45,244 Speaker 2: off the bike. Empirically, I think like when the GPS 219 00:11:45,604 --> 00:11:49,004 Speaker 2: says something that like you as an experienced driver in 220 00:11:49,004 --> 00:11:50,004 Speaker 2: that area. 221 00:11:50,164 --> 00:11:53,004 Speaker 3: Disagree with I think it's kind of fifty to fifty. 222 00:11:53,004 --> 00:11:55,364 Speaker 3: Who's right? Yeah and who isn't? You know what I mean? 223 00:11:55,524 --> 00:11:57,444 Speaker 2: From like little things like knowing, oh, this street is 224 00:11:57,484 --> 00:12:00,004 Speaker 2: closed when we used to live near Penn Station. Right, 225 00:12:01,084 --> 00:12:02,804 Speaker 2: there's a nixt game starting. You don't want to fucking 226 00:12:02,844 --> 00:12:05,684 Speaker 2: drive anywhere near Massion Square Garden when there's a next 227 00:12:05,684 --> 00:12:08,964 Speaker 2: game starting. Or I mean, look, obviously, if you're in 228 00:12:09,004 --> 00:12:12,204 Speaker 2: a new or a lift, then the estimate of when 229 00:12:12,244 --> 00:12:16,884 Speaker 2: you'll arrive at your destination is strongly biased toward optimism. 230 00:12:17,084 --> 00:12:18,404 Speaker 3: You know, this is why I want, like, I want 231 00:12:18,404 --> 00:12:21,964 Speaker 3: prediction Mark. We need polymarket nation, we can bet. 232 00:12:22,484 --> 00:12:25,564 Speaker 1: Yeah, we need psychology actually in this because I think 233 00:12:25,644 --> 00:12:29,244 Speaker 1: that I actually think that cars get this absolutely backwards, 234 00:12:29,484 --> 00:12:32,804 Speaker 1: because it's incredibly frustrating when you see, like you go 235 00:12:32,924 --> 00:12:35,604 Speaker 1: to order an uber a lift and it's like two 236 00:12:35,604 --> 00:12:38,084 Speaker 1: minutes and then it ends up being eight minutes because 237 00:12:38,124 --> 00:12:40,444 Speaker 1: they basically have lied to you and they know that 238 00:12:40,524 --> 00:12:41,804 Speaker 1: right away, and then they're like, oh, it's going to 239 00:12:41,844 --> 00:12:44,164 Speaker 1: take you, you know, ten minutes to get here, or this 240 00:12:44,204 --> 00:12:47,484 Speaker 1: traffic slow down is a fifteen minute slow down, whatever 241 00:12:47,524 --> 00:12:50,284 Speaker 1: it is, it's always optimistic right, And the way you 242 00:12:50,324 --> 00:12:53,324 Speaker 1: need to do it psychologically speaking is you want to 243 00:12:53,364 --> 00:12:57,564 Speaker 1: do the old underpromise over deliver, right. They should actually 244 00:12:57,644 --> 00:13:01,324 Speaker 1: know that this is statistically speaking, you're probably going to 245 00:13:01,364 --> 00:13:03,644 Speaker 1: be at the other side of the distribution. So instead, 246 00:13:03,724 --> 00:13:06,444 Speaker 1: let me say you know that your uber is coming 247 00:13:06,484 --> 00:13:08,844 Speaker 1: in eight minutes, and then when it arrives in four minutes, 248 00:13:08,884 --> 00:13:11,644 Speaker 1: you'll be happy instead of frustrating. Let me tell you 249 00:13:11,684 --> 00:13:14,284 Speaker 1: that the slowdown is going to be twenty five minutes, 250 00:13:14,324 --> 00:13:16,244 Speaker 1: and then when it's over in twenty minutes, you're gonna 251 00:13:16,244 --> 00:13:17,724 Speaker 1: be like, oh my god, it was much shorter. 252 00:13:18,124 --> 00:13:18,804 Speaker 3: This is great. 253 00:13:18,924 --> 00:13:21,644 Speaker 1: So I actually think that psychologically they get this totally wrong. 254 00:13:21,884 --> 00:13:24,644 Speaker 1: We're talking about self driving cars in your. 255 00:13:25,444 --> 00:13:28,124 Speaker 2: Room at your at Bellagio and I'm gonna, I'm gonna, 256 00:13:28,364 --> 00:13:32,124 Speaker 2: I'm gonna order a lift in six minutes, right, and 257 00:13:32,204 --> 00:13:33,484 Speaker 2: it's like they're in two minutes. You have to get 258 00:13:33,524 --> 00:13:36,724 Speaker 2: on the fucking ride share thing. You don't take these 259 00:13:36,764 --> 00:13:38,644 Speaker 2: zubers in Vegas, I guess very much, right. 260 00:13:39,284 --> 00:13:42,324 Speaker 1: No, that's it's yeah, No, that's a different thing. But 261 00:13:42,484 --> 00:13:47,644 Speaker 1: I think you just right now, there's no reliable gauge 262 00:13:47,644 --> 00:13:50,884 Speaker 1: for how quickly or not a car will actually come. 263 00:13:50,924 --> 00:13:53,164 Speaker 1: But I wonder if autonomous vehicles will actually solve this 264 00:13:53,244 --> 00:13:55,644 Speaker 1: problem or if they're just gonna lie the exact same way, 265 00:13:55,804 --> 00:13:59,084 Speaker 1: right that this is a problem with the underlying algorithm 266 00:13:59,204 --> 00:14:02,444 Speaker 1: as opposed to I think anything else. The other thing 267 00:14:02,684 --> 00:14:05,244 Speaker 1: that with autonomous vehicle is actually where you might need 268 00:14:05,284 --> 00:14:08,564 Speaker 1: a human. So talking about GPS and all of these things, 269 00:14:09,204 --> 00:14:12,404 Speaker 1: for whatever reason, where I live in Vegas for a 270 00:14:12,524 --> 00:14:17,684 Speaker 1: very long time was for some reason misrepresented on the GPS, 271 00:14:18,124 --> 00:14:21,644 Speaker 1: and drivers would actually like go to the totally wrong place, 272 00:14:21,684 --> 00:14:24,044 Speaker 1: and it was an absolute nightmare wanting you needed to 273 00:14:24,044 --> 00:14:30,004 Speaker 1: get picked up because somehow the GPS just completely fucked up, right, 274 00:14:30,244 --> 00:14:32,284 Speaker 1: and so people would just not be able to find it. 275 00:14:32,724 --> 00:14:36,044 Speaker 1: And so if there's a human there, you can talk 276 00:14:36,084 --> 00:14:38,124 Speaker 1: to them and be like, hey, okay, no, you need 277 00:14:38,164 --> 00:14:39,604 Speaker 1: to do this, you need to take a turn here, 278 00:14:39,684 --> 00:14:41,684 Speaker 1: you need to do this to get me. What do 279 00:14:41,724 --> 00:14:43,764 Speaker 1: I do if it's a weimo that's coming to pick 280 00:14:43,804 --> 00:14:46,444 Speaker 1: me up? How do I make sure that when you 281 00:14:46,604 --> 00:14:50,244 Speaker 1: need kind of that human intervention because the technology is 282 00:14:50,284 --> 00:14:53,404 Speaker 1: fucking up. Like I said, I've never taken weimos, so 283 00:14:53,604 --> 00:14:56,484 Speaker 1: I don't know if you can interact with them in 284 00:14:56,524 --> 00:14:59,764 Speaker 1: the same way. Maybe this isn't a problem, but it 285 00:14:59,844 --> 00:15:03,404 Speaker 1: seems like it might be a point of friction. 286 00:15:05,484 --> 00:15:10,364 Speaker 2: I mean, look, one problem with algorithms, really with algorithms there, 287 00:15:10,404 --> 00:15:13,484 Speaker 2: with people who designed the algorithms, right, is like there's 288 00:15:13,604 --> 00:15:18,844 Speaker 2: like a lot of over literalness. 289 00:15:18,044 --> 00:15:20,564 Speaker 3: You know what I mean. Like I used to live near. 290 00:15:21,884 --> 00:15:25,964 Speaker 2: A corner near Penn Station, right, and this is a 291 00:15:26,164 --> 00:15:31,564 Speaker 2: very busy all times of days intersection, right, But there 292 00:15:31,604 --> 00:15:34,404 Speaker 2: are lots of times when if not makes sense to 293 00:15:34,444 --> 00:15:37,724 Speaker 2: loot all the way around the whole avenue when you 294 00:15:37,724 --> 00:15:40,844 Speaker 2: can just drop me on like one block north or 295 00:15:40,884 --> 00:15:43,964 Speaker 2: one block south right, and like so like that logic 296 00:15:44,044 --> 00:15:47,404 Speaker 2: that like a human driver, I'm just gonna drop you 297 00:15:47,484 --> 00:15:49,244 Speaker 2: twenty ninth and seventh and not loop. 298 00:15:49,084 --> 00:15:51,884 Speaker 1: Around because otherwise it's gonna add five minutes to fifteen 299 00:15:51,924 --> 00:15:55,284 Speaker 1: fucking minutes, yeah, fifteen minutes whatever it or like hey, pick. 300 00:15:55,084 --> 00:15:56,804 Speaker 2: I'm gonna pick you up one seventh and not on 301 00:15:57,044 --> 00:15:59,964 Speaker 2: on thirtieth. Like that kind of thing has never been 302 00:16:00,764 --> 00:16:02,764 Speaker 2: a strength of any of these, right, because you're like 303 00:16:02,804 --> 00:16:05,604 Speaker 2: overly literal about like okay, get me the general area, 304 00:16:05,684 --> 00:16:07,364 Speaker 2: and like this is you know a lot of the 305 00:16:07,404 --> 00:16:11,644 Speaker 2: pickoff and drop off locations can be overly literal at 306 00:16:11,684 --> 00:16:13,804 Speaker 2: some properties and things like that. 307 00:16:13,924 --> 00:16:14,644 Speaker 3: I mean, I don't know. 308 00:16:15,404 --> 00:16:19,564 Speaker 1: Yeah, presumably algorithms will get better, and they'll they'll get 309 00:16:19,564 --> 00:16:22,164 Speaker 1: better at things like that, but there are other safety 310 00:16:22,204 --> 00:16:24,444 Speaker 1: issues that I do wonder about. And by the way, 311 00:16:24,724 --> 00:16:27,644 Speaker 1: a few weeks back, there was a big op ed 312 00:16:27,884 --> 00:16:31,604 Speaker 1: from a surgeon in the New York Times about autonomous 313 00:16:31,684 --> 00:16:36,244 Speaker 1: vehicles and you know, accident rates and not just fatalities, 314 00:16:36,284 --> 00:16:40,604 Speaker 1: but the types of injuries that the physician you know 315 00:16:40,684 --> 00:16:44,124 Speaker 1: has seen in different situations. And it was arguing for 316 00:16:44,364 --> 00:16:48,284 Speaker 1: more autonomous vehicles because you know of how these insane 317 00:16:48,324 --> 00:16:50,604 Speaker 1: situations that human drivers get into, and that their safety 318 00:16:50,644 --> 00:16:55,284 Speaker 1: record is much better. Obviously this is a tiny fraction, 319 00:16:55,764 --> 00:16:57,564 Speaker 1: you know, et cetera, et cetera, et cetera. But that 320 00:16:57,644 --> 00:17:00,964 Speaker 1: made me much more kind of pro autonomous vehicles than 321 00:17:01,004 --> 00:17:03,084 Speaker 1: I might otherwise have been, because I think that this 322 00:17:03,164 --> 00:17:05,724 Speaker 1: is this is a big, a big deal, and we 323 00:17:05,804 --> 00:17:08,924 Speaker 1: do want to be thinking about, you know, about those 324 00:17:08,924 --> 00:17:13,004 Speaker 1: types of questions. But some of the other kind of 325 00:17:13,044 --> 00:17:16,764 Speaker 1: algorithmic questions are you know, how do you and this 326 00:17:16,844 --> 00:17:19,284 Speaker 1: is you know, I don't want to get all like 327 00:17:19,324 --> 00:17:22,324 Speaker 1: philosophical here, But it does like get into kind of 328 00:17:22,324 --> 00:17:26,924 Speaker 1: these morality problems, types of situations where like whose life 329 00:17:26,924 --> 00:17:29,244 Speaker 1: do you prioritize? Right? Like how do you make those 330 00:17:29,284 --> 00:17:31,724 Speaker 1: types of judgment calls? What is the algorithm going to 331 00:17:31,724 --> 00:17:35,444 Speaker 1: do in different situations where there's no good answer right 332 00:17:35,484 --> 00:17:38,644 Speaker 1: where you might endanger your passenger or engager someone crossing 333 00:17:38,724 --> 00:17:41,444 Speaker 1: the street, or do this or do that, especially as 334 00:17:41,484 --> 00:17:44,724 Speaker 1: we started off the show talking about in this transitional 335 00:17:44,764 --> 00:17:47,404 Speaker 1: period where you're not dealing with other autonomous vehicles, right, 336 00:17:47,484 --> 00:17:49,364 Speaker 1: you're dealing with a lot of humans on the road. 337 00:17:49,604 --> 00:17:51,284 Speaker 1: So how do you make those trade offs? How do 338 00:17:51,324 --> 00:17:55,004 Speaker 1: you program that? You know, who's responsible for programming those algorithms? 339 00:17:55,284 --> 00:17:57,244 Speaker 1: How are they thinking about it? I think these are 340 00:17:57,284 --> 00:17:59,964 Speaker 1: all really important questions where it's not a question of 341 00:18:00,004 --> 00:18:03,004 Speaker 1: tweaking the algorithm and saying, oh, like let the person 342 00:18:03,044 --> 00:18:05,524 Speaker 1: tell them to like pick me up at this corner instead. 343 00:18:06,084 --> 00:18:07,724 Speaker 1: These are much deeper questions. 344 00:18:09,484 --> 00:18:11,764 Speaker 2: Yeah, I mean I some of our listeners are probably 345 00:18:11,764 --> 00:18:13,964 Speaker 2: familiar with with the trolley problem. 346 00:18:14,084 --> 00:18:17,804 Speaker 1: Yeah. Yeah, the trolley problem is your is your gold 347 00:18:17,844 --> 00:18:20,164 Speaker 1: standard here and all the iterations of. 348 00:18:20,084 --> 00:18:24,684 Speaker 2: It, which is a philosophical dilemma. I guess where there 349 00:18:24,724 --> 00:18:28,044 Speaker 2: is a trolley on a busy track. It's going to 350 00:18:28,204 --> 00:18:30,284 Speaker 2: kill I guess some people who are tied to the track. 351 00:18:30,364 --> 00:18:34,644 Speaker 2: You can also divert it and instead of killing these 352 00:18:34,684 --> 00:18:36,884 Speaker 2: five people, you kill like two workers. 353 00:18:36,924 --> 00:18:38,244 Speaker 3: I mean, there are there are lots. 354 00:18:38,044 --> 00:18:39,244 Speaker 1: Of infinite variations of it. 355 00:18:39,284 --> 00:18:41,124 Speaker 2: I think it seems pretty us like diverted kill the 356 00:18:41,124 --> 00:18:43,044 Speaker 2: two and not the five. Right, But Like the point is, 357 00:18:43,044 --> 00:18:46,364 Speaker 2: if you're kind of like intervening and making a decision, 358 00:18:46,924 --> 00:18:51,124 Speaker 2: then it begins to feel like I feel like, yeah, 359 00:18:51,164 --> 00:18:53,564 Speaker 2: you've made a moral choice, right, And. 360 00:18:53,724 --> 00:18:55,324 Speaker 1: And what if you know that one of these is 361 00:18:55,364 --> 00:18:58,364 Speaker 1: an infant one of these is a Nobel Prize winning set, Right, 362 00:18:58,444 --> 00:19:00,844 Speaker 1: what if you actually know something about the identities? Like 363 00:19:00,884 --> 00:19:03,124 Speaker 1: you said, there are infinite variations of this, of the 364 00:19:03,164 --> 00:19:07,324 Speaker 1: trolley problem, and there's no definitive answer. Like philosophers disagree 365 00:19:07,324 --> 00:19:09,644 Speaker 1: on the answers to this to this day. There are 366 00:19:09,644 --> 00:19:13,644 Speaker 1: different philosophical schools that have that advocate different approaches. And 367 00:19:14,324 --> 00:19:17,604 Speaker 1: this is I mean, autonomous vehicles have to deal with 368 00:19:17,924 --> 00:19:21,604 Speaker 1: one big trolley problem, right in some sense. 369 00:19:21,444 --> 00:19:23,924 Speaker 2: Yeah, when and when you have to quantify things, it's 370 00:19:24,004 --> 00:19:27,164 Speaker 2: one of the things that I will defend some of 371 00:19:27,164 --> 00:19:30,884 Speaker 2: the effective altruists and like rationalists about right is Thelle 372 00:19:30,884 --> 00:19:33,564 Speaker 2: will be like, well, what's what's the value of a 373 00:19:33,604 --> 00:19:37,684 Speaker 2: ship's life compared to like a human life? Yep, And 374 00:19:37,724 --> 00:19:40,924 Speaker 2: it can be a little off putting to think about. 375 00:19:41,204 --> 00:19:42,964 Speaker 2: On the other hand, we make the trade offs all 376 00:19:43,004 --> 00:19:44,524 Speaker 2: the time, right. I talk about an example in my 377 00:19:44,564 --> 00:19:47,484 Speaker 2: book from a few years ago where like the entire 378 00:19:47,724 --> 00:19:50,164 Speaker 2: New York City, not the entire New York City subway system, 379 00:19:50,324 --> 00:19:55,044 Speaker 2: several lines, like the f whatever lines were shut down 380 00:19:55,044 --> 00:19:57,084 Speaker 2: because of a loose dog that I had in somewhere 381 00:19:57,084 --> 00:20:00,764 Speaker 2: in Brooklyn Heights. I think, right, And like, you know, 382 00:20:01,244 --> 00:20:04,924 Speaker 2: I mean, there is a big cost to shutting down 383 00:20:04,964 --> 00:20:05,484 Speaker 2: the subway. 384 00:20:05,524 --> 00:20:06,404 Speaker 3: It makes people late. 385 00:20:06,444 --> 00:20:09,124 Speaker 2: It means that can include like emergency workers and things 386 00:20:09,124 --> 00:20:12,004 Speaker 2: like that, because often in New York the fastest way 387 00:20:12,044 --> 00:20:14,124 Speaker 2: to get around is with the subway, no matter, you know, 388 00:20:14,364 --> 00:20:16,244 Speaker 2: unless you have an ambulance with the right of way 389 00:20:16,604 --> 00:20:20,484 Speaker 2: and things like that. Right, and people are comfortable talking 390 00:20:20,524 --> 00:20:22,644 Speaker 2: about like what is the cost of a lost hour 391 00:20:22,684 --> 00:20:25,204 Speaker 2: of economic productivity? And how's that way against a dog's 392 00:20:25,244 --> 00:20:26,204 Speaker 2: life and things like that. 393 00:20:26,244 --> 00:20:28,204 Speaker 3: But the point is that like that. 394 00:20:28,124 --> 00:20:30,204 Speaker 2: When you kind of force people to make decisions that 395 00:20:30,244 --> 00:20:34,644 Speaker 2: are explicit and if playing their logic then then that 396 00:20:34,724 --> 00:20:37,844 Speaker 2: becomes problematic in some ways. 397 00:20:37,924 --> 00:20:42,404 Speaker 1: Right, Yeah, And you know when you are talking about 398 00:20:42,444 --> 00:20:45,484 Speaker 1: all of these autonomous vehicles and all of these different companies, 399 00:20:45,684 --> 00:20:48,924 Speaker 1: you have to realize. I mean we've talked about this before, 400 00:20:48,964 --> 00:20:51,924 Speaker 1: but I think it's a really important point that algorithms 401 00:20:51,964 --> 00:20:53,364 Speaker 1: do not exist in a vacuum. 402 00:20:53,684 --> 00:20:53,844 Speaker 3: Right. 403 00:20:53,884 --> 00:20:58,004 Speaker 1: They are based on they're humans who are programming them. 404 00:20:58,244 --> 00:21:00,964 Speaker 1: There are people who are waiting different inputs. Those are 405 00:21:01,004 --> 00:21:04,524 Speaker 1: all judgment choices. Even when you're looking at you know 406 00:21:04,644 --> 00:21:08,724 Speaker 1: AI and like lms, that's also trained on some set 407 00:21:08,764 --> 00:21:11,444 Speaker 1: of data. Right, which data is it trained on? Who 408 00:21:11,564 --> 00:21:14,964 Speaker 1: created that data? And like those weights really matter, right, 409 00:21:14,964 --> 00:21:17,444 Speaker 1: they can make a really big difference. And like all 410 00:21:17,484 --> 00:21:19,404 Speaker 1: of a sudden, something's going to go off the rails 411 00:21:19,644 --> 00:21:23,244 Speaker 1: because of a tiny, tiny tweak and there's no such 412 00:21:23,364 --> 00:21:26,684 Speaker 1: it's there's no such thing as objectivity, and because it's 413 00:21:26,764 --> 00:21:29,804 Speaker 1: being made by an algorithm, it's objective and it's perfect. 414 00:21:30,364 --> 00:21:32,044 Speaker 1: Like that just doesn't that's that's r I don't know. 415 00:21:32,084 --> 00:21:35,764 Speaker 1: I just think that human beings are absolutely I just 416 00:21:35,804 --> 00:21:38,844 Speaker 1: mean in algorithms like it's this, it's this myth and 417 00:21:38,884 --> 00:21:41,964 Speaker 1: this you know, misc not misconception, because I think a 418 00:21:41,964 --> 00:21:44,044 Speaker 1: lot of people agree with me. But it's this just 419 00:21:44,924 --> 00:21:47,684 Speaker 1: thinking that, oh, like, because it's like an algorithm and 420 00:21:47,724 --> 00:21:50,644 Speaker 1: a computer, it's free of human bias. And that's just 421 00:21:50,684 --> 00:21:52,964 Speaker 1: not true. It's not free of human bias. It's not 422 00:21:53,004 --> 00:21:55,564 Speaker 1: free of human error. And so you have to think about, 423 00:21:55,564 --> 00:21:58,004 Speaker 1: like who are the people in charge, what kinds of 424 00:21:58,084 --> 00:22:00,844 Speaker 1: choices are they making, what weights are they putting on? 425 00:22:01,524 --> 00:22:04,124 Speaker 1: You know, I'm going to value the life of this 426 00:22:04,204 --> 00:22:07,324 Speaker 1: person versus that person, Nate, if I know that the 427 00:22:07,364 --> 00:22:10,924 Speaker 1: passenger of a weim is a Democrat or a Republican, 428 00:22:11,564 --> 00:22:14,564 Speaker 1: you know, am I going to Am I going to 429 00:22:14,564 --> 00:22:17,684 Speaker 1: make different choices based off the value of their life? 430 00:22:17,684 --> 00:22:21,164 Speaker 1: That sounds I mean, I'm being kind of silly on 431 00:22:21,244 --> 00:22:25,244 Speaker 1: purpose because like, you know, I'm trying to make a 432 00:22:25,324 --> 00:22:31,244 Speaker 1: point where like you understand that these are all objective 433 00:22:31,524 --> 00:22:34,444 Speaker 1: judgments and who knows right in the black box of 434 00:22:34,484 --> 00:22:37,404 Speaker 1: the algorithm how people end up coming up with them, 435 00:22:37,644 --> 00:22:41,364 Speaker 1: And and now imagine that, Okay, fine, like we've we've 436 00:22:41,404 --> 00:22:43,524 Speaker 1: tried to do the best we can we being the 437 00:22:43,644 --> 00:22:46,324 Speaker 1: engineers and the people putting this together. I think we've 438 00:22:46,364 --> 00:22:49,284 Speaker 1: got it, you know, down, and then what happens if 439 00:22:49,284 --> 00:22:53,164 Speaker 1: there's like a camera malfunction or that something is misperceived 440 00:22:53,164 --> 00:22:57,244 Speaker 1: as something else, and you know, car does something horrible 441 00:22:57,284 --> 00:22:59,044 Speaker 1: because it thinks that it's about to hit a dog, 442 00:22:59,084 --> 00:23:01,084 Speaker 1: but it's really not about to hit a dog, right, 443 00:23:01,164 --> 00:23:04,204 Speaker 1: Someone just like left a scooter or something like that. 444 00:23:04,524 --> 00:23:07,364 Speaker 1: Those types of things happen as well, and so you 445 00:23:07,444 --> 00:23:10,724 Speaker 1: have just whenever you a lot of technology, there are 446 00:23:10,764 --> 00:23:14,924 Speaker 1: lots of parts where it can malfunction, it can go wrong. 447 00:23:16,164 --> 00:23:20,004 Speaker 1: So last week, please don't judge me, Nate, I rewatched 448 00:23:20,164 --> 00:23:24,444 Speaker 1: the movie Speed. It didn't hold up very well. I 449 00:23:24,484 --> 00:23:28,484 Speaker 1: have to say I was very disappointed. I'm sorry if 450 00:23:28,524 --> 00:23:30,604 Speaker 1: anyone in the audience is a huge fan of Speed. 451 00:23:30,924 --> 00:23:33,044 Speaker 1: Just it did not hold up very well. The acting 452 00:23:33,164 --> 00:23:36,524 Speaker 1: was horrible. Just everyone was a little bit off. Anyway, 453 00:23:36,724 --> 00:23:39,964 Speaker 1: there's a scene this is relevant. There's a scene in 454 00:23:40,044 --> 00:23:42,884 Speaker 1: Speed and the conceit of the movie. This is not 455 00:23:42,924 --> 00:23:46,124 Speaker 1: a spoiler for anyone who hasn't seen it, because you'll 456 00:23:46,164 --> 00:23:48,524 Speaker 1: figure it out within the first ten minutes. Literally, the 457 00:23:48,564 --> 00:23:50,604 Speaker 1: only conceit of the movie. There's a bomb on the 458 00:23:50,604 --> 00:23:52,564 Speaker 1: bus and the bus cannot go below fifty miles an 459 00:23:52,604 --> 00:23:54,444 Speaker 1: hour if it goes below fifty miles an hour, the 460 00:23:54,484 --> 00:23:56,724 Speaker 1: bus will blow up with everyone on it. That's the 461 00:23:56,844 --> 00:24:01,204 Speaker 1: entire movie. That's the plot. So the bus is going 462 00:24:01,284 --> 00:24:05,404 Speaker 1: and they're all in city streets in La, which obviously 463 00:24:05,524 --> 00:24:07,644 Speaker 1: is not very conducive to going fifty miles an hour. 464 00:24:07,884 --> 00:24:11,364 Speaker 1: By the way night the only implausable point, well, there 465 00:24:11,364 --> 00:24:13,804 Speaker 1: are lots of implausible points of that plot, but somehow, 466 00:24:13,844 --> 00:24:16,364 Speaker 1: on a highway in La it managed to go faster 467 00:24:16,444 --> 00:24:18,164 Speaker 1: than fifty miles an hour. I was like, guys, that's 468 00:24:18,244 --> 00:24:22,244 Speaker 1: just never happening. That is never happening. There's way too 469 00:24:22,364 --> 00:24:25,124 Speaker 1: much traffic in La. Anyway, we're on surface streets, we're 470 00:24:25,164 --> 00:24:27,484 Speaker 1: going over fifty miles an hour, and there's a woman 471 00:24:27,564 --> 00:24:32,164 Speaker 1: with a baby carriage and she is going to get hit. 472 00:24:32,324 --> 00:24:34,244 Speaker 1: The baby carriage is going to get hit by the bus, 473 00:24:34,524 --> 00:24:38,404 Speaker 1: and what we can see as humans, which autonomous vehicle 474 00:24:38,444 --> 00:24:40,444 Speaker 1: might not realize, is that there's no baby in the 475 00:24:40,444 --> 00:24:42,804 Speaker 1: carriage is filled with bottles. So this is, you know, 476 00:24:42,884 --> 00:24:46,164 Speaker 1: someone who was collecting bottles from the street putting them 477 00:24:46,204 --> 00:24:48,084 Speaker 1: in a baby carriage. So the carriage goes flying and 478 00:24:48,124 --> 00:24:51,484 Speaker 1: the bottles shatter, but no baby is actually hurt. But 479 00:24:51,564 --> 00:24:54,924 Speaker 1: I was actually just thinking of this just now when 480 00:24:54,964 --> 00:24:57,724 Speaker 1: we're discussing this. You know, what if that happens and 481 00:24:57,804 --> 00:25:01,484 Speaker 1: you're in an autonomous vehicle and the autonomous vehicle doesn't realize, 482 00:25:01,524 --> 00:25:05,164 Speaker 1: you know, baby carriage doesn't necessarily mean baby. You have 483 00:25:05,204 --> 00:25:07,324 Speaker 1: to kind of figure it out and like does something 484 00:25:07,364 --> 00:25:10,524 Speaker 1: that will endanger all the people in the car on 485 00:25:10,564 --> 00:25:12,484 Speaker 1: the bus because they don't want to hit the baby 486 00:25:12,564 --> 00:25:16,164 Speaker 1: because priority says save baby. It's not even a baby, right, 487 00:25:16,204 --> 00:25:18,724 Speaker 1: but it's in a baby carriage. So you know, all 488 00:25:18,764 --> 00:25:21,164 Speaker 1: of these things. This goes back to what we talk 489 00:25:21,204 --> 00:25:23,604 Speaker 1: about often that a lot of the AI can be 490 00:25:23,804 --> 00:25:26,284 Speaker 1: and all of these things can be really really smart, 491 00:25:26,444 --> 00:25:30,204 Speaker 1: but there are certain things that are second nature to humans. 492 00:25:30,244 --> 00:25:32,404 Speaker 1: We don't even realize we're making these judgments that are 493 00:25:32,404 --> 00:25:35,644 Speaker 1: really really difficult to replicate. Right, there's something that we 494 00:25:35,764 --> 00:25:39,324 Speaker 1: just do unthinkingly, but that it's very hard to program 495 00:25:39,364 --> 00:25:41,484 Speaker 1: into an algorithm, into a computer or teach them how 496 00:25:41,524 --> 00:25:42,324 Speaker 1: to do that correctly. 497 00:25:44,004 --> 00:25:44,444 Speaker 3: Yeah. 498 00:25:44,604 --> 00:25:49,804 Speaker 2: On the flip side, when algorithms make counterintuitive quote unquote 499 00:25:49,844 --> 00:25:53,804 Speaker 2: decisions even when they're correct. And believe me, I'm not 500 00:25:53,964 --> 00:26:00,244 Speaker 2: the world's most unskeptical admirer of algorithms. I think there 501 00:26:00,244 --> 00:26:02,604 Speaker 2: are lots of bad algorithms and bad models out there. 502 00:26:02,684 --> 00:26:05,684 Speaker 2: I've built lots of good models and some bad models myself, 503 00:26:05,724 --> 00:26:08,764 Speaker 2: probably over the years and seeing and seeing them from others. 504 00:26:08,844 --> 00:26:09,004 Speaker 3: Right. 505 00:26:09,044 --> 00:26:12,524 Speaker 2: But you know, I was in Florida, like I said, 506 00:26:12,684 --> 00:26:16,084 Speaker 2: and my friend was driving me back to my hotel 507 00:26:16,164 --> 00:26:19,524 Speaker 2: from dinner. Nice of him, but we'd had a little 508 00:26:19,564 --> 00:26:21,564 Speaker 2: wine at dinner, and he's like, I have a test. 509 00:26:21,724 --> 00:26:26,124 Speaker 2: I'm just gonna turn automated mode on and not touch 510 00:26:26,164 --> 00:26:28,324 Speaker 2: the steering wheel, and like it did pretty well, but 511 00:26:28,364 --> 00:26:31,204 Speaker 2: like it took a route that he would never have 512 00:26:31,244 --> 00:26:33,924 Speaker 2: taken back to the hotel, right, And like he's. 513 00:26:33,764 --> 00:26:34,764 Speaker 3: Like, oh, actually this makes sense. 514 00:26:34,804 --> 00:26:36,964 Speaker 2: I understand why it would do it, and so and so, 515 00:26:37,124 --> 00:26:39,964 Speaker 2: like you know, and that's a very low stakes decision, right, 516 00:26:40,044 --> 00:26:43,284 Speaker 2: But like you know, often with these directions, it's like, Okay, 517 00:26:43,404 --> 00:26:45,364 Speaker 2: if people have this intuition that you want to take 518 00:26:45,444 --> 00:26:49,324 Speaker 2: the most direct route, whereas oftentimes going out of your 519 00:26:49,364 --> 00:26:53,204 Speaker 2: way is in principle faster, you could probably test that 520 00:26:53,364 --> 00:26:56,684 Speaker 2: robustly or not or things like that, right, or maybe 521 00:26:56,724 --> 00:26:59,284 Speaker 2: there's a safety thing that seemed like it's dangerous, Like, 522 00:26:59,484 --> 00:27:00,604 Speaker 2: but if you deviate. 523 00:27:00,284 --> 00:27:01,564 Speaker 3: From like the norm. 524 00:27:01,844 --> 00:27:05,444 Speaker 2: Yea, then then you get punished for it, right, and 525 00:27:05,484 --> 00:27:08,724 Speaker 2: so like it might be that like, obviously these vehicle 526 00:27:08,724 --> 00:27:13,004 Speaker 2: should be regular. It might be that regulations that require 527 00:27:13,244 --> 00:27:18,164 Speaker 2: more like explicit and transparent thinking, yeah, are actually less safe. 528 00:27:18,164 --> 00:27:20,004 Speaker 2: I want to be very careful with that. I don't 529 00:27:20,044 --> 00:27:23,964 Speaker 2: think that's the case for like really large language models, 530 00:27:24,004 --> 00:27:26,804 Speaker 2: for example. I think we need far more transparency with those. 531 00:27:26,844 --> 00:27:28,404 Speaker 2: I think they fail in ways that. 532 00:27:28,324 --> 00:27:30,724 Speaker 3: Are a little unpredictable. 533 00:27:31,924 --> 00:27:35,884 Speaker 2: But the problem is, like, yeah, people will blame these 534 00:27:36,004 --> 00:27:39,084 Speaker 2: cars even if they have like a you know whatever 535 00:27:39,164 --> 00:27:41,884 Speaker 2: twenty or twenty five percent the accident rates that humans do, 536 00:27:42,124 --> 00:27:47,044 Speaker 2: and and and plus you have vested interest, right. You know, 537 00:27:47,284 --> 00:27:49,004 Speaker 2: in New York it's a city with a lot of 538 00:27:49,124 --> 00:27:54,364 Speaker 2: union activity. You know, New York also has these you know, 539 00:27:54,564 --> 00:27:58,684 Speaker 2: amazingly expensive frankly ubers and lifts and things like that. 540 00:27:58,684 --> 00:28:01,284 Speaker 2: It's amazing when you go to like a city like 541 00:28:01,284 --> 00:28:03,244 Speaker 2: like in Florida and you're like, oh, man, eight bucks, 542 00:28:03,284 --> 00:28:04,924 Speaker 2: you get across town. That's amazing, right, Not in the 543 00:28:04,924 --> 00:28:06,164 Speaker 2: case in Manhattan and Brooklyn. 544 00:28:06,644 --> 00:28:07,564 Speaker 1: No, certainly not. 545 00:28:14,164 --> 00:28:16,084 Speaker 3: And we'll be right back after this break. 546 00:28:26,084 --> 00:28:29,644 Speaker 1: Even simple algorithms like you know, what's the best route 547 00:28:29,644 --> 00:28:34,924 Speaker 1: to take. I've been completely screwed over in recent months 548 00:28:34,924 --> 00:28:40,324 Speaker 1: by Google, which will default instead of the fastest it 549 00:28:40,364 --> 00:28:43,924 Speaker 1: will go to the eco friendly way to drive. And 550 00:28:43,924 --> 00:28:47,044 Speaker 1: I'm like, how are you even defining eco friendly? Like 551 00:28:47,324 --> 00:28:49,884 Speaker 1: are you is it because I'm saving gas? Because I'm 552 00:28:49,924 --> 00:28:52,004 Speaker 1: not breaking as much? Am I saving the environment? What 553 00:28:52,164 --> 00:28:55,044 Speaker 1: exactly am I doing? And first of all, I didn't 554 00:28:55,044 --> 00:28:57,244 Speaker 1: say I wanted to do this. Secondly, I don't know 555 00:28:57,284 --> 00:28:59,844 Speaker 1: what you mean by eco friendly, Like third of all, 556 00:28:59,884 --> 00:29:01,764 Speaker 1: just tell me that I'm not going the fastest way 557 00:29:01,804 --> 00:29:04,244 Speaker 1: instead of defaulting to it. So I think that in 558 00:29:04,284 --> 00:29:07,844 Speaker 1: that sense, we want to be explicit in the algorithms. 559 00:29:07,844 --> 00:29:09,964 Speaker 1: And I think you need to make think there. Does 560 00:29:11,044 --> 00:29:13,684 Speaker 1: I see a future where like there's human input into 561 00:29:13,724 --> 00:29:15,724 Speaker 1: it where you're like, I want to take the way mo, 562 00:29:16,044 --> 00:29:18,644 Speaker 1: and I like this is my priority, right, like get 563 00:29:18,644 --> 00:29:21,284 Speaker 1: me here fastest or I don't care. I want to 564 00:29:21,324 --> 00:29:25,284 Speaker 1: see neck drive like whatever it is, Like that would 565 00:29:25,324 --> 00:29:27,884 Speaker 1: be that would be an interesting way of doing it, 566 00:29:27,924 --> 00:29:30,604 Speaker 1: as long as you know the model doesn't lie to 567 00:29:30,644 --> 00:29:32,804 Speaker 1: you and it's not like, oh, this is the fastest way, 568 00:29:32,964 --> 00:29:37,444 Speaker 1: right because I'm programmed to take these other things into consideration, 569 00:29:37,564 --> 00:29:39,804 Speaker 1: but I don't want to piss you off. 570 00:29:40,764 --> 00:29:42,284 Speaker 2: Yeah, look, I mean part of the question is like 571 00:29:42,284 --> 00:29:47,684 Speaker 2: what is the algorithm solving for? Right, So, like I 572 00:29:47,724 --> 00:29:52,084 Speaker 2: now live in the East Village, which is very dependent 573 00:29:52,244 --> 00:29:55,284 Speaker 2: on a couple of crosstown subway lines, like the L, 574 00:29:55,364 --> 00:29:58,884 Speaker 2: and so it's actually pretty good overall for subway service, 575 00:29:58,924 --> 00:30:03,284 Speaker 2: but you're almost always making transfers to get most places, 576 00:30:04,084 --> 00:30:07,204 Speaker 2: and like you know, there is a lot of redundancy 577 00:30:07,324 --> 00:30:10,004 Speaker 2: typically in which transfer or you make. 578 00:30:10,044 --> 00:30:12,084 Speaker 3: And like I don't think. 579 00:30:11,924 --> 00:30:16,524 Speaker 2: The Google subway directions are often very good in terms 580 00:30:16,564 --> 00:30:19,164 Speaker 2: of matching my intuitions about like where's the right place 581 00:30:19,204 --> 00:30:21,604 Speaker 2: to make it? Yeah, a transfer, right, Like you know, 582 00:30:21,724 --> 00:30:23,324 Speaker 2: just the kind of combination of the fact that, like 583 00:30:23,764 --> 00:30:30,444 Speaker 2: some lines are more reliable than others, some some actual transitions, right, 584 00:30:30,644 --> 00:30:33,724 Speaker 2: Like you know Union Square, if you ever in a 585 00:30:33,804 --> 00:30:36,524 Speaker 2: Union Square in Manhattan, all those you know, it's very 586 00:30:36,564 --> 00:30:38,564 Speaker 2: crowd of those platforms, right, And like there are some 587 00:30:38,604 --> 00:30:40,364 Speaker 2: other stops where just like you know, I don't want 588 00:30:40,364 --> 00:30:43,164 Speaker 2: to have to change there if there's another alternative, and 589 00:30:43,204 --> 00:30:48,044 Speaker 2: like I don't think it quite understands that adequately, And 590 00:30:48,084 --> 00:30:50,124 Speaker 2: like the fuzziness around, like I don't actually have to 591 00:30:50,124 --> 00:30:53,764 Speaker 2: get this exact location, just somewhere kind of like near there, right, 592 00:30:54,004 --> 00:30:55,924 Speaker 2: I don't think it's fantastic, but like, what are are 593 00:30:55,964 --> 00:30:56,724 Speaker 2: you trying to sell it for? 594 00:30:56,844 --> 00:30:59,244 Speaker 3: Like time or like ease of use? 595 00:30:59,364 --> 00:31:01,084 Speaker 2: You know, in theory, it's like, Okay, you can get 596 00:31:01,084 --> 00:31:04,724 Speaker 2: there two minutes faster if you transfer twice instead of once. Okay, 597 00:31:04,764 --> 00:31:06,364 Speaker 2: Well a I don't really buy I think there's too 598 00:31:06,404 --> 00:31:10,644 Speaker 2: much risk in each transfer, right, that something's wrong, right, 599 00:31:10,684 --> 00:31:11,524 Speaker 2: take a wrong turn. 600 00:31:11,684 --> 00:31:11,844 Speaker 1: Right. 601 00:31:12,044 --> 00:31:14,044 Speaker 3: Depends on a lot of things. 602 00:31:14,124 --> 00:31:16,524 Speaker 2: But yeah, if you don't know what the algorithm is 603 00:31:16,564 --> 00:31:20,484 Speaker 2: solving for exactly, then it's not garbage in garbage out, 604 00:31:20,484 --> 00:31:23,204 Speaker 2: but it's inherently inadequate to the problem unless you're kind 605 00:31:23,204 --> 00:31:25,284 Speaker 2: of like and probably implicitly it's like some notion of 606 00:31:25,324 --> 00:31:30,564 Speaker 2: like utility or satisfaction or whatever else. 607 00:31:30,924 --> 00:31:33,124 Speaker 1: Yeah, No, I mean I think that and a subway 608 00:31:33,164 --> 00:31:36,444 Speaker 1: problem is much simpler than all the different things that 609 00:31:36,484 --> 00:31:39,244 Speaker 1: a self driving car is going to need to solve 610 00:31:39,284 --> 00:31:43,084 Speaker 1: for on any given basis and at any given point 611 00:31:43,084 --> 00:31:46,084 Speaker 1: in time. So I think that these are just illustrations 612 00:31:46,124 --> 00:31:48,364 Speaker 1: of a lot of the things, a lot of the 613 00:31:48,444 --> 00:31:50,804 Speaker 1: challenges that remain a lot of the value judgments that 614 00:31:50,924 --> 00:31:54,644 Speaker 1: have to be made as this technology gets off the ground. Now, 615 00:31:54,844 --> 00:31:58,124 Speaker 1: I'm also worried, Nate, Let's just imagine, for the sake 616 00:31:58,124 --> 00:32:05,804 Speaker 1: of argument, that everything, like all these algorithms are evolved perfectly, 617 00:32:06,084 --> 00:32:09,124 Speaker 1: right that, And like I said, for the sake of argument, 618 00:32:09,444 --> 00:32:12,244 Speaker 1: let's just say that like they've solved the trolley problem 619 00:32:12,524 --> 00:32:16,364 Speaker 1: and they are optimized, like they're doing everything that we 620 00:32:16,444 --> 00:32:19,284 Speaker 1: think they should be doing, and like the technology is 621 00:32:19,364 --> 00:32:22,724 Speaker 1: now like beautiful and the cars can be driving beautifully. 622 00:32:22,884 --> 00:32:25,564 Speaker 1: There's one other element that I worry about, which is 623 00:32:25,764 --> 00:32:30,564 Speaker 1: security of the actual systems, in terms of hacking and 624 00:32:30,644 --> 00:32:34,764 Speaker 1: in terms of bad actors being able to get into 625 00:32:35,164 --> 00:32:38,964 Speaker 1: these autonomous vehicles, take them over, sabotage them, whatever it is. 626 00:32:39,084 --> 00:32:43,164 Speaker 1: This is not me being conspiracy theory minded. Basically every 627 00:32:43,164 --> 00:32:47,044 Speaker 1: single smart car system has been hacked into at some point, 628 00:32:47,644 --> 00:32:50,804 Speaker 1: including things that seem ridiculous, and not just smart cars, 629 00:32:50,844 --> 00:32:53,924 Speaker 1: smart homes. Right, Like, there are people who like hack 630 00:32:53,964 --> 00:32:57,124 Speaker 1: into homes and like turn your heat to one hundred 631 00:32:57,244 --> 00:32:59,724 Speaker 1: or make it freezing, and do things that seem like 632 00:32:59,764 --> 00:33:02,164 Speaker 1: silly pranks, but they're not right, Like they're doing this 633 00:33:02,244 --> 00:33:06,244 Speaker 1: with a very malicious intent, Like people's toasters, Like anything 634 00:33:06,284 --> 00:33:10,444 Speaker 1: that's smart that's connected to the grid, like has been 635 00:33:10,524 --> 00:33:13,524 Speaker 1: actually taken over at some point, whether to prove a 636 00:33:13,564 --> 00:33:17,364 Speaker 1: point or sometimes for a malicious purpose. And cars, like 637 00:33:17,444 --> 00:33:20,204 Speaker 1: if you think about an autonomous system that can be 638 00:33:20,284 --> 00:33:24,644 Speaker 1: taken over and whether it's cars or something else, like 639 00:33:24,844 --> 00:33:26,804 Speaker 1: this isn't the I mean, it could be the plot 640 00:33:26,844 --> 00:33:29,124 Speaker 1: of a movie, but like this is something that's actually 641 00:33:30,484 --> 00:33:34,324 Speaker 1: not It's quite reasonable to assume that there are bad 642 00:33:34,364 --> 00:33:37,004 Speaker 1: actors who will try to break into any system, just 643 00:33:37,044 --> 00:33:41,244 Speaker 1: like energy system, powagrids, et cetera. Like that's that's something 644 00:33:41,284 --> 00:33:45,124 Speaker 1: that has happened. It will continue to happen, and people, 645 00:33:45,644 --> 00:33:48,084 Speaker 1: you know, in their quest to kind of evolve and 646 00:33:48,124 --> 00:33:52,644 Speaker 1: move forward, safety in that sense is not often the 647 00:33:52,764 --> 00:33:56,324 Speaker 1: number one concern, and so they're end up being you know, 648 00:33:56,324 --> 00:33:59,684 Speaker 1: there end up being vulnerabilities that are only discovered after 649 00:33:59,724 --> 00:34:03,684 Speaker 1: the fact. And if you think about what might happen, 650 00:34:03,724 --> 00:34:06,164 Speaker 1: like it's a constant arms race, you know, between bad 651 00:34:06,204 --> 00:34:09,244 Speaker 1: actor hackers and people trying to you know, the white 652 00:34:09,284 --> 00:34:13,364 Speaker 1: hackers like good actors who are trying to prevent vulnerabilities, 653 00:34:13,444 --> 00:34:18,524 Speaker 1: prevent security technology breaches, and it's this constant war and 654 00:34:18,644 --> 00:34:21,524 Speaker 1: if you think about now a future where you know, 655 00:34:21,604 --> 00:34:24,604 Speaker 1: we have lots of fleets of autonomous vehicles, I think 656 00:34:24,604 --> 00:34:26,564 Speaker 1: that's something that we do need to think about and 657 00:34:26,684 --> 00:34:29,084 Speaker 1: worry about as well. And like I said, like I 658 00:34:29,084 --> 00:34:31,804 Speaker 1: don't think that this is being conspiracy theorist or just 659 00:34:32,564 --> 00:34:36,484 Speaker 1: are very alarmist. I actually think that this is quite real. 660 00:34:36,604 --> 00:34:40,284 Speaker 1: It happens all the time, you know, people things are 661 00:34:40,284 --> 00:34:43,684 Speaker 1: hacked all the time, and as you know, as we 662 00:34:43,724 --> 00:34:46,684 Speaker 1: are more and more reliant on it, you realize just 663 00:34:46,764 --> 00:34:51,204 Speaker 1: how big of an impact hacks can have. They If 664 00:34:51,244 --> 00:34:54,284 Speaker 1: you remember, like six months ago, Las Vegas kind of 665 00:34:54,284 --> 00:34:57,364 Speaker 1: came the strip came to a standstill when someone hacked 666 00:34:57,404 --> 00:35:00,164 Speaker 1: into all the MGM systems. Right, no one could check 667 00:35:00,204 --> 00:35:02,364 Speaker 1: in to their rooms. People couldn't even get into their 668 00:35:02,444 --> 00:35:05,364 Speaker 1: rooms like people could. It was just absolutely a shit 669 00:35:05,404 --> 00:35:07,804 Speaker 1: shown that happened to the Caesar's properties. Turns out, by 670 00:35:07,804 --> 00:35:10,764 Speaker 1: the way, I was a teenager, but like it was 671 00:35:10,924 --> 00:35:13,564 Speaker 1: just an absolute nightmare. And that's just like that's a 672 00:35:13,604 --> 00:35:14,204 Speaker 1: tiny thing. 673 00:35:16,244 --> 00:35:20,124 Speaker 2: Yeah, not fully automating things with AI, because when AI 674 00:35:20,164 --> 00:35:27,524 Speaker 2: systems fail, they don't fail particularly robustly necessarily. 675 00:35:27,644 --> 00:35:29,084 Speaker 3: I mean, there's also just some shit like. 676 00:35:30,804 --> 00:35:34,684 Speaker 2: You know, probably if humans are picking out driving routes 677 00:35:34,724 --> 00:35:37,964 Speaker 2: like from JFK back to Brooklyn or Manhattan, right, probably 678 00:35:38,004 --> 00:35:40,844 Speaker 2: you want some degree your randomization. You know, you probably 679 00:35:40,884 --> 00:35:43,324 Speaker 2: have a mixed strategy in game theory terms, where like 680 00:35:43,364 --> 00:35:45,884 Speaker 2: if everybody takes the most optimal route, then it's still 681 00:35:45,924 --> 00:35:48,084 Speaker 2: longer the most optmal route because it gets more crowded. 682 00:35:48,204 --> 00:35:50,244 Speaker 3: Right, you know, I don't know if there's some of that. 683 00:35:50,284 --> 00:35:52,124 Speaker 3: I was just in. 684 00:35:51,564 --> 00:35:54,324 Speaker 2: Miami estate that can have horrible traffic problems, so they're 685 00:35:54,324 --> 00:35:57,724 Speaker 2: also very like patchy. I think it's more an infrastructure 686 00:35:57,764 --> 00:36:01,964 Speaker 2: matter than like a matter of AI decision making per se. Right, 687 00:36:02,004 --> 00:36:07,084 Speaker 2: But like, but yeah, no, look anything where there's not 688 00:36:07,164 --> 00:36:10,844 Speaker 2: like a kill switch, get I get nervous about, yeah, 689 00:36:10,964 --> 00:36:12,604 Speaker 2: or a redundant backup. 690 00:36:12,404 --> 00:36:15,644 Speaker 1: And oh, by the way, it doesn't it doesn't even 691 00:36:15,684 --> 00:36:16,964 Speaker 1: have to be hacking. By the way, I was just 692 00:36:16,964 --> 00:36:22,324 Speaker 1: thinking about the AWS outage that recently happened, where people's 693 00:36:22,484 --> 00:36:25,484 Speaker 1: smart beds got affected and like people got woken up 694 00:36:25,484 --> 00:36:27,444 Speaker 1: in the middle of the night because the beds started 695 00:36:27,484 --> 00:36:30,044 Speaker 1: doing something really weird and then got caught in these 696 00:36:30,164 --> 00:36:33,684 Speaker 1: ridiculous positions and there was nothing they could do because 697 00:36:33,724 --> 00:36:36,204 Speaker 1: there was no redundancy, and like the entire system had 698 00:36:36,244 --> 00:36:36,644 Speaker 1: gone off. 699 00:36:36,644 --> 00:36:41,044 Speaker 2: It's a smart mattress or smart toothbrush or smart clipper. 700 00:36:40,924 --> 00:36:43,124 Speaker 1: But exactly no, neither do I. But now imagine if 701 00:36:43,164 --> 00:36:45,044 Speaker 1: you have to be in a smart car, right, because 702 00:36:45,564 --> 00:36:47,964 Speaker 1: autonomous vehicles are the thing. That's the thing that I 703 00:36:48,044 --> 00:36:51,604 Speaker 1: actually like that. I think that we don't worry enough about, right, 704 00:36:51,644 --> 00:36:53,364 Speaker 1: Like what happens if there are these outages? 705 00:36:53,444 --> 00:36:55,324 Speaker 3: Well, and we have these coercive like if you go 706 00:36:55,404 --> 00:36:59,204 Speaker 3: I've been in where was I in? Norway? Right? 707 00:36:59,804 --> 00:37:04,964 Speaker 2: When ever the car goes over this feed limit, you 708 00:37:05,044 --> 00:37:11,124 Speaker 2: get like a little same thing in Korea, right, So 709 00:37:11,324 --> 00:37:13,164 Speaker 2: like so that can be a. 710 00:37:13,084 --> 00:37:15,564 Speaker 3: Little bit coercive. 711 00:37:15,604 --> 00:37:18,444 Speaker 2: I don't know, I know how I feel about that, right, Yeah, 712 00:37:18,564 --> 00:37:21,204 Speaker 2: but you're certainly giving like the government a lot of control. 713 00:37:20,884 --> 00:37:24,404 Speaker 1: Over, Yeah, you are, and you're giving up and you're 714 00:37:24,444 --> 00:37:27,844 Speaker 1: also giving up a ton of privacy because you know, 715 00:37:27,924 --> 00:37:31,644 Speaker 1: you're presumably you are associated with you know, all the 716 00:37:31,724 --> 00:37:34,044 Speaker 1: routes you take, everything you do, all your habits. So 717 00:37:34,124 --> 00:37:37,924 Speaker 1: I think there are these downstream concerns that we haven't 718 00:37:37,964 --> 00:37:40,804 Speaker 1: really reached yet because we're still figuring out by we, 719 00:37:40,964 --> 00:37:43,724 Speaker 1: I mean the companies who are designing autonomous cars. 720 00:37:43,884 --> 00:37:45,564 Speaker 3: I don't mean me. 721 00:37:45,604 --> 00:37:48,964 Speaker 1: Personally, but I think we're still figuring out like the 722 00:37:49,044 --> 00:37:52,124 Speaker 1: algorithms and all this stuff. But like, even downstream, there 723 00:37:52,124 --> 00:37:55,644 Speaker 1: are lots of things to be thinking about, safety, security, 724 00:37:56,084 --> 00:37:59,964 Speaker 1: redundancy of you know, all of this technology, privacy concerns, 725 00:38:00,284 --> 00:38:04,164 Speaker 1: control like when when people can actually like override, who 726 00:38:04,204 --> 00:38:07,244 Speaker 1: can override your car? Who can make those types of decisions. 727 00:38:07,924 --> 00:38:11,044 Speaker 1: I think that these are really important questions. By the way, 728 00:38:11,364 --> 00:38:13,964 Speaker 1: there was a huge you know, there have been huge 729 00:38:14,884 --> 00:38:18,484 Speaker 1: outcries in the last few years where all of a sudden, 730 00:38:18,484 --> 00:38:21,044 Speaker 1: features of cars that you paid for stopped working because 731 00:38:21,044 --> 00:38:22,924 Speaker 1: the company decided they wanted you to pay for them, 732 00:38:22,964 --> 00:38:26,684 Speaker 1: like heated seats. Right. Like now, imagine you buy a 733 00:38:26,724 --> 00:38:29,444 Speaker 1: car and you don't have a choice anymore to buy 734 00:38:29,684 --> 00:38:32,724 Speaker 1: something where it's not smart and you can control everything, 735 00:38:33,084 --> 00:38:34,924 Speaker 1: and you're just giving all of that up, and all 736 00:38:34,924 --> 00:38:38,004 Speaker 1: of a sudden, they're like, oh, it's now a subscription model. Sorry, Nate, 737 00:38:38,124 --> 00:38:40,844 Speaker 1: you can't go to where you're going today because you 738 00:38:40,964 --> 00:38:44,124 Speaker 1: haven't paid for the upgrades to this feature. I am 739 00:38:44,164 --> 00:38:46,644 Speaker 1: going to lock your car down. So there are all 740 00:38:46,724 --> 00:38:49,364 Speaker 1: of these different considerations that I don't think you know, 741 00:38:49,444 --> 00:38:51,564 Speaker 1: we've only touched the tip of the iceberg when it 742 00:38:51,564 --> 00:39:04,684 Speaker 1: comes to this. Let us know what you think of 743 00:39:04,724 --> 00:39:07,404 Speaker 1: the show. Reach out to us at Risky Business at 744 00:39:07,444 --> 00:39:12,564 Speaker 1: pushkin dot Fm. Risky Business is hosted by me Maria Kanakova. 745 00:39:12,284 --> 00:39:13,604 Speaker 3: And by me Nate Silver. 746 00:39:14,244 --> 00:39:17,484 Speaker 2: The show was a cool production of Pushing Industries and iHeartMedia. 747 00:39:17,964 --> 00:39:21,284 Speaker 2: This episode was produced by Isaac Carter. Our associate producer 748 00:39:21,364 --> 00:39:25,124 Speaker 2: is Sonya gerwit Lydia, Jean Kott and Daphne Chen are 749 00:39:25,124 --> 00:39:29,284 Speaker 2: our editors, and our executive producer is Jacob Goldstein. Mixing 750 00:39:29,364 --> 00:39:30,244 Speaker 2: by Sarah Bruger. 751 00:39:30,804 --> 00:39:33,004 Speaker 1: If you like the show, please rate and review us 752 00:39:33,044 --> 00:39:35,684 Speaker 1: so other people can find us too, But once again, 753 00:39:35,844 --> 00:39:37,644 Speaker 1: only if you like us. We don't want those bad 754 00:39:37,684 --> 00:39:39,844 Speaker 1: reviews out there. Thanks for tuning in.