WEBVTT 00:00:01.960 --> 00:00:05.540 You're listening to There Auto Be A Law, the 00:00:05.540 --> 00:00:08.259 Center for Auto Safety podcast with Executive 00:00:08.259 --> 00:00:11.720 Director Michael Brooks, Chief Engineer Fred 00:00:11.720 --> 00:00:15.640 Perkins, and hosted by me, Anthony Cimino. For 00:00:15.640 --> 00:00:18.300 over 50 years, the Center for Auto Safety has 00:00:18.300 --> 00:00:28.480 worked to make cars safer. Hey, listeners, welcome 00:00:28.480 --> 00:00:31.079 to another exciting episode of what's going on 00:00:31.079 --> 00:00:33.579 with the reflecting pool. Wait, that's not the 00:00:33.579 --> 00:00:37.100 show. Anyway, today is Wednesday, June 24th, 00:00:37.100 --> 00:00:39.579 and we're going to start off with a little publication 00:00:39.579 --> 00:00:44.259 called The New York Times. They they have this. 00:00:44.299 --> 00:00:46.619 I'm very happy. They have this big article on 00:00:46.619 --> 00:00:49.520 how SUVs in particular have gotten bigger and 00:00:49.520 --> 00:00:53.039 there's. The problem with rear over and forward 00:00:53.039 --> 00:00:54.979 over, which we've, I think, talked about since 00:00:54.979 --> 00:00:57.000 episode one. The Center for No Savings has been 00:00:57.000 --> 00:01:00.920 covering this for 20 plus years. But the New 00:01:00.920 --> 00:01:03.859 York Times just discovered this. And their writers 00:01:03.859 --> 00:01:07.819 are teenagers. Because I'm going to quote directly 00:01:07.819 --> 00:01:10.319 from this. The surge in pedestrian deaths has 00:01:10.319 --> 00:01:14.810 baffled researchers. You know. After analyzing 00:01:14.810 --> 00:01:17.010 federal and industry records, including never 00:01:17.010 --> 00:01:19.709 before examined data on vehicle dimensions, we 00:01:19.709 --> 00:01:22.469 found that the rise of large pickups and SUVs 00:01:22.469 --> 00:01:26.950 is an important factor. What? That's a little 00:01:26.950 --> 00:01:29.409 over the top, right? They're trying to present 00:01:29.409 --> 00:01:32.390 this as a novel idea, I guess, that they discovered. 00:01:32.510 --> 00:01:34.670 But obviously, if you've been listening to us 00:01:34.670 --> 00:01:36.489 for four years, we've talked about a lot. Yeah. 00:01:36.750 --> 00:01:40.290 And so further quoting, our estimate is. Estimate 00:01:40.290 --> 00:01:42.569 is that about 200 to 400 pedestrians a year would 00:01:42.569 --> 00:01:44.469 not have died if vehicles had remained approximately 00:01:44.469 --> 00:01:46.209 the same size over the past quarter century. 00:01:46.549 --> 00:01:49.469 That represents about 10 % of the recent increase 00:01:49.469 --> 00:01:52.650 in pedestrian deaths. Well, hey, New York Times, 00:01:52.790 --> 00:01:56.750 round of applause for discovering the year 2000. 00:01:57.349 --> 00:02:01.390 I mean, look, it's a great article. It is. I 00:02:01.390 --> 00:02:04.049 mean, I really like a lot of these articles that 00:02:04.049 --> 00:02:06.549 have been kind of coming out in the last two 00:02:06.549 --> 00:02:10.099 to three or four years. Scroll through it and 00:02:10.099 --> 00:02:13.080 they integrate video and explanation, visual 00:02:13.080 --> 00:02:15.400 explanations other than text into the article. 00:02:15.460 --> 00:02:17.120 I think those are super cool. I think they did 00:02:17.120 --> 00:02:20.860 a good job showing the vision of the driver making 00:02:20.860 --> 00:02:23.599 left turns when a pedestrian is there, you know, 00:02:23.599 --> 00:02:28.560 kind of showing how this giant vehicle problem 00:02:28.560 --> 00:02:31.060 that we're the growing vehicle problem that we're 00:02:31.060 --> 00:02:33.659 experiencing in America is impacting both drivers 00:02:33.659 --> 00:02:37.680 and pedestrians. So great article. Yeah, but 00:02:37.680 --> 00:02:39.680 we also have to give the New York Times credit 00:02:39.680 --> 00:02:42.780 for giving the interns from the Columbia School 00:02:42.780 --> 00:02:44.840 of Journalism a chance to actually see their 00:02:44.840 --> 00:02:48.039 work in print. I think that's too generous. No, 00:02:48.139 --> 00:02:50.860 I OK. Enough sarcasm aside, Michael, you're absolutely 00:02:50.860 --> 00:02:52.800 right. I think this is a great article. They 00:02:52.800 --> 00:02:55.240 really highlight something that most people don't 00:02:55.240 --> 00:02:57.240 think about. Unfortunately, the three of us are 00:02:57.240 --> 00:03:00.259 not most people. Or kids in car safety, like, 00:03:00.340 --> 00:03:02.919 you know, this is, you know, or Consumer Reports. 00:03:03.180 --> 00:03:05.139 Wait a second. I was going to say Consumer Reports 00:03:05.139 --> 00:03:07.340 wrote about this like 10 years, 15 years ago. 00:03:07.520 --> 00:03:09.400 But anyway, hey, it's great. The New York Times 00:03:09.400 --> 00:03:12.199 put this out there. I think one very positive 00:03:12.199 --> 00:03:15.060 thing that came out of this is right after this, 00:03:15.120 --> 00:03:17.240 a couple of Congress people are like, wait a 00:03:17.240 --> 00:03:19.979 second, NHTSA, you should look into this, which 00:03:19.979 --> 00:03:24.419 is great. Congressman Mary Gay Scanlon, Mike 00:03:24.419 --> 00:03:27.990 Thompson, Laura Friedman. They introduced on 00:03:27.990 --> 00:03:31.550 last week the Pedestrian Protection Act to address 00:03:31.550 --> 00:03:33.889 the rapid rise of pedestrian fatalities. Yes. 00:03:33.930 --> 00:03:36.550 They actually they reintroduced that. It was 00:03:36.550 --> 00:03:38.889 introduced last year, two years ago, I believe. 00:03:38.949 --> 00:03:41.169 But they reintroduced that actually before the 00:03:41.169 --> 00:03:44.770 New York Times article. So it wasn't a response. 00:03:44.870 --> 00:03:48.710 It was it preceded it. But the Pedestrian Protection 00:03:48.710 --> 00:03:52.169 Act, you know, essentially makes NHTSA dive into 00:03:52.169 --> 00:03:54.280 this issue. They've got to look at. visibility. 00:03:54.479 --> 00:03:56.219 They've got to look at vehicle height, vehicle 00:03:56.219 --> 00:03:58.439 weight. They've got to look at essentially everything 00:03:58.439 --> 00:04:03.439 that is contributing to this problem and also 00:04:03.439 --> 00:04:06.180 incorporate, you know, the larger vehicle testing 00:04:06.180 --> 00:04:09.710 and all of this into NCAP. so that consumers 00:04:09.710 --> 00:04:13.150 have that availability as a rating when they're 00:04:13.150 --> 00:04:15.349 buying their vehicle. At least those consumers 00:04:15.349 --> 00:04:18.209 who care about their impact on other folks on 00:04:18.209 --> 00:04:21.089 the roads would pay attention to it, which is 00:04:21.089 --> 00:04:24.870 great. But you see in the New York Times article 00:04:24.870 --> 00:04:27.430 kind of why we're at where we're at, because 00:04:27.430 --> 00:04:30.949 the industry is passing the buck, right? It's 00:04:30.949 --> 00:04:33.370 nothing to see here, move along. So we've got, 00:04:33.509 --> 00:04:36.290 and these would be great gas lights. I hope I'm 00:04:36.290 --> 00:04:38.970 not stepping on any of y 'all's toes, but. We've 00:04:38.970 --> 00:04:42.129 got Mike Levine or Mike Levin at Ford saying, 00:04:42.290 --> 00:04:45.329 while vehicle safety is critical, blaming larger 00:04:45.329 --> 00:04:48.230 vehicles for pedestrian deaths overlooks systemic 00:04:48.230 --> 00:04:52.639 issues, including the design of roads. Yeah, 00:04:52.639 --> 00:04:55.399 thanks, bud. You can address both at once, right? 00:04:55.620 --> 00:04:58.319 I mean, you know, nothing to see here, Ford says. 00:04:58.579 --> 00:05:01.680 We need better roads, not the ability for the 00:05:01.680 --> 00:05:03.360 driver to actually see people out of the damn 00:05:03.360 --> 00:05:07.199 car. Makes no sense. And then, you know, General 00:05:07.199 --> 00:05:10.439 Motors came in and said, you know, and here's 00:05:10.439 --> 00:05:12.040 part of the article, I'll just quote it. Automakers 00:05:12.040 --> 00:05:14.540 say that new technology designed to detect and 00:05:14.540 --> 00:05:17.759 avoid pedestrians, including systems that automatically 00:05:17.759 --> 00:05:20.079 apply the brakes, would dramatically improve 00:05:20.079 --> 00:05:23.610 safety. For example, Bill Grotz, a spokesman 00:05:23.610 --> 00:05:25.870 for General Motors, pointed to a recent study 00:05:25.870 --> 00:05:28.250 that found that GM vehicles with so -called front 00:05:28.250 --> 00:05:30.990 pedestrian braking reduced the frequency of injuries 00:05:30.990 --> 00:05:36.410 by 35 percent. Hey, Bill, General Motors is part 00:05:36.410 --> 00:05:38.709 of the Alliance for Automotive Innovation, which 00:05:38.709 --> 00:05:41.170 is suing it to stop the pedestrian automatic 00:05:41.170 --> 00:05:43.509 emergency braking rule from progressing right 00:05:43.509 --> 00:05:46.470 now. So I don't know what to tell you other than, 00:05:46.470 --> 00:05:50.889 you know. you can't you can't tout a technology 00:05:50.889 --> 00:05:53.230 while at the same time trying to tear it down 00:05:53.230 --> 00:05:57.089 from a regulatory standpoint um so you know again 00:05:57.089 --> 00:06:00.029 i guess the industry is backwards on this um 00:06:00.029 --> 00:06:02.629 but when you look at these crashes i think i 00:06:02.629 --> 00:06:04.250 guess one of the one of the big takes from the 00:06:04.250 --> 00:06:05.649 new york times article and something that everyone 00:06:05.649 --> 00:06:09.240 should get out of it is that When you see a pedestrian 00:06:09.240 --> 00:06:13.879 in a collision with a vehicle with a lower hood, 00:06:13.959 --> 00:06:18.740 a car, a sedan, or an older pickup or older SUV 00:06:18.740 --> 00:06:26.019 model, the human body essentially tends to roll 00:06:26.019 --> 00:06:30.579 over the hood and by doing so escapes the rest 00:06:30.579 --> 00:06:33.339 of the force that the vehicle is potentially 00:06:33.339 --> 00:06:37.500 transmitting to them. When you look at the higher 00:06:37.500 --> 00:06:39.720 hooded vehicles, what they're essentially doing 00:06:39.720 --> 00:06:41.800 is instead of the person rolling up onto the 00:06:41.800 --> 00:06:45.839 hood and out of the path of the massive amount 00:06:45.839 --> 00:06:49.379 of mass and momentum, I guess, that's there. 00:06:50.879 --> 00:06:52.899 What the big vehicles do is they essentially, 00:06:53.120 --> 00:06:54.740 and the article said it this way, I think it's 00:06:54.740 --> 00:06:58.740 a good way to phrase it. It punts the human in 00:06:58.740 --> 00:07:00.980 front of the vehicle, but the human remains in 00:07:00.980 --> 00:07:05.550 the path of the truck. can then be run over by 00:07:05.550 --> 00:07:07.829 the truck and subject to even more injuries and 00:07:07.829 --> 00:07:12.009 potential death. So that's really a lot of what 00:07:12.009 --> 00:07:14.829 we're looking at here beyond the visibility problem 00:07:14.829 --> 00:07:18.310 where manufacturers are blaming roof crush standards 00:07:18.310 --> 00:07:21.290 and the A -pillar, which I disagree with. I think 00:07:21.290 --> 00:07:23.689 that manufacturers could design smaller A -pillars. 00:07:23.730 --> 00:07:26.050 They're just doing it cheaply, which is why the 00:07:26.050 --> 00:07:30.230 A -pillars are so large. But ultimately, I think 00:07:30.230 --> 00:07:33.069 we really need to consider the height of hoods 00:07:33.069 --> 00:07:36.449 striking humans here as a big part of the problem. 00:07:36.949 --> 00:07:40.189 What I want to know is, so manufacturers will 00:07:40.189 --> 00:07:42.569 put in, you know, we've got rear view cameras 00:07:42.569 --> 00:07:45.189 and we've got things like that for pedestrian 00:07:45.189 --> 00:07:49.189 thing. Does this make people complacent? Because 00:07:49.189 --> 00:07:51.269 these systems, as we know, rear view cameras 00:07:51.269 --> 00:07:54.680 fail all the time. I imagine the pedestrian forward 00:07:54.680 --> 00:07:57.180 detection thing fails all the time. Does this 00:07:57.180 --> 00:08:00.759 start making drivers complacent to not turn around 00:08:00.759 --> 00:08:03.980 and actually look, and they don't use these tech 00:08:03.980 --> 00:08:08.399 add -ons as a supplement to what they should 00:08:08.399 --> 00:08:10.970 be doing as a driver? Are you trying to segue 00:08:10.970 --> 00:08:13.370 into the Tesla stories this week? No, no, no, 00:08:13.430 --> 00:08:17.230 no. That's the same process here, right? I mean, 00:08:17.230 --> 00:08:19.550 people become so comfortable and entrust their 00:08:19.550 --> 00:08:23.009 cars so much that when it does act up, they don't 00:08:23.009 --> 00:08:25.250 have the ability to take over and correct the 00:08:25.250 --> 00:08:27.449 problem, right? Right, because the example I'm 00:08:27.449 --> 00:08:30.509 thinking of is last week, I was traveling in 00:08:30.509 --> 00:08:33.429 a bunch of rental cars. And we're through some 00:08:33.429 --> 00:08:36.570 tight passages where you're backing up and I'm 00:08:36.570 --> 00:08:38.590 noticing the drivers, they're just relying on 00:08:38.590 --> 00:08:40.509 the rear view camera. And I was having a blast 00:08:40.509 --> 00:08:44.649 because rental car is not in my name. And they'd 00:08:44.649 --> 00:08:46.269 ask questions like, wait, is the yellow line 00:08:46.269 --> 00:08:47.889 or the red line? Should we be worried about? 00:08:48.690 --> 00:08:50.809 And I'm like, how about the stone wall you're 00:08:50.809 --> 00:08:54.830 aiming towards? Yeah, you're probably screwing 00:08:54.830 --> 00:08:58.490 up if you're just looking at the indicators that 00:08:58.490 --> 00:09:00.350 the technology puts on your screen and not at 00:09:00.350 --> 00:09:02.870 the actual objects. Yeah, I think that's something 00:09:02.870 --> 00:09:04.830 that all drivers have to realize that it doesn't 00:09:04.830 --> 00:09:07.149 matter what the tech is and how amazing and great 00:09:07.149 --> 00:09:10.450 it is. The best technology you have is the Mark 00:09:10.450 --> 00:09:15.190 I eyeball. Just look around. Mark one eyeball. 00:09:15.350 --> 00:09:17.210 Is that a brand name? I don't know. That's what 00:09:17.210 --> 00:09:19.610 a sailing instructor told me. He turned off all 00:09:19.610 --> 00:09:23.169 the electronics on the boat. All right. So you 00:09:23.169 --> 00:09:26.110 said it, Michael. You said the T word. It's time 00:09:26.110 --> 00:09:29.250 for us to jump into the world of Tesla. We have 00:09:29.250 --> 00:09:32.309 a couple of links to these stories. This is from 00:09:32.309 --> 00:09:34.570 the New York Times. And on the New York Times 00:09:34.570 --> 00:09:37.870 article, you can see the doorbell camera. What 00:09:37.870 --> 00:09:40.509 happens is a driver said they were using some 00:09:40.509 --> 00:09:42.960 sort of. They said they were an autopilot. They 00:09:42.960 --> 00:09:44.480 said they were using an autopilot, but autopilot 00:09:44.480 --> 00:09:48.019 doesn't exist anymore. It's renamed to full self 00:09:48.019 --> 00:09:52.759 -driving, supervised self -drive. Whatever the 00:09:52.759 --> 00:09:55.720 thing is, they were using Tesla nonsense. And 00:09:55.720 --> 00:09:58.940 the car accelerated to, I think they said like 00:09:58.940 --> 00:10:02.139 73 miles per hour residential area. And it went 00:10:02.139 --> 00:10:04.379 flying through a brick wall and unfortunately 00:10:04.379 --> 00:10:07.279 killed a woman inside. You can see the doorbell 00:10:07.279 --> 00:10:09.679 camera footage of this. And it's like nothing, 00:10:09.820 --> 00:10:12.179 nothing, nothing, nothing. Is this what? Nothing. 00:10:12.200 --> 00:10:17.159 And then, wow, a car come flying at it. The driver's 00:10:17.159 --> 00:10:21.360 not drunk, totally sober and saying, hey, what's 00:10:21.360 --> 00:10:24.679 going, you know, completely cooperative. Quoting 00:10:24.679 --> 00:10:26.980 from New York Times articles, the autopilot function, 00:10:27.220 --> 00:10:30.080 again, the exact name, don't know. Tesla's driver 00:10:30.080 --> 00:10:32.080 assistance technology is a popular feature among 00:10:32.080 --> 00:10:34.240 drivers, but it's had its issues over the years. 00:10:36.200 --> 00:10:39.980 That's playing a little mildly with it, that 00:10:39.980 --> 00:10:43.259 it's had its issues over the years. From an article 00:10:43.259 --> 00:10:47.100 in Electric, I'm going to quote, says, On Monday, 00:10:47.220 --> 00:10:50.120 CEO Elon Musk pushed back on the idea that the 00:10:50.120 --> 00:10:53.220 car was driving itself at all, posting that FSD 00:10:53.220 --> 00:10:55.320 drives slowly through neighborhood streets and 00:10:55.320 --> 00:10:58.019 this was a high -speed crash. Now Tesla's own 00:10:58.019 --> 00:10:59.820 account places its driver assistance software 00:10:59.820 --> 00:11:01.659 in the loop at the moment of the crash, with 00:11:01.659 --> 00:11:04.909 the driver's pedal input as the override. So 00:11:04.909 --> 00:11:08.370 Tesla's making the claim is that, hey, the driver 00:11:08.370 --> 00:11:12.450 overrode our software. Now, a question I have, 00:11:12.450 --> 00:11:16.909 and maybe this is my gaslight, sure, is that 00:11:16.909 --> 00:11:20.269 if Tesla or any company can create software that 00:11:20.269 --> 00:11:23.710 can do object detection in front of them, why 00:11:23.710 --> 00:11:27.669 would it allow the driver to override a clear 00:11:27.669 --> 00:11:31.009 inherent danger? Was this vehicle on Kool -Aid 00:11:31.009 --> 00:11:35.360 mode? Oh, yeah. I know that was lame, but you're 00:11:35.360 --> 00:11:37.759 all. Well, I mean, I guess there's a simple answer 00:11:37.759 --> 00:11:40.460 to that, right? If it's a supervised automation 00:11:40.460 --> 00:11:43.000 with the driver supervision, you've got to give 00:11:43.000 --> 00:11:45.100 the driver the ability to override vehicle commands, 00:11:45.279 --> 00:11:50.250 right? But in such a dangerous way. Well, that's 00:11:50.250 --> 00:11:52.230 what that's see, that's kind of the crux here. 00:11:52.269 --> 00:11:55.029 What happened there? Like, you know, Musk has 00:11:55.029 --> 00:11:57.610 come out and, you know, done the usual Twitter 00:11:57.610 --> 00:12:00.450 blame the human game. You know, yes, you know, 00:12:00.450 --> 00:12:02.269 admitted, at least in this case, which is not 00:12:02.269 --> 00:12:04.289 always the case that the vehicle is on full self 00:12:04.289 --> 00:12:06.450 -drive or autopilot, whatever the full self -driving 00:12:06.450 --> 00:12:10.110 system that's working here is. But, you know, 00:12:10.129 --> 00:12:13.970 they don't address at all what happens when these, 00:12:14.070 --> 00:12:18.139 you know, these situations occur. You know, they 00:12:18.139 --> 00:12:21.460 say there was pedal misapplication. There's still 00:12:21.460 --> 00:12:23.799 some questions around that and Tesla's data. 00:12:23.940 --> 00:12:26.120 Are the pedal misapplications they're seeing 00:12:26.120 --> 00:12:29.139 on their end real? Are they actually, you know, 00:12:29.179 --> 00:12:31.379 what's going on there? We never get to see the 00:12:31.379 --> 00:12:33.740 data. And so we're never really going to know 00:12:33.740 --> 00:12:39.039 that. And it's a very similar. excuse that they've 00:12:39.039 --> 00:12:41.820 given, you know, in, in, in the NHTSA investigation 00:12:41.820 --> 00:12:44.539 around, you know, sudden acceleration, you know, 00:12:44.600 --> 00:12:47.080 it's, you know, a hundred percent pedal misapplication. 00:12:47.080 --> 00:12:50.840 We've seen it in the data. Well, have we, you 00:12:50.840 --> 00:12:52.059 know, we're going to, we're going to be talking 00:12:52.059 --> 00:12:55.899 about this in the towel coming up. Can we, can 00:12:55.899 --> 00:13:00.840 we trust what's actually going on there? Should 00:13:00.840 --> 00:13:03.600 I leap into that now? Yeah, why don't you, well, 00:13:03.759 --> 00:13:06.919 give us a few minutes and we'll jump to your 00:13:06.919 --> 00:13:11.320 towel. But essentially, you know, it goes back 00:13:11.320 --> 00:13:15.320 to what we were talking about with the other 00:13:15.320 --> 00:13:18.240 vehicles, right? Are they kind of breeding or 00:13:18.240 --> 00:13:24.580 creating the situation that the human, that requires 00:13:24.580 --> 00:13:28.919 a human to essentially possess? super human capabilities 00:13:28.919 --> 00:13:31.539 to overcome the safety problem that the vehicle 00:13:31.539 --> 00:13:35.559 has created? And are they creating situations 00:13:35.559 --> 00:13:38.759 that, you know, the human can't recover the vehicle 00:13:38.759 --> 00:13:41.539 from? In this case, you know, it's a neighborhood 00:13:41.539 --> 00:13:44.519 with low speeds and somehow, you know, we know 00:13:44.519 --> 00:13:46.700 it tells us it can accelerate incredibly fast, 00:13:46.879 --> 00:13:48.399 right? That's a big part of their marketing. 00:13:48.600 --> 00:13:52.519 So, you know, a vehicle leaping from, you know, 00:13:52.539 --> 00:13:57.629 35 miles per hour to 70 miles per hour. takes, 00:13:57.750 --> 00:14:00.889 you know, a split, you know, half a second maybe. 00:14:00.990 --> 00:14:03.389 So we're not talking about it, you know, yeah, 00:14:03.470 --> 00:14:05.429 we're not really talking about, you know, the 00:14:05.429 --> 00:14:08.129 driver holding their foot down the pedal for 00:14:08.129 --> 00:14:10.549 30 seconds and flying to the neighborhood here. 00:14:10.649 --> 00:14:12.149 I mean, this certainly is something that could 00:14:12.149 --> 00:14:14.789 have occurred in, you know, a matter of a second 00:14:14.789 --> 00:14:20.610 or two, right? So that is well within the timing 00:14:20.610 --> 00:14:23.850 that we think is problematic for humans to recover 00:14:23.850 --> 00:14:28.029 or to, you know. take back attention and take 00:14:28.029 --> 00:14:31.950 back control of the vehicle. How long do you 00:14:31.950 --> 00:14:34.429 think it'll be before Tesla uses this as a testament 00:14:34.429 --> 00:14:38.370 to the structural integrity of their car? I mean, 00:14:38.470 --> 00:14:42.169 you know, pretty soon I'm sure we'll see advertising 00:14:42.169 --> 00:14:46.230 that says Tesla's such a strong car that it protected 00:14:46.230 --> 00:14:48.889 the driver when it penetrated the wall of a house. 00:14:49.490 --> 00:14:51.490 Yeah, that woman should have been sitting in 00:14:51.490 --> 00:14:53.809 her living room inside of a Tesla. I get it. 00:14:54.590 --> 00:14:56.789 Absolutely. Yeah, it was her fault for not having 00:14:56.789 --> 00:14:59.269 bought a Tesla that she could bring into her 00:14:59.269 --> 00:15:01.909 living room. Well, fortunately for them, they've 00:15:01.909 --> 00:15:03.750 got a, you know, I think another Tesla went off 00:15:03.750 --> 00:15:07.389 a cliff recently and people survived. So they've 00:15:07.389 --> 00:15:10.850 got at least two, you know, 200 foot, 300 foot 00:15:10.850 --> 00:15:13.799 cliff drops in their favor. Over a house that, 00:15:13.820 --> 00:15:16.179 you know, that kind of brings I don't know. I 00:15:16.179 --> 00:15:17.820 don't think the situation is one they want to 00:15:17.820 --> 00:15:21.100 highlight. No, but so related to my initial question 00:15:21.100 --> 00:15:23.580 is like, OK, why would you allow someone to override 00:15:23.580 --> 00:15:25.919 so dramatically? And also, since you're in a 00:15:25.919 --> 00:15:28.799 residential neighborhood again, why would you 00:15:28.799 --> 00:15:32.879 allow the driver to override to such a speed? 00:15:32.980 --> 00:15:35.500 It seems to me that, hey, if you guys are the 00:15:35.500 --> 00:15:38.480 most software, the best software engineers and 00:15:38.480 --> 00:15:40.559 hardware engineers in all time. wouldn't you 00:15:40.559 --> 00:15:43.159 at least be like, hey, driver, are you sure you 00:15:43.159 --> 00:15:46.419 want to go this fast in this neighborhood? I 00:15:46.419 --> 00:15:48.820 think there's that gap missing there. I understand, 00:15:48.919 --> 00:15:50.500 Michael, your point of saying, yes, at times, 00:15:50.500 --> 00:15:52.019 if it's supervised by humans, you want to be 00:15:52.019 --> 00:15:54.620 able to override this. But there should be warnings 00:15:54.620 --> 00:15:56.860 presented to users because there's no scenario 00:15:56.860 --> 00:15:58.480 where you're in a residential street where you're 00:15:58.480 --> 00:16:00.299 doing 30 and you're like, I'm going to do 75 00:16:00.299 --> 00:16:02.179 all of a sudden. Yeah, you know what? It should 00:16:02.179 --> 00:16:04.779 even be, I don't know, every time that kind of 00:16:04.779 --> 00:16:08.350 subject comes up, if the vehicle... is capable 00:16:08.350 --> 00:16:11.570 of seeing that the driver is exceeding the speed 00:16:11.570 --> 00:16:14.309 limit in a neighborhood, why is the vehicle slowing 00:16:14.309 --> 00:16:17.070 itself down in the first place? That's my problem. 00:16:17.070 --> 00:16:19.590 We've talked about this before. Why are people 00:16:19.590 --> 00:16:22.049 using a list technology to warn not only the 00:16:22.049 --> 00:16:26.370 driver, but also the people in the neighborhood 00:16:26.370 --> 00:16:28.830 of the car that something dangerous is happening? 00:16:29.169 --> 00:16:31.409 You know, if somebody's speeding this badly, 00:16:31.629 --> 00:16:33.850 there should be lights that's going off and lights 00:16:33.850 --> 00:16:38.799 flashing and klaxons klaxing. I mean, in this 00:16:38.799 --> 00:16:40.799 case, you know, granted that this may have only 00:16:40.799 --> 00:16:44.179 taken a second or two. So warnings are not going 00:16:44.179 --> 00:16:47.120 to help maybe as much as, you know, as you might 00:16:47.120 --> 00:16:50.320 think. But I don't know. It's definitely something 00:16:50.320 --> 00:16:52.120 that I think we should consider with vehicles 00:16:52.120 --> 00:16:56.480 that are being robots being supervised by humans. 00:16:57.220 --> 00:17:01.019 It makes me wonder if we if we need to put put 00:17:01.019 --> 00:17:04.079 in some type of warning signs or at least. You 00:17:04.079 --> 00:17:05.660 know, the manufacturer of these vehicles, that's 00:17:05.660 --> 00:17:07.640 the last thing they want because they think that's 00:17:07.640 --> 00:17:10.740 a beacon saying this car isn't safe. Right. It's 00:17:10.740 --> 00:17:13.000 like, hey, why? I don't have to put in seatbelts. 00:17:13.059 --> 00:17:17.420 My car is safe. So but out of this. So Senators 00:17:17.420 --> 00:17:20.680 Markey and Blumenthal, they wrote a letter to 00:17:20.680 --> 00:17:22.720 NHTSA saying, hey, you got to give us some answers. 00:17:23.180 --> 00:17:26.900 What's going on with Tesla's FSD? And they told 00:17:26.900 --> 00:17:30.400 NHTSA you have July 7th to respond. And who knows 00:17:30.400 --> 00:17:35.039 what will happen there? Their it's essentially 00:17:35.039 --> 00:17:37.640 was their response to that Reuters article. I 00:17:37.640 --> 00:17:39.160 think we talked about a couple of weeks ago that 00:17:39.160 --> 00:17:41.819 was taking a look at Tesla's bullshit data that 00:17:41.819 --> 00:17:44.140 they're submitting to the European authorities 00:17:44.140 --> 00:17:46.240 and that they're advertising on their website 00:17:46.240 --> 00:17:49.579 and that no one else on Earth has been able to 00:17:49.579 --> 00:17:54.359 independently review and verify. So Tesla's data 00:17:54.359 --> 00:17:58.380 is supremely problematic. And with that, Mr. 00:17:58.539 --> 00:18:02.380 Perkins, you were eager. And excited to jump 00:18:02.380 --> 00:18:04.240 into your towel. I hear it's been peer reviewed 00:18:04.240 --> 00:18:07.160 this week. It's been peer reviewed. Yes, this 00:18:07.160 --> 00:18:11.240 is a peer reviewed study, which, as we all know 00:18:11.240 --> 00:18:14.960 from Waymo's experiences, really important to 00:18:14.960 --> 00:18:18.720 get people to publicly accept or to get the public 00:18:18.720 --> 00:18:21.339 to accept the results. So maybe we'll get into 00:18:21.339 --> 00:18:23.819 the New York Times, too. Who knows? But anyway, 00:18:24.839 --> 00:18:28.000 you know, we hear a lot about edge cases. So 00:18:28.000 --> 00:18:30.680 I started to wonder whether an edge case is a 00:18:30.680 --> 00:18:35.759 bug or a feature of the AI -controlled car. So 00:18:35.759 --> 00:18:42.599 ran a few numbers on this. And so a couple of 00:18:42.599 --> 00:18:47.220 assumptions that basically there's 100 safety 00:18:47.220 --> 00:18:51.140 -critical features that you can identify in the 00:18:51.140 --> 00:18:55.700 car that are potentially failure modes, but only 00:18:55.700 --> 00:18:58.410 100. In fact, the car's got many more. There 00:18:58.410 --> 00:19:03.150 might be several hundred electronic control units. 00:19:03.289 --> 00:19:06.269 There's sensors. Sensors can fail in different 00:19:06.269 --> 00:19:10.509 ways. But to make the numbers easy, I just selected, 00:19:10.670 --> 00:19:13.609 you know, assumed that there were a hundred different 00:19:13.609 --> 00:19:16.329 things that are safety critical that can fail. 00:19:16.529 --> 00:19:19.210 Can you give me an example of what one of these 00:19:19.210 --> 00:19:26.259 would be? Oh, sure. A camera. That would be used 00:19:26.259 --> 00:19:29.519 in the cars. So, you know, Tesla says they've 00:19:29.519 --> 00:19:33.299 got 23 cameras. Excuse me, Waymo says they have 00:19:33.299 --> 00:19:37.339 23 cameras. Or excuse me, 13 cameras. They have 00:19:37.339 --> 00:19:41.920 four lidars and six radars. And they've got at 00:19:41.920 --> 00:19:46.299 least one CPU. Each of the cameras has got at 00:19:46.299 --> 00:19:49.480 least one perception module associated with it. 00:19:49.539 --> 00:19:52.849 So you get the raw data. Then the data flows 00:19:52.849 --> 00:19:56.329 through a perception module to say, okay, what 00:19:56.329 --> 00:19:59.410 is this object that I'm looking at? Is it a cow 00:19:59.410 --> 00:20:03.910 or is it a house? And then there's the main computer. 00:20:04.029 --> 00:20:07.230 You've got engine control units. You've got units 00:20:07.230 --> 00:20:10.369 that are electronically controlling the steering 00:20:10.369 --> 00:20:12.829 and the brakes. So all of these things, if they 00:20:12.829 --> 00:20:17.759 fail, they have safety consequences. Right. And 00:20:17.759 --> 00:20:20.099 so that's what we're looking at. So there's hundreds 00:20:20.099 --> 00:20:25.660 in any vehicle. But again, to just make the arithmetic 00:20:25.660 --> 00:20:28.140 a little easier, I said, let's assume there's 00:20:28.140 --> 00:20:32.619 100. And let's further assume that the car is 00:20:32.619 --> 00:20:37.400 designed to be two -fault tolerant. So in other 00:20:37.400 --> 00:20:40.460 words, if you've got one failure, the car will 00:20:40.460 --> 00:20:43.759 be safe. It will safely come to a halt or something. 00:20:46.480 --> 00:20:50.440 it could potentially cause injury to the occupants 00:20:50.440 --> 00:20:52.960 or somebody around it. So two -fault tolerant 00:20:52.960 --> 00:20:56.960 would mean you can fail two and you'll still 00:20:56.960 --> 00:21:00.160 be safe or you'll still be able to preserve the 00:21:00.160 --> 00:21:03.799 safety of the occupant. If you do that, what 00:21:03.799 --> 00:21:09.319 you find is that there are at least 5 ,000 combinations, 00:21:09.839 --> 00:21:12.920 100 features, two failures. There's at least 00:21:12.920 --> 00:21:17.779 5 ,000 combinations. That can result in injury 00:21:17.779 --> 00:21:22.480 to the occupants or to the people, vulnerable 00:21:22.480 --> 00:21:28.180 users around the vehicle. Now, in order to assure 00:21:28.180 --> 00:21:31.039 that your vehicle is going to be safe, you'd 00:21:31.039 --> 00:21:34.900 have to validate all of those combinations. It's 00:21:34.900 --> 00:21:39.099 not within anybody's range of dollars or time 00:21:39.099 --> 00:21:45.009 who's developing a vehicle to do that. Just there 00:21:45.009 --> 00:21:47.109 isn't enough time and money in the world to do 00:21:47.109 --> 00:21:50.170 that. And that's, again, assuming that you've 00:21:50.170 --> 00:21:52.690 only got 100 features. If you've got many more 00:21:52.690 --> 00:21:54.589 features, of course, you've got many more combinations 00:21:54.589 --> 00:21:59.109 and a lot more to do. So in practice, what the 00:21:59.109 --> 00:22:03.569 developers do is they look at, as best they understand, 00:22:03.670 --> 00:22:05.849 the risk profile associated with each of the 00:22:05.849 --> 00:22:08.069 potential failures and say, well, we're going 00:22:08.069 --> 00:22:12.650 to look at a couple of the really high risk features. 00:22:13.289 --> 00:22:16.049 High probability and high consequence, for example, 00:22:16.170 --> 00:22:20.230 or very high probability and medium consequence. 00:22:20.589 --> 00:22:23.109 So there's a way of managing the risk. And what 00:22:23.109 --> 00:22:27.089 they do is they pick a few. But notably, they 00:22:27.089 --> 00:22:31.150 do not pick all of them. So if you look at a 00:22:31.150 --> 00:22:36.650 couple of the leading risk factors, like is this 00:22:36.650 --> 00:22:39.789 car possibly going to run into a house automatically? 00:22:40.890 --> 00:22:44.369 That's something you might want to address. That 00:22:44.369 --> 00:22:49.789 leaves 5 ,048 potential combinations that are 00:22:49.789 --> 00:22:53.190 risky that you have not looked at. And each of 00:22:53.190 --> 00:22:55.230 these has got a certain probability of happening. 00:22:55.869 --> 00:22:59.869 But if you combine the certain probability of 00:22:59.869 --> 00:23:02.690 it happening with the huge number of potential 00:23:02.690 --> 00:23:06.450 consequences, you end up saying, you know, this 00:23:06.450 --> 00:23:09.769 is inevitable that something's going to happen. 00:23:10.650 --> 00:23:15.069 risk the lives of the occupants of the car. But 00:23:15.069 --> 00:23:18.450 another factor is that if you're using AI to 00:23:18.450 --> 00:23:23.410 run your car, you have no idea even how to develop 00:23:23.410 --> 00:23:27.910 a really reliable way of figuring out whether 00:23:27.910 --> 00:23:29.769 or not this is going to happen, what the risk 00:23:29.769 --> 00:23:33.069 profile is, because we know that AI will inject 00:23:33.069 --> 00:23:37.569 false results into any analysis that it's doing. 00:23:38.240 --> 00:23:40.240 And you've got AI sprinkled throughout the car. 00:23:40.380 --> 00:23:43.220 You've got it in the perception module, right? 00:23:43.279 --> 00:23:46.380 How is it going to figure out what is in front 00:23:46.380 --> 00:23:48.980 of the car based upon the inputs from the cameras? 00:23:49.619 --> 00:23:52.380 You've got it involved in the driving itself. 00:23:53.640 --> 00:23:56.859 How do you interpret the speed of the car, the 00:23:56.859 --> 00:23:59.720 location of the car, the acceleration of the 00:23:59.720 --> 00:24:02.319 car? How do you know that what you've been training 00:24:02.319 --> 00:24:06.960 the car on is actually going to provide a reasonable 00:24:06.960 --> 00:24:10.259 result? out of the computer, because we know 00:24:10.259 --> 00:24:14.559 that the AI just sprinkles false results into 00:24:14.559 --> 00:24:18.720 whatever it's doing. So I've come to the conclusion, 00:24:19.019 --> 00:24:22.099 and I'm sure a lot of people may not come to 00:24:22.099 --> 00:24:25.039 the same conclusion, that what people call edge 00:24:25.039 --> 00:24:29.259 cases are in fact intrinsic to the architecture 00:24:29.259 --> 00:24:32.400 you've built the car around. So they're not bugs, 00:24:32.519 --> 00:24:35.920 they're features. They're going to happen. You 00:24:35.920 --> 00:24:38.269 know they're going to happen. And they will not 00:24:38.269 --> 00:24:41.490 happen repeatedly. They'll not happen reliably. 00:24:42.089 --> 00:24:44.569 But they're sure going to happen. This is just 00:24:44.569 --> 00:24:49.509 a feature of the AI architecture that's driving 00:24:49.509 --> 00:24:53.569 the vehicle. And it's inevitable. And so I think 00:24:53.569 --> 00:24:56.490 that when people talk about edge cases, and I 00:24:56.490 --> 00:24:57.869 know a lot of people are going to beat me up 00:24:57.869 --> 00:24:59.869 over this, but they're really just providing 00:24:59.869 --> 00:25:03.069 cover for the industry to say, well, no, we never 00:25:03.069 --> 00:25:07.690 bothered to. properly evaluate and properly validate 00:25:07.690 --> 00:25:09.630 all of the things that can happen in the car. 00:25:10.309 --> 00:25:14.430 Because this is just an edge case. It's not something 00:25:14.430 --> 00:25:17.750 that's plausibly ever going to happen. My argument 00:25:17.750 --> 00:25:21.769 is that it has to happen because you've built 00:25:21.769 --> 00:25:24.269 an architecture that makes it inevitable that 00:25:24.269 --> 00:25:26.390 something really bad is going to happen. And 00:25:26.390 --> 00:25:28.690 you've never put the proper amount of time and 00:25:28.690 --> 00:25:32.660 energy into validating all of the... safety results 00:25:32.660 --> 00:25:37.180 as you need to to make sure that well for example 00:25:37.180 --> 00:25:39.180 that the car is not going to run into a house 00:25:39.180 --> 00:25:43.140 in a residential neighborhood so that's that 00:25:43.140 --> 00:25:45.779 makes sense is that clear i think i'm wandering 00:25:45.779 --> 00:25:48.160 a little bit yeah but you started wandering into 00:25:48.160 --> 00:25:51.980 english um so i i get you so fred you've spent 00:25:51.980 --> 00:25:54.559 you know a quite a bit of time doing software 00:25:54.559 --> 00:25:57.759 development hardware development what's the worst 00:25:57.759 --> 00:26:00.279 part of both of those things that no one wants 00:26:00.279 --> 00:26:03.069 to do Well, the worst part is the uncertainty. 00:26:04.349 --> 00:26:09.710 Because if you look at AI, right, it's not like 00:26:09.710 --> 00:26:12.930 an input is going to give you a certain output. 00:26:13.390 --> 00:26:17.210 You have an input that you're giving it, an image, 00:26:17.329 --> 00:26:19.849 for example, of a human being dressed in a cow 00:26:19.849 --> 00:26:25.269 costume. And somehow the AI is going to interpret 00:26:25.269 --> 00:26:27.890 that one way or another, and the odds that it's 00:26:27.890 --> 00:26:31.029 going to interpret it wrong. are reasonably high. 00:26:31.410 --> 00:26:34.569 But you as a human being have no way of figuring 00:26:34.569 --> 00:26:38.630 out what the AI is going to do. So the only way 00:26:38.630 --> 00:26:42.670 you can responsibly validate its results is to 00:26:42.670 --> 00:26:46.569 run a huge number of tests to see where the boundaries 00:26:46.569 --> 00:26:50.049 are, to see if this has a one in a million chance 00:26:50.049 --> 00:26:52.910 of happening, or if it has a one in 10 chance 00:26:52.910 --> 00:26:57.170 of misinterpreting the results. That's very expensive 00:26:57.170 --> 00:27:00.769 to do, right? And in fact, what the AI companies 00:27:00.769 --> 00:27:04.690 are doing is ignoring all that. They have never 00:27:04.690 --> 00:27:11.109 set up test apparatus that allow them to go and 00:27:11.109 --> 00:27:14.769 look at the things that have occurred that are 00:27:14.769 --> 00:27:18.589 either lethal or very hazardous from the AI control 00:27:18.589 --> 00:27:21.069 system to say, okay, the minimum we should do 00:27:21.069 --> 00:27:25.200 is have a test system that looks at Events that 00:27:25.200 --> 00:27:26.920 have already happened that have either killed 00:27:26.920 --> 00:27:29.279 people or threatened the lives of people and 00:27:29.279 --> 00:27:32.920 put our AI through the paces to make sure, as 00:27:32.920 --> 00:27:37.180 a minimum, that it will have a very low probability 00:27:37.180 --> 00:27:42.279 of repeating that lethal consequence because 00:27:42.279 --> 00:27:45.720 I've now fixed the software. What is, I mean, 00:27:45.740 --> 00:27:48.440 it's just from that point, what, what is it? 00:27:49.259 --> 00:27:51.200 Do you have any guarantee, even if you perform 00:27:51.200 --> 00:27:53.119 those tests, that the AI isn't going to change 00:27:53.119 --> 00:27:54.940 its mind and start behaving differently the next 00:27:54.940 --> 00:27:58.299 day? No. No, I have no guarantee. So what's the 00:27:58.299 --> 00:28:04.200 point? Is AI just not ever going to be trustworthy 00:28:04.200 --> 00:28:07.160 in this area? Well, it depends on what you mean 00:28:07.160 --> 00:28:09.500 by trustworthy. If you say trustworthy means 00:28:09.500 --> 00:28:12.480 there's a 1 ,000 chance it's going to come up 00:28:12.480 --> 00:28:15.039 with the wrong result, you've got to run several 00:28:15.039 --> 00:28:17.819 thousand tests to verify that, yeah, it's probably 00:28:17.819 --> 00:28:21.980 not going to happen. When an AI company releases 00:28:21.980 --> 00:28:25.240 a software update two days after a lethal event 00:28:25.240 --> 00:28:28.380 and says, well, now it's fixed, it is pure bullshit. 00:28:28.440 --> 00:28:30.920 There's no way in the world you can validate 00:28:30.920 --> 00:28:33.880 that with real hardware and a real system in 00:28:33.880 --> 00:28:37.000 two days. That's just not going to happen. So 00:28:37.000 --> 00:28:44.769 really, this is all based upon people's... really 00:28:44.769 --> 00:28:47.049 anthropomorphizing, is that a word? The computer 00:28:47.049 --> 00:28:50.089 itself to say, well, it's a human being and we 00:28:50.089 --> 00:28:53.369 slapped it and now it's going to behave properly. 00:28:53.789 --> 00:28:56.849 But when you've got AI in the mix, it's never 00:28:56.849 --> 00:28:58.950 going to be deterministic and you as a human 00:28:58.950 --> 00:29:02.829 being will never be able to understand what the 00:29:02.829 --> 00:29:06.609 computer is going to do to interpret this new 00:29:06.609 --> 00:29:09.289 input or this new rule that you put in place. 00:29:09.990 --> 00:29:12.119 This ain't going to happen. There's no time. 00:29:13.000 --> 00:29:16.039 The point of doing this, Michael, is to try and 00:29:16.039 --> 00:29:20.079 get that error rate down as small as possible. 00:29:20.160 --> 00:29:23.880 But even that's not reliable if the AI is changing. 00:29:24.259 --> 00:29:28.099 Well, OK, so that's how you do a testing. Yeah, 00:29:28.200 --> 00:29:30.900 there's a bunch of weird things here. Yeah, those 00:29:30.900 --> 00:29:34.299 are like the way we understand testing completely 00:29:34.299 --> 00:29:38.150 change for AI. I think, yes, right? Well, that's 00:29:38.150 --> 00:29:39.950 the problem. The way we understand testing has 00:29:39.950 --> 00:29:42.789 not changed, but the way we should have interpreted 00:29:42.789 --> 00:29:48.849 it has changed significantly. Okay, so there 00:29:48.849 --> 00:29:50.569 really should be a completely different approach. 00:29:51.849 --> 00:29:54.369 Now, the companies understand this. That's why 00:29:54.369 --> 00:29:56.769 they say, well, we need to drive a jillion miles 00:29:56.769 --> 00:29:59.710 to validate this. This is why we need to threaten 00:29:59.710 --> 00:30:03.009 the public with untested vehicles to make sure 00:30:03.009 --> 00:30:06.230 that, you know, it doesn't kill too many people. 00:30:06.889 --> 00:30:09.750 That's because that's the best they can do. Or 00:30:09.750 --> 00:30:11.769 they could say, well, we're going to put it on 00:30:11.769 --> 00:30:14.349 a test track. We know certain things are very 00:30:14.349 --> 00:30:17.849 hazardous and very difficult to live through. 00:30:17.990 --> 00:30:20.309 So that's what we're going to put in place on 00:30:20.309 --> 00:30:23.569 the test track. Not the easy stuff, the hard 00:30:23.569 --> 00:30:28.230 stuff. And nobody's doing that. Well, I mean, 00:30:28.250 --> 00:30:29.930 they make the absurd claim that we don't know 00:30:29.930 --> 00:30:31.569 what the hard stuff is. These are edge cases. 00:30:31.730 --> 00:30:33.920 And then they claim like, you know. Flooded streets 00:30:33.920 --> 00:30:36.460 are an edge case. It's not. Or school buses with 00:30:36.460 --> 00:30:38.480 their stop arm extended. It's an edge case. It's 00:30:38.480 --> 00:30:41.720 not. But breaking away from this, so it's separate 00:30:41.720 --> 00:30:44.420 from AI. Just programming in itself is never 00:30:44.420 --> 00:30:47.680 perfect. It never will be. Or if it's, you know, 00:30:47.720 --> 00:30:49.759 you're doing object -oriented or functional or 00:30:49.759 --> 00:30:51.759 AI, whatever, you'll never get to perfection. 00:30:52.160 --> 00:30:55.660 I don't care what you claim. So the example I 00:30:55.660 --> 00:30:58.740 think of, again, is like Airbus and Boeing. So 00:30:58.740 --> 00:31:01.579 they have these mission -critical systems that 00:31:01.579 --> 00:31:05.720 can fail. But they have and there's no way they 00:31:05.720 --> 00:31:08.259 will ever claim that everything in their systems 00:31:08.259 --> 00:31:10.740 are perfect. There are complex systems. And every 00:31:10.740 --> 00:31:12.579 time there's a crash, they use that as a learning 00:31:12.579 --> 00:31:15.019 experience to get things better. I remember there 00:31:15.019 --> 00:31:19.579 was an Airbus incident in in in Russia where 00:31:19.579 --> 00:31:22.779 one of the pilots let their kids sit it there. 00:31:22.900 --> 00:31:24.420 And like, you know, it was an autopilot and he 00:31:24.420 --> 00:31:25.920 just, you know, play with the wheel and nothing 00:31:25.920 --> 00:31:28.099 happened. But they didn't know that he threw 00:31:28.099 --> 00:31:30.940 the wheel all the way to port. And that set some 00:31:30.940 --> 00:31:34.640 flag inside Airbus' systems, like autopilot is 00:31:34.640 --> 00:31:37.720 now disabled. And the pilots had no clue what 00:31:37.720 --> 00:31:39.319 was going on. They didn't know how to recover 00:31:39.319 --> 00:31:41.720 from this. And that was not programmed that way. 00:31:41.779 --> 00:31:44.500 That was not some intentional thing. It was just 00:31:44.500 --> 00:31:47.420 a weird glitch. Or maybe it wasn't. It was never 00:31:47.420 --> 00:31:49.940 explained to people. But either way, these are... 00:31:50.410 --> 00:31:54.910 That's an edge case. You let a teenager sit behind 00:31:54.910 --> 00:32:00.349 a $1 million jet. You let a non -trained pilot 00:32:00.349 --> 00:32:05.069 in the cockpit behind the joystick. It's the 00:32:05.069 --> 00:32:09.269 yoke. That's no different than putting your car 00:32:09.269 --> 00:32:12.390 out, an autonomous vehicle out in a flash flood, 00:32:12.589 --> 00:32:15.529 right? See, that's the difference. So the airline 00:32:15.529 --> 00:32:18.410 industry. recognize this shit's going to happen 00:32:18.410 --> 00:32:21.809 so they've got a process through which when this 00:32:21.809 --> 00:32:25.190 shit happens everybody gets together and discusses 00:32:25.190 --> 00:32:28.809 what happened and they put you know and they 00:32:28.809 --> 00:32:31.529 get to the root cause and then they do their 00:32:31.529 --> 00:32:34.069 best to address the root cause and and they're 00:32:34.069 --> 00:32:37.089 pretty damn good at doing that okay the automobile 00:32:37.089 --> 00:32:41.089 industry doesn't do that they assume that every 00:32:41.089 --> 00:32:44.440 safety architecture is proprietary They're not 00:32:44.440 --> 00:32:46.319 going to share it with anybody, much less the 00:32:46.319 --> 00:32:48.180 public. They're certainly not going to share 00:32:48.180 --> 00:32:51.740 it with their competitors. So this is a structural 00:32:51.740 --> 00:32:55.819 failure of the way the automobile industry is 00:32:55.819 --> 00:32:59.619 set up and the way that they handle information 00:32:59.619 --> 00:33:05.519 that is never going to result in rapid identification 00:33:05.519 --> 00:33:10.180 and correction of the failures they've got. And 00:33:10.180 --> 00:33:12.460 they're stuck with this AI and they're stuck 00:33:12.460 --> 00:33:15.680 with their... PR that says it's perfect. It's 00:33:15.680 --> 00:33:17.640 better than human. It never sleeps. It never 00:33:17.640 --> 00:33:21.059 shits. It never does any of this stuff that distracts 00:33:21.059 --> 00:33:23.440 you from the driving. I also think there's a 00:33:23.440 --> 00:33:26.440 cultural shift, too, in terms of software engineers. 00:33:26.680 --> 00:33:28.880 So the engineers that I've spoken to that are 00:33:28.880 --> 00:33:30.460 10 years younger than me that work for these 00:33:30.460 --> 00:33:33.339 large multinational corporations, I've asked 00:33:33.339 --> 00:33:35.220 them basic questions like, oh, who does your 00:33:35.220 --> 00:33:38.619 testing process? Some of them at very large corporations 00:33:38.619 --> 00:33:42.720 that we all use are like, no one. A friend of 00:33:42.720 --> 00:33:44.940 mine at a large financial institution that we've 00:33:44.940 --> 00:33:48.359 all used, she said she gets to test her own code. 00:33:48.619 --> 00:33:52.960 I'm like, no, no, no, no. That's not a no. No, 00:33:53.059 --> 00:33:56.119 you can't. No, that's that's. Wait, is this peer 00:33:56.119 --> 00:34:00.200 review? You're your own peer. And so I I mean, 00:34:00.240 --> 00:34:02.319 I remember when I was, you know, coming up in 00:34:02.319 --> 00:34:04.579 this in the late 90s, that was not a situation 00:34:04.579 --> 00:34:07.690 like. You're absolutely like, and I was specialized 00:34:07.690 --> 00:34:09.429 in a specific thing where people didn't have 00:34:09.429 --> 00:34:11.670 these skills. I couldn't review my own stuff. 00:34:12.010 --> 00:34:14.489 Like still other people had to review it and 00:34:14.489 --> 00:34:16.550 say, hey, is this input and output correct? Is 00:34:16.550 --> 00:34:18.730 this what we expect? And if it wasn't, they'd 00:34:18.730 --> 00:34:21.510 tell me to fix it. Like you can't test your own 00:34:21.510 --> 00:34:23.670 thing. And I think there's been a winnowing away 00:34:23.670 --> 00:34:27.489 of that in the software. From what is anecdotal, 00:34:27.510 --> 00:34:29.710 but in the software engineering field is, oh, 00:34:29.789 --> 00:34:31.750 we have these automated tools that do some sort 00:34:31.750 --> 00:34:33.590 of testing. We have this automated release pipeline 00:34:33.590 --> 00:34:38.599 that does things. But there's no QA teams anymore, 00:34:38.760 --> 00:34:41.219 as best I can tell. You don't need QA because 00:34:41.219 --> 00:34:44.239 you've got AI. Yeah, but even prior to AI tools 00:34:44.239 --> 00:34:46.599 doing this stuff, they just kind of gave up on 00:34:46.599 --> 00:34:49.340 that because that's expensive and the managers, 00:34:49.500 --> 00:34:52.559 they don't understand this shit. Well, I mean, 00:34:52.599 --> 00:34:55.739 I guess in the financial arena, you'd stand to 00:34:55.739 --> 00:34:57.760 lose a lot of money if you screw some of that 00:34:57.760 --> 00:35:00.360 up, right? Oh, yeah. I mean, and then but if 00:35:00.360 --> 00:35:02.340 you're so that doesn't make sense that they wouldn't 00:35:02.340 --> 00:35:04.360 want to test it there. But even more so when 00:35:04.360 --> 00:35:06.719 you've got a life on the line, a car. Yeah. So 00:35:06.719 --> 00:35:09.039 I think that's part of the issue. And I would 00:35:09.039 --> 00:35:11.960 love to hear from anybody at these, you know, 00:35:11.960 --> 00:35:15.639 at the GM or Ford or Waymo and whatnot. We're 00:35:15.639 --> 00:35:17.079 not going to talk about your name or anything 00:35:17.079 --> 00:35:19.119 like that. I just want some engineers there and 00:35:19.119 --> 00:35:22.380 say, no, no, no, no. We have this is our detailed 00:35:22.380 --> 00:35:24.960 process. We're not going to reveal any of your 00:35:24.960 --> 00:35:29.539 your secret sauce. But I make me. And our listeners 00:35:29.539 --> 00:35:31.920 feel slightly more comfortable that it's not 00:35:31.920 --> 00:35:34.059 just, oh, they're just automated testing. It's, 00:35:34.059 --> 00:35:36.760 you know, we just release things and, you know, 00:35:36.780 --> 00:35:39.420 there's a test case. Because that's my impression 00:35:39.420 --> 00:35:43.059 from talking to people that that's been going 00:35:43.059 --> 00:35:45.380 on for at least the last 10, 15 years, which 00:35:45.380 --> 00:35:49.460 is. Well, to be fair to Waymo, they have published 00:35:49.460 --> 00:35:51.519 a lot of information about how they do their 00:35:51.519 --> 00:35:55.099 safety releases. Okay. And you can go to their 00:35:55.099 --> 00:35:57.239 website and it talks all about the structures 00:35:57.239 --> 00:36:00.289 that they've got. And they've got review teams 00:36:00.289 --> 00:36:05.070 and they have three levels or whatever of people 00:36:05.070 --> 00:36:08.329 reviewing it. What they do not have is independence. 00:36:08.949 --> 00:36:13.570 And what they do not have is some way of a person 00:36:13.570 --> 00:36:16.670 to stand up and say, this is wrong. This is just 00:36:16.670 --> 00:36:19.889 not the right way to go. What instead they have 00:36:19.889 --> 00:36:23.849 are all Waymo employees reviewing the Waymo work 00:36:23.849 --> 00:36:26.469 and objectives, which is a lot like what you 00:36:26.469 --> 00:36:29.250 said. I would like to point out that they do 00:36:29.250 --> 00:36:31.889 have a way. It's the NHTSA whistleblower regulations 00:36:31.889 --> 00:36:34.750 that, you know, if they come forward and provide 00:36:34.750 --> 00:36:37.949 that information, they can actually get a percentage 00:36:37.949 --> 00:36:40.550 of any penalties that are that are, you know, 00:36:40.550 --> 00:36:43.309 assessed against Waymo for violations of the 00:36:43.309 --> 00:36:45.510 Safety Act. So there is one one thing they can 00:36:45.510 --> 00:36:47.730 do there. Right. Well, how would that be implemented? 00:36:48.210 --> 00:36:50.289 Well, I mean, what would a person do? Who do 00:36:50.289 --> 00:36:56.039 they call? They call NHTSA's chief counsel. And 00:36:56.039 --> 00:36:58.659 they start the process there. Okay. Isn't it 00:36:58.659 --> 00:37:01.679 legally bound to provide them with protection? 00:37:02.420 --> 00:37:05.300 Yeah. That's as legally bound to keep their information 00:37:05.300 --> 00:37:09.000 confidential. And after you do that whistleblower, 00:37:09.019 --> 00:37:11.980 go to autosafety .org and click on donate. Bam. 00:37:12.239 --> 00:37:15.500 Do we have that whistleblower phone number on 00:37:15.500 --> 00:37:18.010 our website? Yep. Yep. And you can look at it. 00:37:18.050 --> 00:37:19.570 We have a little more information, too, on the 00:37:19.570 --> 00:37:22.130 new whistleblower act that was passed in 2021. 00:37:22.329 --> 00:37:24.349 I think that was part of the IJA. You can go 00:37:24.349 --> 00:37:26.449 to autosave .org and type in whistleblowers and 00:37:26.449 --> 00:37:31.130 you'll find it. Yeah. And that expanded the classification 00:37:31.130 --> 00:37:33.809 of people that can actually be whistleblowers 00:37:33.809 --> 00:37:36.829 to the extent that it covers, you know, the type 00:37:36.829 --> 00:37:39.550 of, you know, maybe not even employees, but contractors 00:37:39.550 --> 00:37:43.650 that are involved in some of these vehicle manufacturers, 00:37:44.030 --> 00:37:47.090 but definitely employees. All right. With that, 00:37:47.110 --> 00:37:51.630 I think that was an excellent towel. As somebody 00:37:51.630 --> 00:37:53.510 who peer -reviewed the paper you sent to Michael 00:37:53.510 --> 00:37:56.070 and I, I didn't understand it, but I'm glad I 00:37:56.070 --> 00:37:58.409 got the opportunity to peer -review it. Thank 00:37:58.409 --> 00:38:01.849 you. No problem. Thank you for the review. Sure. 00:38:02.010 --> 00:38:04.190 Let's jump into gaslights. I mean, I kind of 00:38:04.190 --> 00:38:06.829 vaguely threw one in there, but I'm going to 00:38:06.829 --> 00:38:10.090 start off. It's not really a gaslight. Maybe 00:38:10.090 --> 00:38:14.070 vaguely, it's a complaint against all auto manufacturers. 00:38:14.969 --> 00:38:19.090 And this past week, I was in a car called – it 00:38:19.090 --> 00:38:22.809 was made in Malaysia called a Proton. And I did 00:38:22.809 --> 00:38:25.250 what every smart person does in this situation. 00:38:25.530 --> 00:38:27.869 I texted Michael Brooks and said, Michael Brooks, 00:38:27.969 --> 00:38:29.949 I'm in a Malaysian -made car. Should I get out? 00:38:30.050 --> 00:38:32.269 And his response was, it's safer than being on 00:38:32.269 --> 00:38:36.590 a motorcycle. And he was right. So this Malaysian 00:38:36.590 --> 00:38:38.329 -made Proton – I've never heard of this before. 00:38:38.630 --> 00:38:42.690 Where were you? I was in Barbados. Okay. So it 00:38:42.690 --> 00:38:45.960 had a – It had a couple of two fascinating features 00:38:45.960 --> 00:38:48.699 that I think every auto manufacturer should have. 00:38:48.920 --> 00:38:52.940 One, you could put the car in drive, but it would 00:38:52.940 --> 00:38:56.199 not move unless every occupant's seatbelt was 00:38:56.199 --> 00:39:00.119 clicked. All right. Which was awesome. Two, you 00:39:00.119 --> 00:39:05.039 could not open the doors unless the car was stopped 00:39:05.039 --> 00:39:09.389 and in park. These are very simple things. Oh, 00:39:09.469 --> 00:39:11.090 you don't wait. Michael's making a face like 00:39:11.090 --> 00:39:12.389 he doesn't like that one. Well, that second one 00:39:12.389 --> 00:39:14.849 gives me serious concerns about post -crash escape. 00:39:16.789 --> 00:39:19.510 Okay, but I... You're not shifting into park 00:39:19.510 --> 00:39:22.849 after a crash to open your door, right? So presumably... 00:39:22.849 --> 00:39:27.590 Yeah, presumably that would have to be deactivated 00:39:27.590 --> 00:39:30.750 if a crash pulse was sensed. But even then, that 00:39:30.750 --> 00:39:33.670 didn't give me a whole lot of hope, right? Well, 00:39:33.769 --> 00:39:36.750 who knows? That may have been implemented. Because 00:39:36.750 --> 00:39:39.369 you didn't crash. No, but I liked it because 00:39:39.369 --> 00:39:42.130 one of the occupants of the car, my wife, was 00:39:42.130 --> 00:39:45.650 very excited by a fruit stand. And tried to jump 00:39:45.650 --> 00:39:47.369 out of the vehicle. I mean, the movie was like 00:39:47.369 --> 00:39:50.030 moving a mile per hour. And our friend who was 00:39:50.030 --> 00:39:51.590 driving, it was like, she's like, why can't I 00:39:51.590 --> 00:39:53.070 open the door? Why can't I open the door? That's 00:39:53.070 --> 00:39:54.670 great for some things. Like it's an automatic 00:39:54.670 --> 00:39:56.829 child lock. But, you know, just like some of 00:39:56.829 --> 00:40:02.039 the, one of the major problems with the. electric 00:40:02.039 --> 00:40:05.179 door latch problem that we've seen recently that 00:40:05.179 --> 00:40:07.880 there's some um movement on congress hopefully 00:40:07.880 --> 00:40:10.579 to fix problem that we've generally seen with 00:40:10.579 --> 00:40:14.019 teslas and some other electric vehicles and battery 00:40:14.019 --> 00:40:17.619 fires and escape is that there's a there's a 00:40:17.619 --> 00:40:19.760 kind of a conflict when it comes to the child 00:40:19.760 --> 00:40:23.030 safety locks right you're essentially When you 00:40:23.030 --> 00:40:25.889 turn on a child lock on a rear door, you're essentially 00:40:25.889 --> 00:40:29.010 disabling the ability of that person, that child, 00:40:29.070 --> 00:40:30.849 or anyone else who happens to sit in the back 00:40:30.849 --> 00:40:32.869 while your child locks are on to escape during 00:40:32.869 --> 00:40:36.429 a crash. And so there's an inherent conflict 00:40:36.429 --> 00:40:39.389 in any regulation that comes out of that process 00:40:39.389 --> 00:40:41.590 because you're going to have to figure out how 00:40:41.590 --> 00:40:45.550 to make child locks stay operable, protecting 00:40:45.550 --> 00:40:48.989 children in the rear while also being able to 00:40:48.989 --> 00:40:52.500 unlatch in case of a crash. Well, but, but you 00:40:52.500 --> 00:40:54.119 like the first one where it wouldn't, the car 00:40:54.119 --> 00:40:55.559 wouldn't drive unless all the seatbelts were 00:40:55.559 --> 00:40:57.880 clicked. Absolutely. Yeah. And I think I, you 00:40:57.880 --> 00:40:59.500 know, one of the, one of the areas I think where 00:40:59.500 --> 00:41:02.400 autonomous vehicles are beating humans for sure 00:41:02.400 --> 00:41:05.480 is that many of them require all occupants to 00:41:05.480 --> 00:41:09.079 be latched before they can proceed. Okay. Unlike, 00:41:09.159 --> 00:41:12.820 unlike our cars. So my, my simple gas light is, 00:41:12.820 --> 00:41:15.599 I don't know why any auto manufacturer is not 00:41:15.599 --> 00:41:18.639 doing this in their, you know. How much do you 00:41:18.639 --> 00:41:20.699 think that car actually cost? I don't know. Okay, 00:41:20.739 --> 00:41:22.400 so the rest of the car sucked, let's be honest. 00:41:22.500 --> 00:41:24.380 Like, yeah, we were stopped at a speed hump, 00:41:24.460 --> 00:41:27.219 like to pay airport parking, and you had to literally 00:41:27.219 --> 00:41:30.119 gun it to get over the speed hump, or else it 00:41:30.119 --> 00:41:33.340 just would not make it. It was a weird vehicle, 00:41:33.519 --> 00:41:35.780 but those two features I like, but now I only 00:41:35.780 --> 00:41:38.880 like one of those features. So that's my vague, 00:41:38.940 --> 00:41:41.139 not real gaslight. I'll let one of you guys win 00:41:41.139 --> 00:41:42.800 this week. Michael Brooks, would you like to 00:41:42.800 --> 00:41:46.250 do your gaslight? Yeah, I mean, I would love 00:41:46.250 --> 00:41:48.250 to just go ahead and say that those two quotes 00:41:48.250 --> 00:41:50.889 from the New York Times article from Ford and 00:41:50.889 --> 00:41:53.250 General Motors are gaslights. But this week, 00:41:53.269 --> 00:41:56.369 I'm going to venture into Fred territory and 00:41:56.369 --> 00:42:01.570 just do an easy one and give a gaslight to Waymo's. 00:42:03.909 --> 00:42:07.110 Waymo's recall remedies that are just, you know, 00:42:07.150 --> 00:42:09.030 they're continuing to do things on the school 00:42:09.030 --> 00:42:11.829 buses. They're continuing to do things on the 00:42:11.829 --> 00:42:14.909 problem with running into water. They're continuing 00:42:14.909 --> 00:42:19.949 to do things on the speeding. What was it? The 00:42:19.949 --> 00:42:23.289 freeway issue. Right. And they're putting out 00:42:23.289 --> 00:42:26.530 recalls and, you know, the recalls inevitably 00:42:26.530 --> 00:42:30.289 are failing down the road. And so, you know. 00:42:30.980 --> 00:42:35.599 We're being gaslit by this pretend world that 00:42:35.599 --> 00:42:37.460 Waymo's created where the recall remedies are 00:42:37.460 --> 00:42:40.179 going to do anything to change the safety characteristics 00:42:40.179 --> 00:42:43.920 of these vehicles. And, you know, it's concerning. 00:42:44.159 --> 00:42:46.179 And a lot of this has to do with what we were 00:42:46.179 --> 00:42:49.260 talking about earlier with AI and training. How 00:42:49.260 --> 00:42:51.599 do you ensure that consistent performance over 00:42:51.599 --> 00:42:55.500 time? And I just think that needs to be acknowledged 00:42:55.500 --> 00:42:58.099 more in this case. I feel like Waymo, when they 00:42:58.099 --> 00:43:00.320 experience a big problem like this, instead of 00:43:00.320 --> 00:43:02.019 pushing forward and saying, we're going to continue 00:43:02.019 --> 00:43:05.420 operating, pulls back, says, we're going to stop 00:43:05.420 --> 00:43:08.940 this until we have at least some certainty that 00:43:08.940 --> 00:43:13.019 this problem isn't going to continue. Issuing 00:43:13.019 --> 00:43:15.539 a recall that's destined to fail, as they've 00:43:15.539 --> 00:43:17.719 done in a lot of cases recently, is not going 00:43:17.719 --> 00:43:20.840 to work. Pretty good. All right, Mr. Perkins, 00:43:21.039 --> 00:43:24.260 it's your time to shine. I'm going with Tesla's 00:43:24.260 --> 00:43:27.380 homepage because interesting things on there. 00:43:27.920 --> 00:43:30.639 One of the assertions they make is there's seven 00:43:30.639 --> 00:43:33.599 times fewer collisions. They don't say relative 00:43:33.599 --> 00:43:36.800 to what, but there's no references or analysis 00:43:36.800 --> 00:43:40.519 or caveats provided. But what they show there 00:43:40.519 --> 00:43:45.440 are six drivers who are shown using FSD underway. 00:43:46.280 --> 00:43:49.679 And one of them is making espresso in the console 00:43:49.679 --> 00:43:53.000 while looking away from the road. One of them 00:43:53.000 --> 00:43:55.480 is butt dancing while looking away from the road 00:43:55.480 --> 00:43:57.820 and talking to the videographer. Wait, what is 00:43:57.820 --> 00:44:00.719 butt dancing? I would assume they're shifting 00:44:00.719 --> 00:44:03.219 cheeks, right? You know, it's what you do when 00:44:03.219 --> 00:44:05.460 Jumpin' Jack Flash comes on the radio. Okay. 00:44:06.500 --> 00:44:08.739 I'm more of a steering wheel drummer, but you 00:44:08.739 --> 00:44:12.039 know. One is an elderly woman who's shown looking 00:44:12.039 --> 00:44:14.800 from left to right away from the direction of 00:44:14.800 --> 00:44:18.510 travel. Another is looking away from the direction 00:44:18.510 --> 00:44:21.250 of travel. And all of them are distracted by 00:44:21.250 --> 00:44:24.789 communicating with the videographer while underway. 00:44:26.670 --> 00:44:30.429 One of their other video shows a Tesla passing 00:44:30.429 --> 00:44:33.690 at a relatively high speed, passing another car 00:44:33.690 --> 00:44:38.469 at a relatively high speed during rain that certainly 00:44:38.469 --> 00:44:40.530 degrades the sensors and makes it difficult to 00:44:40.530 --> 00:44:43.889 look out of the windshield. So that was... Pure 00:44:43.889 --> 00:44:47.670 gaslight, I think, on top of the claim of 7x 00:44:47.670 --> 00:44:50.989 fewer collisions, which is meaningless. But they 00:44:50.989 --> 00:44:55.070 go on to say that they don't share any of your 00:44:55.070 --> 00:44:57.650 data and they don't keep any of your data. And 00:44:57.650 --> 00:45:00.429 then if you dive in a little deeper, it says 00:45:00.429 --> 00:45:03.489 how to obtain a copy of the data associated with 00:45:03.489 --> 00:45:06.909 your Tesla account. There should just be an asterisk 00:45:06.909 --> 00:45:09.469 about everything Tesla puts out. Just kidding. 00:45:09.650 --> 00:45:12.170 Then it says the data associated with your Tesla 00:45:12.170 --> 00:45:14.840 account includes, Order details, account information, 00:45:15.260 --> 00:45:18.559 customer support, activity, service history, 00:45:18.760 --> 00:45:21.860 vehicle usage information, safety events, infotainment 00:45:21.860 --> 00:45:25.980 system settings, mobile app, supercharging history. 00:45:26.420 --> 00:45:28.880 So this is all part of the data that they say 00:45:28.880 --> 00:45:32.400 they don't collect. It's very interesting. And 00:45:32.400 --> 00:45:37.699 then they say, you know, in their consumer privacy 00:45:37.699 --> 00:45:41.219 notice, information we may collect. We may collect 00:45:41.219 --> 00:45:45.679 three main types. You use the product's information 00:45:45.679 --> 00:45:48.940 from or about you, information from or about 00:45:48.940 --> 00:45:54.260 your Tesla vehicle. So I give Tesla my nomination 00:45:54.260 --> 00:45:58.900 for complete doublespeak and complete internal 00:45:58.900 --> 00:46:03.000 contradiction of their safety and state of privacy 00:46:03.000 --> 00:46:06.880 standards. Fred, you won because you made Michael 00:46:06.880 --> 00:46:10.800 and I laugh. So that's how you win. But I'm surprised 00:46:10.800 --> 00:46:14.710 neither of you did. a different Waymo take. So 00:46:14.710 --> 00:46:16.409 we're going to jump into this, whereas we've 00:46:16.409 --> 00:46:17.889 talked about in the past, Fred in particular 00:46:17.889 --> 00:46:20.090 has talked in the past about Waymo citing their 00:46:20.090 --> 00:46:22.070 data about how this is how we're safer than a 00:46:22.070 --> 00:46:25.489 human. We paid our employees to tell us that 00:46:25.489 --> 00:46:27.929 we're safer than a human, or we paid these researchers 00:46:27.929 --> 00:46:30.230 to tell us the world that we're safer than a 00:46:30.230 --> 00:46:33.090 human, and we're going to use our data from 2020 00:46:33.090 --> 00:46:37.670 to back this up, even though it's 2026. So while 00:46:37.670 --> 00:46:40.449 Waymo is, I guess maybe they've heard the complaints 00:46:40.449 --> 00:46:43.800 of Fred and... And now Waymo has come up with 00:46:43.800 --> 00:46:46.840 a new model to decide if how to show the world 00:46:46.840 --> 00:46:48.519 how they're safer than a human. According from 00:46:48.519 --> 00:46:51.099 a tech crunch piece, the new model, Waymo said, 00:46:51.239 --> 00:46:54.519 evolves this concept serving as a behavioral 00:46:54.519 --> 00:46:56.900 benchmark for autonomous driving systems able 00:46:56.900 --> 00:46:59.599 to realistically represent reasonable expectations 00:46:59.599 --> 00:47:03.139 on how a careful and competent human driver responds 00:47:03.139 --> 00:47:06.719 to traffic conflicts. What first year law student 00:47:06.719 --> 00:47:11.780 wrote that? Like, this is just, I mean. There 00:47:11.780 --> 00:47:15.219 are so many caveats on here to realistically 00:47:15.219 --> 00:47:21.119 represent reasonable expectations. A lot of weasel 00:47:21.119 --> 00:47:24.199 words. Yeah, 100%. So Waymo is saying, wait, 00:47:24.300 --> 00:47:26.920 people are catching on that we're talking bullshit. 00:47:26.940 --> 00:47:29.139 So instead, we're going to invent a fictitious 00:47:29.139 --> 00:47:33.019 person and compare our driver to this fictitious 00:47:33.019 --> 00:47:37.260 person in a reasonable kind of way, sort of, 00:47:37.260 --> 00:47:40.639 depending if it's Tuesday. And again, I want 00:47:40.639 --> 00:47:43.079 to repeat, this is the company that hit a child 00:47:43.079 --> 00:47:45.300 and said, we hit it at a lower speed than a human 00:47:45.300 --> 00:47:48.280 would have. Instead of saying, fuck, we hit a 00:47:48.280 --> 00:47:51.760 child. You know, like that's the correct response. 00:47:52.659 --> 00:47:55.699 So that would have been the greatest gaslight. 00:47:55.840 --> 00:47:59.019 I mean, I'm interested in this to the extent 00:47:59.019 --> 00:48:02.619 that... You need another gaslight? No, to the 00:48:02.619 --> 00:48:05.920 extent that they're comparing the performance 00:48:05.920 --> 00:48:09.199 of Waymo vehicles to the performance of... good 00:48:09.199 --> 00:48:12.280 human drivers, right? All we hear about from 00:48:12.280 --> 00:48:14.599 the autonomous vehicle industry is about the 00:48:14.599 --> 00:48:17.099 drunks and the distracted people and how they're 00:48:17.099 --> 00:48:18.900 better than them. Well, of course they are. We're 00:48:18.900 --> 00:48:21.719 all better. I mean, a lot of us. A vast majority 00:48:21.719 --> 00:48:24.179 of drivers are better than those people, right? 00:48:24.340 --> 00:48:27.260 So comparing your autonomous vehicle to a shitty 00:48:27.260 --> 00:48:30.300 driver doesn't prove anything to us, right? So 00:48:30.300 --> 00:48:33.400 maybe here they're comparing to a good driver 00:48:33.400 --> 00:48:35.619 and we'll get to see what happens. I'm not sure 00:48:35.619 --> 00:48:39.880 that this is all going to be an exercise we can 00:48:39.880 --> 00:48:44.130 rely on, but it's. It's the progress, I think, 00:48:44.130 --> 00:48:46.730 from what has been going on in the autonomous 00:48:46.730 --> 00:48:49.789 vehicle industry, going back to Kyle and Cruz, 00:48:49.969 --> 00:48:54.389 where they are insistent that the autonomous 00:48:54.389 --> 00:48:56.889 vehicles are better than humans, but they're 00:48:56.889 --> 00:49:00.429 only looking at the worst of us. All right. From 00:49:00.429 --> 00:49:04.230 a technical perspective, it's a simulation, right? 00:49:04.309 --> 00:49:07.090 And every simulation is doomed to failure or 00:49:07.090 --> 00:49:10.690 doomed to success, I should say, in every circumstance. 00:49:11.990 --> 00:49:14.730 what's key here is that they haven't talked about 00:49:14.730 --> 00:49:17.570 how they validated this, what test sequences 00:49:17.570 --> 00:49:20.710 they used, what situations they use, and why 00:49:20.710 --> 00:49:22.909 should anybody believe that? It's just another 00:49:22.909 --> 00:49:26.210 computer model. So, yeah. I look forward to this 00:49:26.210 --> 00:49:32.530 reasonable driver to change over time to help 00:49:32.530 --> 00:49:35.309 their numbers. I think it's hilarious. But Michael 00:49:35.309 --> 00:49:39.570 mentioned Kyle and Cruz, which jumps us to General 00:49:39.570 --> 00:49:43.000 Motors. From an article from thenextweb .com, 00:49:43.079 --> 00:49:47.219 I'm going to quote, GM is rehiring its cruise 00:49:47.219 --> 00:49:50.159 employees. Quoting, the rehiring is a reversal 00:49:50.159 --> 00:49:53.280 from a painful exit. GM acquired Cruise in 2016 00:49:53.280 --> 00:49:56.900 for $581 million and spent more than $10 billion 00:49:56.900 --> 00:50:01.059 on the division before pulling the plug. This 00:50:01.059 --> 00:50:03.320 is when, you know, it hit somebody, dragged a 00:50:03.320 --> 00:50:05.059 person, lied about it, and California regulators 00:50:05.059 --> 00:50:07.679 suspended the company's driver's license. So 00:50:07.679 --> 00:50:10.599 the person running this new autonomous section 00:50:10.599 --> 00:50:15.039 of GM has said, hey, you know, RoboTax is probably 00:50:15.039 --> 00:50:16.980 going to be in our future. Why not? That's not 00:50:16.980 --> 00:50:18.860 how we're going to start. And the person who 00:50:18.860 --> 00:50:20.820 runs this used to run a company called Aurora. 00:50:21.519 --> 00:50:24.360 And we've talked about Aurora, Aurora Trucking, 00:50:24.360 --> 00:50:27.719 and he's getting paid, what was it, $30 million, 00:50:28.119 --> 00:50:30.460 something like that, $40 million. He was hired 00:50:30.460 --> 00:50:33.559 last in May 2025 with a $40 million compensation 00:50:33.559 --> 00:50:37.409 package. And he ran Aurora, and Aurora, as best 00:50:37.409 --> 00:50:40.429 anyone can tell, has two customers and still 00:50:40.429 --> 00:50:42.590 just operating in Texas. Now they have a route 00:50:42.590 --> 00:50:45.670 to Oklahoma as well, a very limited area. But 00:50:45.670 --> 00:50:52.230 GM is like, hey, he's different than Kyle. Is 00:50:52.230 --> 00:50:55.769 he? I don't know. GM, like. Well, he also, you 00:50:55.769 --> 00:50:57.510 know, let's let's you know, I don't know if I 00:50:57.510 --> 00:50:59.110 can't. I may have missed it if you mentioned 00:50:59.110 --> 00:51:02.050 this, but he also, you know, ran Tesla's autopilot. 00:51:03.110 --> 00:51:06.789 So just a little background. His resume is nicer 00:51:06.789 --> 00:51:09.829 looking than Kyle. But, you know, and it's just 00:51:09.829 --> 00:51:12.829 I don't understand what GM is doing here. Yeah, 00:51:12.829 --> 00:51:16.699 I will say that Aurora is. At least from my perspective, 00:51:16.960 --> 00:51:19.679 a whole hell of a lot more careful than Tesla. 00:51:19.940 --> 00:51:23.420 So I don't know. Yeah. Yeah. But again, as far 00:51:23.420 --> 00:51:25.420 as I can tell, I was doing looking into this 00:51:25.420 --> 00:51:28.320 before saying, like, what's going on with Aurora? 00:51:28.619 --> 00:51:31.539 And like, there's not really any news out of 00:51:31.539 --> 00:51:34.260 that company. It's still doing that, mainly that 00:51:34.260 --> 00:51:38.440 same short run in Texas. Were they the company 00:51:38.440 --> 00:51:40.840 that their customer said put human drivers back 00:51:40.840 --> 00:51:43.650 in the cab? Was that Aurora? Yeah, I had to put 00:51:43.650 --> 00:51:46.070 it back in because I think it was Packer, Packer, 00:51:46.070 --> 00:51:49.050 one of those companies wouldn't essentially said 00:51:49.050 --> 00:51:51.150 the liabilities too much for us. They didn't 00:51:51.150 --> 00:51:52.610 come out public and say that, but it was pretty 00:51:52.610 --> 00:51:54.030 clear that's what they were concerned about. 00:51:54.309 --> 00:51:58.469 So anyway, we'll keep you in touch as GM shows 00:51:58.469 --> 00:52:00.929 how they can set billions of dollars on. Yeah. 00:52:00.949 --> 00:52:03.429 And GM is saying a lot of weird stuff like, you 00:52:03.429 --> 00:52:06.449 know, at least. Sterling Anderson is, you know, 00:52:06.469 --> 00:52:12.570 after years of saying that these autonomous vehicle 00:52:12.570 --> 00:52:14.670 companies saying, you know, you can't build, 00:52:14.670 --> 00:52:18.070 you know, level two, level three, you can't build 00:52:18.070 --> 00:52:20.230 upwards. These are separate technologies, autonomy 00:52:20.230 --> 00:52:22.750 and advanced driver assistance systems. Here's, 00:52:22.750 --> 00:52:24.550 you know, Sterling Anderson saying, ultimately 00:52:24.550 --> 00:52:28.400 the two converge, right? There's just so much 00:52:28.400 --> 00:52:31.420 nonsense out there. It's bizarre. With that, 00:52:31.519 --> 00:52:33.659 we got to run into recalls because we have a 00:52:33.659 --> 00:52:35.860 massive, massive list of recalls. We're not doing 00:52:35.860 --> 00:52:38.719 all of them. What do you mean? I want to do all 00:52:38.719 --> 00:52:41.860 of them. You want to do all 11 Ford recalls? 00:52:42.119 --> 00:52:44.440 Don't ruin it for the public. First, we're going 00:52:44.440 --> 00:52:46.199 to start off with Waymo because they've got a 00:52:46.199 --> 00:52:50.900 recall of 3 ,871 vehicles, the fifth generation 00:52:50.900 --> 00:52:54.820 ADS. Under certain circumstances, the AV may 00:52:54.820 --> 00:52:57.559 enter and drive at speed in freeway construction 00:52:57.559 --> 00:53:00.239 zones due to inappropriately prioritizing the 00:53:00.239 --> 00:53:02.820 avoidance of other freeway hazard and or failing 00:53:02.820 --> 00:53:07.199 to recognize the construction zone. Safer than 00:53:07.199 --> 00:53:13.400 a human. Waymo. Yeah. I mean, the remedy will 00:53:13.400 --> 00:53:15.659 enhance Waymo's AV performance in and around 00:53:15.659 --> 00:53:18.530 freeway construction zones. How do you know? 00:53:18.730 --> 00:53:22.090 We don't. It's again, they had the simple solution 00:53:22.090 --> 00:53:25.590 of saying, hey, let's take things off the freeway 00:53:25.590 --> 00:53:27.869 for a while. Hey, let's stop. They're still off 00:53:27.869 --> 00:53:29.789 the freeways, right? I mean, that's a geofence 00:53:29.789 --> 00:53:31.989 essentially, right? That works because you're 00:53:31.989 --> 00:53:34.530 not going to be putting yourself in this scenario 00:53:34.530 --> 00:53:37.829 anymore. But now they're saying, you know, we've 00:53:37.829 --> 00:53:40.070 got a software improvement and a remedy. So we're 00:53:40.070 --> 00:53:42.030 going to be able to perform on the freeway. So 00:53:42.030 --> 00:53:44.690 presumably they're going to. put their vehicles 00:53:44.690 --> 00:53:46.909 back into operation on 3As, and then we're going 00:53:46.909 --> 00:53:50.230 to get to see if this recall works, right? Right. 00:53:50.909 --> 00:53:53.570 And then the next recalls, as Michael alluded, 00:53:53.789 --> 00:53:57.389 so he sent an article to us from USA Today, and 00:53:57.389 --> 00:54:01.429 that Ford has more recalls than any other automaker 00:54:01.429 --> 00:54:06.230 by far, as we know this. They have 51 recalls. 00:54:06.789 --> 00:54:09.389 I think that's pre these new 11 ones, right? 00:54:10.659 --> 00:54:13.519 And Michael Brooks. Oh, that's with them. And 00:54:13.519 --> 00:54:15.619 Michael Brooks is quoted in this article. And 00:54:15.619 --> 00:54:17.760 Michael, what's the reason that Ford is leading 00:54:17.760 --> 00:54:24.280 the recall bunch? So in late 2024, NHTSA, I'm 00:54:24.280 --> 00:54:26.599 sure we've discussed this before, but NHTSA basically 00:54:26.599 --> 00:54:29.480 cracked down on Ford for a lot of different violations 00:54:29.480 --> 00:54:33.659 of the Safety Act. And they fined them or. set 00:54:33.659 --> 00:54:36.239 up a consent agreement where Ford could be fined. 00:54:36.239 --> 00:54:40.860 I think it was about $165 million. And they made 00:54:40.860 --> 00:54:45.400 Ford essentially go under Ford's recall and safety. 00:54:46.199 --> 00:54:49.000 programs are having to run through an independent 00:54:49.000 --> 00:54:52.780 monitor type situation for three or four years. 00:54:52.960 --> 00:54:55.980 And that runs until 2027. You know, virtually 00:54:55.980 --> 00:54:58.440 ever since we've seen that consent order go into 00:54:58.440 --> 00:55:01.440 place, Ford has been pumping out recalls left 00:55:01.440 --> 00:55:03.480 and right. They've been going back and fixing 00:55:03.480 --> 00:55:06.739 a lot of bad recall remedies that came out. And 00:55:06.739 --> 00:55:08.780 they've also been going forward and fixing, you 00:55:08.780 --> 00:55:11.159 know, new recall, new safety issues on their 00:55:11.159 --> 00:55:15.690 vehicles. And so since 2024, they have been I 00:55:15.690 --> 00:55:18.289 mean, crushing the competition, I guess you could 00:55:18.289 --> 00:55:20.409 say, on the recall charts. They've significantly 00:55:20.409 --> 00:55:24.510 more recalls than any other manufacturer. And, 00:55:24.630 --> 00:55:26.630 you know, this is the latest story on how they're 00:55:26.630 --> 00:55:28.969 leading this year, but they're leading, you know, 00:55:28.969 --> 00:55:31.710 for the past three years, essentially. And I 00:55:31.710 --> 00:55:35.829 don't see that stopping at least. through 2027, 00:55:35.909 --> 00:55:37.809 which is the year where they're finally going 00:55:37.809 --> 00:55:41.289 to be released, maybe, from the NHTSA consent 00:55:41.289 --> 00:55:43.989 agreement, they have to show that they haven't 00:55:43.989 --> 00:55:46.369 committed any more Safety Act violations during 00:55:46.369 --> 00:55:48.750 that period, which is why I think we're seeing 00:55:48.750 --> 00:55:52.170 Ford do so many recalls and be so involved in 00:55:52.170 --> 00:55:54.809 this area, because if they don't get released 00:55:54.809 --> 00:55:56.250 from that consent order, they're going to lose, 00:55:56.289 --> 00:55:59.329 I think it's another $50 million or so. So that's 00:55:59.329 --> 00:56:01.949 hanging in the balance, and Ford doesn't want 00:56:01.949 --> 00:56:04.579 to lose that money. But why would they on June 00:56:04.579 --> 00:56:07.860 9th of this year release 11 recalls all at once? 00:56:07.960 --> 00:56:09.719 And this is a Tuesday. It's not like a late Friday 00:56:09.719 --> 00:56:12.880 afternoon. No one will see a thing. Yeah, that's 00:56:12.880 --> 00:56:16.159 I mean, that's that's odd that you see a manufacturer. 00:56:16.300 --> 00:56:18.159 You know, some manufacturers you'll see they'll 00:56:18.159 --> 00:56:20.320 have their quarterly recall meeting or their 00:56:20.320 --> 00:56:23.300 monthly recall safety meeting. And you'll see, 00:56:23.320 --> 00:56:25.579 you know, two, two recalls, three recalls come 00:56:25.579 --> 00:56:29.860 out of that. Eleven is. A very large number is 00:56:29.860 --> 00:56:31.780 all at once. Okay, so let's run through some 00:56:31.780 --> 00:56:34.340 of these or maybe all of these. Unexpected downshift 00:56:34.340 --> 00:56:38.239 can cause loss of control. 44 ,963 vehicles. 00:56:39.460 --> 00:56:42.360 Their transmission output shaft speed sensor 00:56:42.360 --> 00:56:46.260 and powertrain control module can cause the vehicle 00:56:46.260 --> 00:56:49.300 to unexpectedly shift to first gear. Oh, God. 00:56:49.659 --> 00:56:51.420 Yeah, well, you're just driving along and all 00:56:51.420 --> 00:56:53.219 of a sudden you shifted to first gear. Can't 00:56:53.219 --> 00:56:56.940 be good, right? This is a recall of a recall. 00:56:57.119 --> 00:56:59.980 So they've already done a recall on this in 2024. 00:57:00.460 --> 00:57:03.800 Not a good recall under the order with NHTSA. 00:57:03.900 --> 00:57:05.659 They've got to make sure they're going back and 00:57:05.659 --> 00:57:08.039 fixing bad recalls from the past. So here we 00:57:08.039 --> 00:57:10.360 go. They're going to put out another program. 00:57:10.920 --> 00:57:13.300 It looks like it's going to be another software 00:57:13.300 --> 00:57:16.139 update of the power control module software that 00:57:16.139 --> 00:57:18.889 owners should hear about. In the next two weeks. 00:57:19.110 --> 00:57:22.230 Okay, I'm going to jump into another one. Transmission 00:57:22.230 --> 00:57:26.030 clutch may fracture and leak fluid. 5 ,252 vehicles. 00:57:26.630 --> 00:57:28.690 And this one's an obvious to me because it's 00:57:28.690 --> 00:57:31.030 focused on the Ford Focus and Fusion vehicles. 00:57:32.030 --> 00:57:35.010 You know, these were not the best made vehicles 00:57:35.010 --> 00:57:38.610 of all time. And the clutch may fracture, resulting 00:57:38.610 --> 00:57:40.829 in damage to the transmission assembly and possibly 00:57:40.829 --> 00:57:44.449 transmission fluid leak. In this case, I'm sorry. 00:57:44.590 --> 00:57:47.510 There have been massive problems with the fusions 00:57:47.510 --> 00:57:49.530 from those model years and their transmission. 00:57:49.809 --> 00:57:54.530 So it's, you know, some of the vehicles being 00:57:54.530 --> 00:57:56.190 serviced under this have been on the road for 00:57:56.190 --> 00:57:59.849 about 14 years. So clearly they're continuing 00:57:59.849 --> 00:58:02.670 to see some significant problems with those transmissions. 00:58:02.730 --> 00:58:04.969 And I'm glad they're taking action there because 00:58:04.969 --> 00:58:07.829 that's an issue that has been out there for about 00:58:07.829 --> 00:58:10.730 a decade or more now. Right. Next up, unintended 00:58:10.730 --> 00:58:15.869 vehicle movement. This is 10 ,742 vehicles. The 00:58:15.869 --> 00:58:20.989 2018 F -150 previously repaired incorrectly under 00:58:20.989 --> 00:58:25.690 another recall. Yep. Oh, no. Quickly moving the 00:58:25.690 --> 00:58:27.809 shift lever from park to drive may cause a loss 00:58:27.809 --> 00:58:29.570 of gear indication on the instrument cluster 00:58:29.570 --> 00:58:32.409 and a momentary selection of an unintended gear 00:58:32.409 --> 00:58:36.070 such as reverse or neutral. Yikes. So they had 00:58:36.070 --> 00:58:39.449 a recall on this, and their recall was a failure. 00:58:40.880 --> 00:58:45.059 Is that what I just read? Yeah. And essentially 00:58:45.059 --> 00:58:48.179 they're saying, quoting from Ford, you know, 00:58:48.179 --> 00:58:50.719 the underlying safety risks specified in the 00:58:50.719 --> 00:58:54.820 2017 recall still exists on this vehicle. So 00:58:54.820 --> 00:58:57.699 apparently the fix that they put into play in 00:58:57.699 --> 00:59:02.219 2017 did not work. So they're going back and 00:59:02.219 --> 00:59:04.420 doing another fix here. Okay. I want to point 00:59:04.420 --> 00:59:06.300 out that computers will not make everything better. 00:59:06.420 --> 00:59:10.760 Engineering is hard. Next up window automatic 00:59:10.760 --> 00:59:15.239 reversal system malfunction. This is 18 ,124 00:59:15.239 --> 00:59:18.980 vehicles, certain 2017 escape vehicles repaired 00:59:18.980 --> 00:59:22.440 incorrectly under another recall. Oh my God. 00:59:22.480 --> 00:59:24.840 And this is like, if you have your arm or your 00:59:24.840 --> 00:59:27.199 neck in the way, it just, it doesn't reverse 00:59:27.199 --> 00:59:32.579 automatically the window. Right. Right. And they're 00:59:32.579 --> 00:59:36.539 not. So there's, there's kind of a weird. Federal 00:59:36.539 --> 00:59:38.699 Motor Vehicle Safety Standard 118 essentially 00:59:38.699 --> 00:59:41.739 prevents, the main goal of this is to prevent 00:59:41.739 --> 00:59:44.559 children from getting entrapped in windows while 00:59:44.559 --> 00:59:46.619 screwing around with the buttons in the car, 00:59:46.780 --> 00:59:48.360 that type of thing. That's why you no longer 00:59:48.360 --> 00:59:52.619 see the rocker switches to control your windows, 00:59:52.719 --> 00:59:54.719 and you're typically having to pull up on a button 00:59:54.719 --> 00:59:58.340 because it prevents actuation of the switch by 00:59:58.340 --> 01:00:01.599 a child's knee or an unintentional actuation. 01:00:02.480 --> 01:00:05.739 And so in this situation, when they upgraded 01:00:05.739 --> 01:00:09.400 that standard in the early 2000s, they provided 01:00:09.400 --> 01:00:11.739 a pathway that essentially let manufacturers 01:00:11.739 --> 01:00:16.199 do one thing and avoid – do one thing to pass 01:00:16.199 --> 01:00:19.179 or install automatic reversing. A lot of vehicles 01:00:19.179 --> 01:00:22.019 have had this technology. I think my 2003 Volkswagen 01:00:22.019 --> 01:00:24.860 Jetta had auto -reverse in it. We used it as 01:00:24.860 --> 01:00:27.420 a prop in one of our press conferences around 01:00:27.420 --> 01:00:30.730 this in the early 2000s. Auto reverse technology 01:00:30.730 --> 01:00:34.730 is old and it's cheap and it should probably 01:00:34.730 --> 01:00:37.789 be going into every vehicle, even though it's 01:00:37.789 --> 01:00:40.250 not explicitly required. All right. I'm going 01:00:40.250 --> 01:00:42.610 to jump to someone related to instrument cluster, 01:00:42.789 --> 01:00:46.769 instrument cluster failures. 4 ,445 vehicles, 01:00:47.110 --> 01:00:51.809 2017 F -150s, they are repaired incorrectly again. 01:00:52.469 --> 01:00:54.670 Yeah. So you're seeing, and this is something 01:00:54.670 --> 01:00:58.010 that I've been trying to point out to a lot of 01:00:58.010 --> 01:01:00.070 reporters when they call about Ford, because 01:01:00.070 --> 01:01:02.389 there's a lot of questions coming in on Ford 01:01:02.389 --> 01:01:04.409 recalls, because the numbers, you know, number 01:01:04.409 --> 01:01:05.949 of recalls they're putting out is crazy. But 01:01:05.949 --> 01:01:08.849 a lot of these recalls they're doing are actually 01:01:08.849 --> 01:01:13.750 fixes of previous recall repairs. They did a 01:01:13.750 --> 01:01:16.449 really bad job on some recalls. If you look through 01:01:16.449 --> 01:01:20.989 all this, you can pick out dozens of recalls 01:01:20.989 --> 01:01:25.170 over the past three years since the NHTSA consent 01:01:25.170 --> 01:01:27.989 order where Ford's going back and saying, oh, 01:01:28.030 --> 01:01:31.250 we screwed this one up. So I think that's reflective 01:01:31.250 --> 01:01:33.969 of just how poorly they were doing and how necessary 01:01:33.969 --> 01:01:37.030 that action by NHTSA was. So what's the bottom 01:01:37.030 --> 01:01:41.199 line here for our listeners? Bought a Ford with 01:01:41.199 --> 01:01:45.019 it, or if their Ford is less than, what, 10 years 01:01:45.019 --> 01:01:47.980 old, they should make sure that they check in 01:01:47.980 --> 01:01:50.039 with their dealer. Yeah, I would stay on top 01:01:50.039 --> 01:01:54.619 of, you know, I would be plugging my vehicle's 01:01:54.619 --> 01:01:56.980 VIN number or my license plate in the NHTSA website 01:01:56.980 --> 01:01:59.980 a lot more because, you know. If you own a Ford, 01:02:00.099 --> 01:02:02.800 frankly, you're going to have more recalls coming 01:02:02.800 --> 01:02:04.679 out on your vehicle than any other manufacturer 01:02:04.679 --> 01:02:08.119 on average, right? So you're going to need to 01:02:08.119 --> 01:02:10.139 stay on top of those because an issue you thought 01:02:10.139 --> 01:02:13.980 might have been fixed in 2017 wasn't fixed. And 01:02:13.980 --> 01:02:16.860 that's got to be confusing for consumers. They 01:02:16.860 --> 01:02:18.559 say they get a recall notice. Oh, my instrument 01:02:18.559 --> 01:02:20.420 cluster. No, no, no. They fixed that two years 01:02:20.420 --> 01:02:23.039 ago. This must be a repeat of the same thing. 01:02:24.139 --> 01:02:26.500 Like when they do a recall of a recall. You just 01:02:26.500 --> 01:02:28.300 need to understand anytime you get a new recall 01:02:28.300 --> 01:02:30.559 on your vehicle that it's going to require a 01:02:30.559 --> 01:02:34.039 remedy, right? Just because you had something 01:02:34.039 --> 01:02:37.099 fixed once doesn't mean that fix was proper or 01:02:37.099 --> 01:02:40.159 appropriate or worked. And so anytime you get 01:02:40.159 --> 01:02:42.480 a new recall notice, you need to do due diligence. 01:02:42.719 --> 01:02:44.500 You know, you can't poo -poo a recall notice. 01:02:45.019 --> 01:02:48.539 So where do people find the VIN for their vehicle? 01:02:49.199 --> 01:02:51.340 Well, there's I mean, there's a lot of places 01:02:51.340 --> 01:02:54.400 you can look on your on your registration. You 01:02:54.400 --> 01:02:57.699 can look on the driver's side windshield. There's 01:02:57.699 --> 01:03:00.159 usually a bar there, which will have your VIN 01:03:00.159 --> 01:03:03.079 on it. And I believe it's on the placard on the 01:03:03.079 --> 01:03:06.219 inside of your driver's door as well. So what 01:03:06.219 --> 01:03:09.449 people should do is. Take the registration. You 01:03:09.449 --> 01:03:11.010 don't even need that. You can go to the NHTSA 01:03:11.010 --> 01:03:13.070 website, put your license plate number in it. 01:03:13.070 --> 01:03:15.909 They've made it so easy to look up active recalls 01:03:15.909 --> 01:03:18.590 on vehicles now without having to dig up your 01:03:18.590 --> 01:03:21.170 VIN number in all 17 digits. Just put your license 01:03:21.170 --> 01:03:22.650 plate number in and it'll look it up for you. 01:03:23.150 --> 01:03:28.010 And the NHTSA website is what? NHTSA .gov. Yep. 01:03:28.309 --> 01:03:30.210 Even simpler. If you discover you have a Ford 01:03:30.210 --> 01:03:32.590 vehicle, sell it off to somebody you don't like. 01:03:32.869 --> 01:03:36.300 You know, give them a deal. And another one, 01:03:36.360 --> 01:03:39.059 another instrument cluster may go blank on startup 01:03:39.059 --> 01:03:43.559 2349 vehicles, the 2019 Mustang Nautilus Navigator. 01:03:43.820 --> 01:03:48.760 And yeah, this is the dealership program, the 01:03:48.760 --> 01:03:51.780 instrument cluster cluster panel cluster assembly. 01:03:52.079 --> 01:03:55.500 This is just come on, guys. Yeah, I mean, we 01:03:55.500 --> 01:03:58.300 continue to see instrument panel failures. I 01:03:58.300 --> 01:04:04.900 there's some. I think between all these instrument 01:04:04.900 --> 01:04:11.039 cluster failures and all of the backup camera 01:04:11.039 --> 01:04:13.460 failures that we're seeing too. All right. So 01:04:13.460 --> 01:04:15.480 I'm just going to do two more of these. And these 01:04:15.480 --> 01:04:19.019 are the ones with big numbers. Engine stall from 01:04:19.019 --> 01:04:23.380 canister purge valve malfunction. This is 255 01:04:23.380 --> 01:04:30.500 ,404 vehicles, 2012 to 2018 Focus vehicles. that 01:04:30.500 --> 01:04:33.420 were previously repaired incorrectly under a 01:04:33.420 --> 01:04:38.480 recall. What is a canister purge valve? I have 01:04:38.480 --> 01:04:42.860 no idea. Oh, no. Well, you know, if Fred has 01:04:42.860 --> 01:04:45.619 no idea, probably just throw it out. You don't 01:04:45.619 --> 01:04:48.079 need it. Yeah, it's got something to do. It's 01:04:48.079 --> 01:04:50.460 got, I think it, I think it. Something to do 01:04:50.460 --> 01:04:53.199 with the power. This is similar to, I think some 01:04:53.199 --> 01:04:55.119 of the, I think it was a Hyundai and Kia recalls 01:04:55.119 --> 01:05:01.880 that we saw where the, the gas tank. The canister 01:05:01.880 --> 01:05:04.820 purge valve has something to do with the evaporative 01:05:04.820 --> 01:05:07.760 emissions in the gas in the fuel tank and preventing 01:05:07.760 --> 01:05:10.039 pressures from building up in the fuel tank. 01:05:10.300 --> 01:05:13.760 So when the CPV sticks open, they're saying it 01:05:13.760 --> 01:05:16.679 can cause excess vacuum in the fuel system, which 01:05:16.679 --> 01:05:18.719 can deform the fuel tank. I think we talked about 01:05:18.719 --> 01:05:21.800 Hyundai and Kia. Right. We had a guest on about 01:05:21.800 --> 01:05:26.639 that years ago. Yeah. I get you. Last one. Center 01:05:26.639 --> 01:05:32.389 console may peel, causing sharp edges. 548 ,463 01:05:32.389 --> 01:05:36.789 vehicles. 2018 to 2024 expedition. The chrome 01:05:36.789 --> 01:05:38.989 plating on the center console may bubble and 01:05:38.989 --> 01:05:44.510 peel resulting in sharp edges. I mean, these 01:05:44.510 --> 01:05:46.190 are one of these things where people are like, 01:05:46.230 --> 01:05:48.829 you know, software will fix everything. These 01:05:48.829 --> 01:05:51.929 guys can't get chrome plating correct. What makes 01:05:51.929 --> 01:05:55.170 you think that autonomy will be solved? Like, 01:05:55.210 --> 01:06:00.239 come on. This is like a technology. Look, chrome 01:06:00.239 --> 01:06:03.139 plating's been around since what, the early 1900s? 01:06:03.380 --> 01:06:07.280 Like, this is... I'm going to walk everywhere. 01:06:08.340 --> 01:06:11.239 Hey, if we mention Piggly Wiggly... I'll walk 01:06:11.239 --> 01:06:14.519 to a Piggly Wiggly. Yeah, and owners on this 01:06:14.519 --> 01:06:17.000 one, you know, good luck getting a remedy anytime 01:06:17.000 --> 01:06:19.340 soon. It looks like that Ford is not going to 01:06:19.340 --> 01:06:22.420 have anything ready on this until late January 01:06:22.420 --> 01:06:26.980 of 2027. With that, ladies and gentlemen, thank 01:06:26.980 --> 01:06:29.579 you for listening. I hope this has been educational 01:06:29.579 --> 01:06:34.099 or entertaining. And go to autosafety .org and 01:06:34.099 --> 01:06:37.679 click on Donate. Until next week, bye -bye. Bye 01:06:37.679 --> 01:06:41.699 -bye. For more information, visit www .autosafety 01:06:41.699 --> 01:06:42.239 .org.