Self driving risk conditions
GM Cruise is the gift that keeps on giving to this podcast. Cruise and their CEO Kyle have repeatedly claimed that their cars are so much safer than humans. How much safer you ask? Well their new software will recognize a firehose. My 3 year old nephew can do the same thing but the DMV won’t let him drive a car. Also in this episode Michael gives Anthony some driving advice, NHTSA investigates Cruise for being as good as a bad human driver when it comes to pedestrians, Hawaii cuts down on catalytic convertor theft, Fred explains what a minimal risk condition is and hopefully the autonomous vehicle industry can establish a standard for one and we discuss some recalls of Fords and Tesla’s.
This weeks links:
Subscribe using your favorite podcast service:
note: this is a machine generated transcript and may not be completely accurate. This is provided for convience and should not be used for attribution.
Anthony: You’re listening to Their Auto Be a Law. The Center for Auto Safety Podcast with Executive Director Michael Brooks, Chief Engineer Fred Perkins, and hosted by me, Anthony Cimino. For over 50 years, the Center for Auto Safety has worked to make cars safer.
Fred: Good morning, world.
Anthony: Good morning, everybody. Good evening. Good afternoon. Good traffic driving time. Speaking of traffic, and this is going to be related to a topic we’re talking about. So I did a lot of driving the last few days. And I kept having people who wanted to hang out in my blind spot.
Three way highway, I’m in the middle lane, I want to get over to the right. And somebody in the right, they either want to speed past me and race me that way. I don’t understand why. There’s plenty of room to go on the other side. But I had two separate drivers who’d hang out in my blind spot and I’m like, what are you doing?
So I try to speed up to get in front of them and they’d speed up with me Then eventually I would just take my foot off the accelerator drift down to god knows what and they’d get past Both of them staring at their phones texting not driving a tesla
Michael: maybe they’re just fans, Anthony, or, I think the speed up tactic is probably not the way to go.
Just hit, just coast down, stay on your brakes and get behind people, especially if they’re on their phones. And that
Anthony: I didn’t know they’re on their phones until I until I took my foot off the gas and just let them coast past. Yeah they zoomed past. That, and there’s a message for the state of Connecticut and residents there.
The speed limit is not a suggestion, okay? The speed limit in the United States is 55 miles per hour, everyone is doing a minimum of 80. What is wrong with you people? I want to live. Hey,
Michael: freedom. Freedom. I don’t think that’s just
Connecticut, home of most insurance companies. But hey let’s look, let’s, that’s not what this podcast is about.
It’s not about my personal life. This podcast is about Cruise. Ha. And the silly things they do. Cruise as frequent listeners of the show will know, I’m a big fan. They updated their RoboTaxi software to stop them from crashing into emergency vehicles. I think that’s great.
Fred: Wait a minute, we saw an advertisement a couple of weeks ago that said it never sleeps, it’s never drunk, it never loses attention, it’s all good, and the world is perfectly fine.
How can they possibly improve it if it’s already perfect?
Anthony: That is an excellent question. But so they tout the things they’ve improved and updated. Hey, our software update will now recognize fire hoses. And to me, that’s like telling somebody, Hey, I just changed my diaper my diet, and I shit my pants less.
Like why are you bragging about, hey, we’re really bad at something. Dangerously bad. But now we think we might have a fix for it. I thought this
Fred: whole, Thing was a laughable admission that things haven’t gone. You know if you read it carefully in the first prior Paragraph it says quote the data demonstrate that autonomous vehicles have the potential close quote So basically what they’re admitting is there’s no data that says they’re currently any good And there’s an aspirational statement that in the future, everything will be better.
Where do you find data on potential?
Michael: That’s what I’m wondering. I know a lot of sports guys that would like that.
Anthony: Yeah, I think my bookie has data on potential. Yeah. So there’s news articles on this relating to, but also we’re going to link to the cruise blog themselves, which has such morsels. With over 5 million miles of driverless experience, our AVs have demonstrated their ability to enhance road safety versus status quo human driving.
Michael: mean, you see the usual canard. It’s in the second sentence. They’re already reminding us of how many humans die every year because of all our terrible driving, which is what they based their entire vision on. And it’s. I don’t know, after a while of seeing that touted over and over again, it makes you start wondering why they’re always pointing to something else to begin their conversations.
Anthony: That’s, that is their MO. It’s hey, we may have driven into a fire truck, hey, we may have parked our vehicle on top of a woman’s leg, but we have 5 million miles, which sounds impressive, and as frequent listeners to the show will know, it’s less than a rounding error. 5 million miles.
We need a hundred million miles for real data, but they’re like, Hey, we’re humans can’t drive. Oh, by the way, we’re so good. We got better at recognizing things that humans recognize easily without even thinking.
Michael: In another point here, it’s just something we’ve probably gone over in the past, but at the point where you make all of these changes to the vehicle software, to those 5 million miles, you’ve driven to that point, even really counting more,
Anthony: that is an excellent point.
I don’t know, Fred, what do you think?
Fred: That’s a great point. And not only that, but has this, the upgrades that you’ve made been compared to the operational data from the past as part of your validation process to make sure that, in fact, you are safer in any circumstance or in all circumstances that, they’ve got a really rich database and whenever I see a software upgrade or change has been rolled out very quickly, it’s, I’ve gotta believe that they haven’t checked
Anthony: it very carefully.
You’re a cynic. You don’t understand. Look, if it’s on a computer, it must be correct.
Michael: This is also just hauntingly familiar, this type of approach. And Cruise does it a lot more effectively, I think, than Tesla. But it’s a similar process of saying, Oh, there was a problem. We fixed it.
Trust us and none of us had any way of verifying whether or not anything, either cruise or Tesla, or really anyone who’s doing a software patch tells us so it’s, there’s a bridge there. I think that needs to be created. There’s a gap there. I should say between the public understanding and public confidence and public trust about these issues and what we’re being provided.
And, I think until that is. Yeah. That bridge is created somehow by companies that are pursuing these types of technologies and you’re going to continue to see Consumers be skeptical about the process.
Anthony: Yeah, I think GM Cruise is like Lucy holding the football and we’re Charlie Brown as the public and hey She’s gonna let us kick it this time and we go for it.
They’re like, nope And then they blame somebody else. It’s a strange situation. One of their other things they talked about in their blog posts, they start off one paragraph with, one paragraph with, although incidences are rare, but hey, as we’ve pointed out, what, do they interfere with emergency vehicles?
What is it, 145, 000 times greater than humans do? What was the matter with that? Yeah, I remember like that, yeah. This is how annoying this is. Fred’s just bored with this conversation already. He’s it’s just more GM Cruise bullshit. Yeah, they’re, he’s taking a nap almost.
Fred: I don’t understand their management.
How can they let this go on? They’re just digging themselves into a deeper and deeper hole.
Michael: Until the point at which all the cars save us from ourselves.
Anthony: Maybe that’s it. Yeah. That could be it. They’re still down to what half their fleet size in San Francisco and San Francisco is still trying to get them off the roads entirely.
And then there’s a little known Organization agency called NHTSA and they came out with a report that suggests cruise vehicles are encroaching on pedestrians present in or entering roadways, it’s a weird word, roadways, including pedestrian crosswalks in the proximity of the intended travel path of the vehicles and the issue could increase the risk of a collision with a pedestrian which may result in severe injury or death.
Basically cruise doesn’t care about people if you’re not in the car and we’re going to hit you. Come on, move along. We’re on the clock.
Michael: This is NHTSA opening an investigation. I would have to guess, based on timing, that is somewhat triggered by the incident we talked about a few weeks back where the pedestrian was struck by a non cruise vehicle, a Nissan, tossed in the path of the cruise vehicle and was trapped in the cruise vehicle.
There were, I believe, two or three other Cases that were cited by NHTSA to came up in SGO reports that were summarily deleted from NHTSA site. We haven’t been able to locate them again, but it was the, I believe both of those were from San Francisco. So it’s. It’s NHTSA investigating another autonomous system that, what I, in this situation, in a lot of the Tesla investigations, Tesla has come back and said, Oh, we fixed that with a software update and NHTSA has said okay I wonder if in this case, they’re going to, verify if they’re going to establish that.
New parameters for the operation of cruise. They do have the authority to do that sort of thing. It’s interesting to see what’s going to happen with this, but, it’s the 2nd investigation that’s been opened by the agency into cruise. And the 1st, I believe, was resolved with a software update of some sort.
We’re talking about 300 vehicles, so it’s a very small population compared to what they’re used to dealing with there. But it’s a very important population because, autonomous vehicles, where we think that the focus of the agency administer, from enforcement perspective is going to be for the next few years as these things are deployed.
And so seeing them respond quickly to things like this is important. I’m glad they’ve opened the investigation, but, as usual, the proof is in the pudding. So we’ll see what happens here.
Anthony: So soon enough, we’re going to see a blog post from Cruise saying we’ve updated our software and we can now recognize most humans and try not to hit them.
Lastly, on one of their updates, they said, Hey, we’ve made it easier for emergency responders to get in there and move our vehicles when they cause problems at a scene. One, why are you making these guys and gals work harder in an emergency? Human driver or safety driver, they can move the damn thing.
But two, how do they move a cruise origin that’s got no wheels, it’s got no steering wheel, it’s got no accelerator pedals, it’s got no brake pedals. They’re just gonna, hey, let’s let the fire go, let’s get the biggest people here and we’ll just pick this thing up and move it. They didn’t give a lot of detail on that.
I don’t know what their thoughts are. Maybe it’s a leash. Ooh, maybe it’s a leash. Hey, Kyle, come on, tell us how you feel about leashes.
Fred: It could be, but easier is not easy. There’s a long way between easier and actually getting the job done.
Anthony: Let’s, continue along this thread of cars and software, and hey, we just released an update that made it better.
Kia! We’ve talked about Kia Hyundai quite a bit, and, how easy it is to borrow one temporarily and leave it parked outside my window for hours idling. So they released a software patch for this, and then Kia’s Hey you know that software patch release? We need to patch that software.
Patch. In a link we have to autoevolution. com. KIA started sending out letters to a bunch of its customers to inform them that an updated anti theft software upgrade is now available at no cost. So kind. The reason is as simple as possible. The original patch needs a patch! As the updates did not resolve the glitch as expected.
That’s the dog didn’t quite eat my homework. I, it’s, I think they, Kia might be using the same PR company as GM Cruise. I’m not really sure, but.
Michael: This is just, this whole situation has been disappointing. We’ve known that there were some issues with the software patch. We’ve speculated that it was rushed out in response to the insurance industries coming out a number of insurance industry players and saying they weren’t going to cover these vehicles anymore.
Something wasn’t done earlier this year and. The software patch came out and we’ve fielded complaints from consumers who said I’ve got the software patch and it’s not my car stolen. And also consumers who had vehicles that couldn’t be patched. We were anticipating an update at some point, but, this suggests that.
As we suspected, there were some issues with the original patch, and now everyone has to basically do it again at least in these certain models of vehicles. We’re not sure how many vehicles this applies to. We don’t know which years and models it applies to because it’s not a recall and there is no public which vehicles are going to have to go back in for an update.
At this point. Thank you. As of October 12th, I believe we, we saw a letter that was a petition to NHTSA yesterday that is documented, I believe, around 42 deaths and over 100 crashes. Over 100 injuries, 44 deaths, 132 crashes, and 126 injuries directly related to Hyundai and Kia thefts. So that’s a number that’s big.
It’s continuing to grow. This really isn’t stopping. It’s apparently going wild on Instagram now in addition to TikTok. The hack allows for these vehicles to be stolen and NHTSA is still sitting there. Saying apparently that they don’t have any authority to do a recall here or to jump in and make sure the software updates are done right, or that they reach all the owners that they need to reach.
So we’re ultimately in a very similar situation here as we are with GM Cruise you’re being saying, trust us, this 1 is going to be right. This 1 is going to fix things and then, we may see it roll out, the last 6 months, I think it rolled out to about 20 percent of their fleet since it came out, which is just not good enough to stop these steps in the 1st place.
So there needs to be a new coordinated effort between the D. O. T. Hyundai and Kia to get a good software patch out there, fix this problem once and for all. And, from this perspective, develop regulations going forward to address this situation and. The, the tension that’s here between cyber security and vehicle theft, if you can’t protect vehicles from this type of thing, what, how are we going to protect the vehicles of the future from more advanced thefts and attacks?
Anthony: I’m going to put out something a little controversial. I know I’m getting a little nervous here. Michael’s I don’t know what is he going to say this time? Software is hard. Software is hard to do. And we look at it, Kia, they’re a big company, bad software that came out there. GM, Cruise, they’re ostensibly an auto manufacturer.
They’re upgrading their software. Tesla, same thing. You look at Apple and Microsoft, who are software companies. We’ve all gotten an OS update, which in a day or two later, they’re like, Oh, no, bad one. Don’t install that. Here’s the correct one. So if you have very large companies that are focused on strictly software, they can’t get it right.
Why would we think that some autonomous vehicle company, run by somebody named Kyle, would get any of this right? If my computer at home dies, that’s really a bad inconvenience. If the software in my RoboTaxi has a glitch, people can die. And without any regulations, this is just going to happen more and more.
Because software is hard. And unlike Fred’s, what he says, and he believes, everything will not be alright in the future. No, I’m just kidding. Of course it will.
Fred: Hey, there’s another problem. We talked about this before. Excuse me. But every simulation is doomed to succeed, and by that I mean that when you’re an engineer doing a simulation, you keep banging on your computer until you get the simulation to work and satisfy whatever objective it is.
The problem is, simulations are not the real world. Simulations are digital, the real world is analog. Whenever you make a digital… simulation of the real world, you are creating an abstraction, which does not have anything like the same degrees of freedom as the real world. So that’s a fundamental problem associated with the software that’s being developed for these automated systems, and in fact, any automated driving system.
It’s really hard to do because it takes literally billions of miles on the road to establish adequate Driving experience to really say that these things work the way they’re supposed to work starting with a simulation is a good way to start, but the simulations should be used to identify critical situations that can be used to validate the software and then test the hell out of it to make sure that the.
Vehicle is actually doing what you expect it to do in that situation. That doesn’t appear to be what’s happening. It appears to be a really slapdash procedure where they’re just throwing the software out and as Anthony said, creating patches over patches and moving that along as quickly as possible.
Due to the pressure of their economic realities, they’ve got to start showing some revenue for these things. It’s a tough situation, and it’s different than when you’re sitting at your desk wondering if Microsoft Word is going to correct that comma that’s in your statement. That how to format, how to work out, how to format your document.
It’s different when it’s driving a car.
Anthony: I’m, so you’re under the impression that perhaps GM spending two billion dollars a year on GMCruise might… Might cause them to rush things a little bit. Is that what I’m getting at?
Fred: Yes, and it’s not nearly enough.
Yeah. They’ve got to start working on real cars and real situations and real critical traffic situations that they can replicate on the proving ground.
If they invested in the proving ground, they should use it. They shouldn’t use San Francisco as a proving ground for unqualified and unverified software. Sorry I’m a proud Luddite, but that’s, that’s just the way things should be.
Anthony: No, I’m very happy that you clarified that the Proving Ground should not be a city with actual humans in it.
Agree, a hundred percent, because unfortunately, that’s what Cruise and Waymo have managed to get away with. But, it’s because there’s weak laws and continuing with maybe laws could do something good in the world. What do you think? Do you think we had proper legislation that might reduce crime, might make things better?
The state of Hawaii has shown that it does because Hawaii passed a law. In 2022, classifying a theft of a catalytic converter and correct me if I’m wrong, that’s when you can buy like a GM Chevette and you get the catalytic converter and now it’s a Cadillac, right? That’s what a catalytic converter is?
Michael: That’s why they’re so valuable. Yeah,
Anthony: okay, good. Yeah. The theft of a catalytic converter is a Class C felony. And another law that came in at the start of 2023 that clamped down on the sale of the parts and their components. Under the new law, people in Hawaii looking to sell catalytic converters must now show photo ID your Costco card might work, sign a form saying the converter wasn’t stolen, and meet a few other essential requirements.
Which is great. And so it turns out this is dramatically reduced the theft of catalytic converters in the state of Hawaii. I have another theory is, there’s only a limited number of cars in the state of Hawaii because you can’t drive out of state and they’ve stolen all of them already.
They’ve all been stolen.
Michael: I think the data showed that, what they’re doing there is working. And the article was pointing to the fact that most. Of the states, the lower, the Eastern 48 don’t have these laws. They don’t have laws that restrict the sale unless you have ID or something to prove, essentially who you are and that type of restriction.
Basically eliminates the chance that we’re going to see, the type of massive crime rings that have sprung up around this business. I think it was in Philadelphia. There was a crime ring in the North Northeast that was busted about six, seven months ago, I think. And they. Have, seen their rates of catalytic converter theft dropped dramatically almost by half since that unit was eliminated.
So there’s, between the enforcement, making stronger laws around the theft of catalytic converters, but also, restricting the sale and the transfer of them between parties who are probably a little shady. That’s one way to really address this and this has been a consumer headache across America for years now.
Anthony: Now is why would you buy a used catalytic converter? Because I, I know what one is. I know what it does, but it’s like the gallbladder of your car. Yeah, I got one. It does something, but I
Michael: don’t even know that people are using them after they’re stolen, but I think there’s platinum or other rare materials in them that can be sold as well.
Fred: So there’s multiple things that can happen after 1 store. Yeah, that’s exactly right. It’s it is does have a significant amount of platinum in it. And so that can be taken out and just sold as, as a pure metal, but they’re also sold in the aftermarket or people who had their guess what catalytic converters stolen.
Anthony: Okay, so there’s a whole bunch of, this is, I’m sorry, I shouldn’t be laughing if this is happening to you, but okay, so you get your catalytic converter stolen, that sucks, and then you’re going to buy one, but you’re basically just buying back your old catalytic converter?
Fred: Could be, a new one costs a thousand bucks, and a stolen one costs you a couple hundred bucks, so what are you going to do?
Anthony: But the stolen one’s missing all the platinum.
Fred: Maybe. And you may not care because once your catalytic converter is stolen, remember it’s part of the exhaust system. So you start the car and all of a sudden, it sounds like a Mac truck. It’s just got this because you’ve got a hole in the exhaust system.
So it would be a big improvement. Particularly in those states that do not have any Killed pipe emission testing to just get the appropriately sized piece of pipe in there whether or not it’s actively working where it’s supposed to, just to keep the noise down. A lot of possibilities to sell those, resell those stolen catalytic converters, I think.
Anthony: Are any other states saying, hey, you know what, Hawaii is a fun place to visit, and I like this idea, let’s take their law. Copy, paste.
Michael: I think that’s one of the points the article made was that states have been slow to pick up on that type of legislation. Not sure why. It makes sense. I guess it’s, something that independent parts folks, it’s one another layer of bureaucracy.
I guess they think they have to wait through. So maybe they’re opposed to that type of thing. But, I don’t when there’s a giant theft crisis going on with catalytic converters. I don’t see any reason why states shouldn’t step in those specific areas to, virtually eliminate theft in the way that Hawaii has shown it can do.
Fred: Yeah, a few years ago, there was an airbag theft ring and that seems to have calmed down. How do people counter that, Michael?
Michael: They had laws, and particularly a law was passed in New York to counter that. And when, New York or California… Getting cahoots on those types of things.
It makes it such a huge dent in the market for it that a lot of people leave the business anyway. So it doesn’t take every state. Sometimes it only takes a few. I don’t think Hawaii is going to have an outsized impact on what’s happening on shore here in the United States. But, this type of law could really be something that I think we should bring bring over the bring over from Hawaii.
Anthony: That and more sunshine. Hey, if you want more sunshine, I mean I do too, but instead I really just want you to donate to the Center for Auto Safety. Go to autosafety. org and click that donate button. And if you don’t do it, I’m just gonna, it’s gonna be, this show’s just gonna be an hour long of me saying that over and over again.
And saying, welcome to the Pledge Drive. Up next, we have a young boy who can juggle dolphin shells. It’s unbelievable. Then followed by that, it’s more of a macrame with things in your pantry. I have no idea what that is. I was just reenacting as a kid, you’re like, I want to watch Sesame Street, and then you’re like, Oh my god, PBS is doing a pledge drive.
Boo! Socialism’s bad. Wait, what? Sorry, I got off track there. So don’t steal catalytic converters. And Deya, if you’re a car rental company, don’t allow people to rent cars that have recalls on them. Cuz,
Michael: Zip That one’s, that’s obvious. And that’s Zipcar. And it’s a… Entered a consent agreement with them and gave them a civil penalty this week.
Now, the civil penalty was, for a company with revenue of over 100 million dollars a year was 300, 000 dollars with half of it deferrable. I’m not sure if it’s enough to really teach them a lesson. And this was an investigation that’s had open for, it’s 5 years now. I’m not sure why it took quite that long to get done.
But. In any event, it’s the 1st time that NHTSA has enforced this provision that was passed a number of years ago to prevent the sale of recalled vehicles and. It’s the rental of recalled vehicles. And there’s also a NHTSA investigation ongoing, I believe, into another rental car company for a similar practice.
And it’s something we received complaints on occasionally, a consumer get a vehicle and look it up on the NHTSA recall finder and find out that it has an open recall, which is a big no, we’d encourage everyone and all of our listeners to do that. When you pick up a rental car is to scan the VIN real quick and see if there are any recalls on it and let us know.
If so, and we can report them NHTSA and get more fines, hopefully a lot larger, just like this.
Anthony: Yeah, because in this case it was one of the NHTSA’s employees said they’d rented a van from Zipcar, only discovered the vehicle had an open recall. Yeah, so I think besides Michael and em You know, employees at NHTSA, yeah, it’d be great if everyone scanned the VIN card because, a couple extra more minutes hanging out in a car rental agency is what everyone wants to do while you sit there and go, wait, do I need liability?
Do I need collision? Wait, don’t already have that? I don’t know. I’m putting this on a credit card. Let’s just see what happens.
Michael: What, I wonder what kind of negotiating ability that would give you if you discovered that your car was under recall when you were picking it up from the rental office.
Anthony: But why would they even have that? Because, recall repairs are free. It’s just because Oh, we don’t have enough transit vans.
Michael: When, I’ve heard complaints from the rental car industry that suggest, they get these huge, busy times and their cars are rented back to back.
They don’t have time to get them in for the recall fix. That’s generally, there’s some type of I think the last time I heard that was during COVID, when they had reduced their fleet significantly, and there was a huge demand for the rental cars that were still available. So that’s 1 excuse, but the fact is a lot of recalls now could be fixed by a software.
Some of these companies are getting the software fixed from the manufacturers and able to deploy it to their fleet. So they can fix recalls faster than consumers can even get the software recall from their dealer. On the mechanical recalls, yes, that’s going to cause some downtime and your rental fleet.
That’s something you probably could have calculated on the front end because every fleet is going to be have vehicles that are out of service for some period of time for safety recalls and maintenance. So there’s really no excuse there. I think generally, They were just allowing vehicles, under recall that have a small chance, but still a significant chance of being defective and hurting someone or killing someone.
They just didn’t find that it was, it made sense for them and it from a business perspective to keep a tight lid on that. And after the law is passed, I’m hoping that they, besides it now being illegal that they are understanding that they need to patrol their. The VINs in their fleet, to make sure that there aren’t recalls being sent out on the road.
Anthony: Yeah, that should be an easy fix because they have all that information, you know stored in a database somewhere easily
Michael: Yeah, it’s easy to check and fleets actually rental fleets can access some of the car facts and that’s the tools around recalls and bulk check there Vins and their system daily, hourly, however quickly they’d like to.
So it’s not an overwhelming burden on them to figure out whether or not a vehicle is recalled. Really it’s just a matter of getting it to a facility for a repair that, that compromises most of the time that a vehicle wouldn’t be available to rent.
Anthony: Hey we rarely talk about on the show that I want to spend some time talking about.
It’s been a long time. Autonomous vehicles. Yeah, I can’t get enough of that. I’m getting back to it. Okay. J. D. Power and MIT. I’ve heard of one of those. One of them gives out little awards. They put out a white paper saying that people are turning against autonomous vehicles that basically the public is ah, they’re they’ve had two years in a row of declining their impression of them.
And some of it is. The quote unquote negative news cycle, but, eh, they glance over the fact that the negative news cycle’s caused by them parking their cars on a woman’s leg and driving into fire trucks and things of that nature, but it’s it’s fascinating because I imagined, I think, before we started this show, I would have imagined that, one, I’d never have to touch the steering wheel again by now two, everyone would love an autonomous vehicle and three, I would be taller.
But none of those things have come true.
Fred: No, but Anthony, it’s, it is not a coincidence that confidence in AVs has been declining ever since we started this podcast.
Anthony: Ah, sorry. I was thinking of something else. But, was this surprising to the two or why you to
Michael: I don’t know. The most interesting thing about this to me, or that I think it was around 55%, 54 percent of the responders thought that autonomous fully automated self driving vehicles.
Are either ready for purchase or lease now or will be in the next three to four years, which I think is insane and it I think it really points to The fact that most people don’t understand What fully automated self driving vehicles means? A lot of that those percentages seem to suggest folks who either think that teslas are fully self driving or that the cruise and waymo experiments that are ongoing or somehow Fully automated and are going to be ready to buy which is wild.
But yes, I think that, a lot of the negatives of autonomous vehicles have come to the fore over the last couple of years that the survey covers, particularly in San Francisco and now moving to other places. So it’s probably to be expected there. Yeah. There’s a gap in public understanding.
I think that’s pretty clear as to what these vehicles are. And, what these, the companies that are behind some of the testing that’s going on haven’t really shown those same people how this stuff is going to work and how it’s going to be safe and that type of thing. We’re all still waiting for some really good evidence that.
What their product is going to actually save all these lives that they claim are going to happen someday in the future. And right now we just don’t have that yet. And I’m, it’s, I think it’s beyond time that, you know, instead of repeating their same mantras over and over again about, how their AI is 1 day going to save us from ourselves.
If they could actually show us some basic data that proves this is something that’s going to happen rather than it’s just potential.
Anthony: Could it be this distrust of autonomous vehicles? Could it be because there’s a lack of autonomous vehicle standards? And with that, it’s welcome to the tau of Fred, or as our Transcription software likes to call it the towel of Fred.
Fred: entered the Dow of Fred. Thank you. I’ve I’m waiting a little bit for the round of applause. No, you got it. Good. We’re good? Okay. Good to go. Michael, I have a question for you. What is the SAE standard defining autonomous vehicle automation levels?
Michael: That is S A E J. No, there is no standard.
Did you say standard? I did. No. Why does he get all the good questions? The one that the report that, that talks about the levels is J 3016. But I don’t believe that rises to the level of a standard or even really close.
Fred: Anthony, or Michael, you rose to that, I was offering you a trick question, you saw right through it, the legal mind at work.
That’s correct, that’s, even though it is used as a standard by a lot of people, it is not a standard, it’s a report. So it’s and in the business, what it’s, what people use to describe it is, they say it is not normative. If you have a normative standard, that means that you have to comply. Okay, I’m not,
Anthony: Let’s back up for a second real quick. Just remind listeners, what is SAE and why do I care?
Fred: SAE is the standard acronym used to describe SAE International, which is a standards creating organization in the automation and mobility business. Now, they used to be called society of automotive engineers, but they’ve gone uptown.
So they’ve upgraded to SAE international
Okay. And so auto manufacturers, ideally they, these bunch of folks get together and say, Hey, here’s some standards. We’re going to comply to this and make cars safer, better, more advanced.
Fred: Not just cars. Have you ever bought a garden hose, Anthony? I know you live in the city.
Anthony: this is a private quote.
Fred: I’m going to switch over to Michael. You’re a city dweller. Michael, have you ever bought a
Michael: garden hose? I have. In fact, I’ve bought one recently. And have you ever
Fred: connected it to a faucet? Yes. Yes.
Michael: Did it work?
Fred: It did actually. The reason that worked is because there’s a thing called an SAE thread standard, and the hose and the faucet were both manufactured in conformance to the normative SAE thread standard.
And you can find, whenever you buy a bolt, like a, we’ll say a number 10, 24 bolt. That is also built to SAE Thread standards. And if you go to another company and you buy an SAE 1024 nut, it’s going to fit because they both conform to the standard. So standards are important. Got it. Reports are a lot less important because they’re not normative.
You can make of them whatever you want to make of them. In this case, in a very important case, there’s a lot of wiggle room in this. So I’m gonna, I got another question for you, Anthony. If you pick up a small object, and it has a label on the outside that says chocolate bar with almonds, and you open it, what do you expect to find inside of it?
Anthony: Happiness, for 30 seconds. I was
Fred: thinking in a more chemical
Anthony: sense. Oh, a chocolate bar with almonds? Wait, don’t you exactly right? Yeah,
Fred: you win on that one. Okay. So there’s a critical parameter in the report Michael correctly identified, which is often mistaken as a standard called the minimal risk condition.
So minimal risk condition is a Basically thought of as a safe harbor for the vehicle to maneuver to whenever there’s any problem in the self driving system. Something craps out, the computer craps out, the vehicle transitions into a minimal risk condition. That’s what’s written in.
Anthony: And right now the Cruise in Waymo’s, their minimal risk condition is, they just stop in the middle of the street.
Apparently, yeah. Okay, just checking.
Fred: Or run over a fire hose or do something. But the reason, so it sounds trivial, but it’s really important because the term minimal risk condition is being written into legislation in a lot of states. And often without attribution, even to the SAE J3016, just people expect it to be exactly what it says, right?
So what does minimum mean? It means the minimum. And what does risk mean? Anthony, what do you think risk means?
Anthony: It was a board game by Parker Brothers, which I never played because I didn’t, no one would play it with me.
Fred: Michael, what do you think risk means?
Michael: Risk is the chance of something bad happening.
Fred: Ah, close. So risk is actually a two parameter, and in business, risk is actually a two parameter data, I guess you would call it. And on the vertical axis, typically, you have the probability of occurrence. And on the horizontal axis, you have the consequence. So you’ve got both some risk and consequence in it.
Okay, so if you’re in the upper right hand corner of that, you’ve got a real problem because you have a high risk or you’ve got a very probable occurrence with a lot of consequence. If you’re in the lower left, low risk, low probability of happening. Low probability, low consequence, you’re in good shape.
Unfortunately, the minimal risk condition never defines the word risk. So risk means whatever you want it to mean when you’re developing the car. Minimal means minimal, right? It means the lowest. Unfortunately, if you go to the definition, the minimal risk condition only refers to reducing the risk of a crash.
Okay, so think of a… A convertible car that’s going over a cliff, ala Thelma and Louise with the doors open. If you close the doors before you go over the cliff, you have technically satisfied the requirement to be in a minimal risk condition. It doesn’t address the underlying fact that you’re hurtling over a cliff, you’ve achieved M R C.
Anthony: No way. That’s how this is. That’s insane. My wife’s
Michael: favorite movie.
It’s based on once you’re off the cliff into the air, right? Yeah, you’ve reduced the risk as you go over the cliff.
Anthony: If I took my seatbelt off, I’d increase the risk.
Michael: Ideally, you would, your minimal risk condition calculation would have been made.
20 seconds earlier when you were still on the road.
Fred: Yeah, ideally, I’m just saying it means whatever you want it to mean, because it only says you’re going to reduce the risk. Okay. And so again, you’ve got the problem of risk versus hazard. So when you’re trying to reduce the risk, does that mean.
That you’re making it less likely to occur, or does that mean you’re reducing the consequence? If you had said hazard, it would be a lot clearer as to what you’re talking about here. Then there’s the problem with the term condition. Okay what’s an example of a condition? A condition would be like it’s 20 degrees outside.
That’s cold, right? Or a condition would be something is wet or something is hot. That’s a condition. A condition is not a procedure. And when you look at the definition of minimal risk condition, it says A condition to which a user or an ADS may bring a vehicle, so you have to make a decision right as part of that.
Are you going to do it or you’re not going to do it? The definition here allows you to optionally do it. Also optionally not do it. So you can do nothing and it can still be in conformance with this definition of minimal risk condition. Okay, so that’s a procedure. It’s not a condition. So the problem with the definition of minimal risk condition is that a, it’s not minimal, b, There’s no definition of risk, and what they’re really talking about is the hazard, and C, they’re talking about a condition versus a procedure, and a procedure is not a condition.
Why should anybody care? Oh, and also it’s voluntary versus normative, as Michael established earlier on. It’s not a standard. It’s not a normative standard. So basically, as far as the consumer is concerned, this gives you the illusion of safety because you said I’m headed towards a minimal risk condition.
They’ve even made up terms like a minimal risk maneuver. I’m using a minimal risk maneuver to get to the minimal risk condition. That’s great, but what does it actually mean? You’re going to find out when you’re in a courtroom somewhere. And when you’re suing a company and The defendant’s lawyer says your honor, we conformed to the minimum risk condition and it says right here that may bring a vehicle.
Yeah, we didn’t do it, but that’s the standard says it may bring a vehicle and the standard says we’re going to reduce the risk is. So is the jury going to be a suit enough to say it’s not a standard. It’s a report, and, it’s a, it’s, it really indemnifies the company by the ambiguity that’s built into this particular term to say, basically, whatever you want it to say.
So I won’t belabor that, but, if you drive to your Piggly Wiggly and you do a lot of things to get there, that is a procedure, right? That’s not a condition. And the same thing holds if you were to say, I’m going to park. the car, make it stationary in the parking lot of a Piggly Wiggly that’s a condition.
Being stationary in a piggly wiggly parking lot. So this distinction is very important and I think it’s completely lost on the people who are using this minimal risk condition, including SAE, including the International Standards Organization, including state governments throughout the country. And by using this term, they’re really opening up the public to a lot of jeopardy.
Anthony: Oh also another good board game by Parker Brothers.
Fred: That’s all I’ve really got to say on that. Is there anything about that you’d like to discuss further?
Anthony: Yes. You’re part of SAE, correct? Yes. Okay. And, obviously, I don’t imagine you’re the only person inside of this organization. And there’s at least seven people in this group, right?
Fred: There are more than one person in this organization, yes.
Anthony: Okay, no, there’s thousands of people, right? Is that about right? At least a thousand? I would, yes, definitely. Okay, so there’s other people who realize, hey, this minimal risk condition, this is nonsense. What do we do about it?
Fred: Oddly enough, until we started this campaign, I don’t think there was.
Anthony: Is it just, people, their eyes glazed over, yeah, this sounds it, and they did exactly what you just explained to us. You just put in your head what it means, and then realize, wait a second, we actually have to define what it means, because my version is closing the doors.
Fred: Once the scales fall from their eyes, then they start to come around.
Surprisingly, not everybody does come around, just because I’m stating a position, but those scales are really firmly attached to a lot of people’s eyes. And… Procedurally, a lot of these, a lot of the work that gets done in developing these standards, particularly the international standards, are really driven hard by the need to conform to their production schedule, right?
So things have got to come out by, June of a certain year. Then there’s a lot of steps that have to take place years before that to conform to that June standard. And one of the things that people are really loathe to do is to go back to the fundaments and the fundamental assumptions associated with the standards and say this falls apart.
It’s natural for people to say this is the way you’ve always done it. What’s wrong with that? Let’s just move
But I imagine the situation we’re in, where we’re just talking about how consumer confidence in autonomous vehicles keeps declining. If somebody said, Hey, we have this and this is what we mean by it.
And it’s not a hard definition. It’s Hey, things failed, we’re pulling over. To the side of the road. Something like that. Whereas instead, GM Cruise is just Nothing failed. Everything’s perfect. Here’s a new software update. That’s part of the issue, why people don’t like these things and people hit them with hammers.
Fred: Yeah, That’s all true, and I guess at a personal level, I don’t have any friends, so I can go out and say this. I don’t piss anybody off, or if I do, it doesn’t really change the constellation of friends in my orbit anytime soon. But it’s, it’s a fight, and we do have some friends in this, we do have some allies, but at the beginning of this it didn’t happen, and procedurally, We still haven’t won the battle because even though there’s a lot happening at different levels, there are other levels of approval before any of these standards get published by an organization.
And there’s a lot of politics that goes on. Sometimes you’re able to make the case. Sometimes you can’t. Sometimes things go on. That you can control and sometimes things go on that you cannot control just like any other political organization. Summarizing people need to be careful in order to build the trust and autonomous vehicles.
I think it’s fundamental to 1st, have a set of standards, controls and documentation that is both transparent and accurate and will adequately protect the public’s interest in. Personal safety, I don’t think we’re there yet. And I don’t want to say that this is an intended result of any particular organizations activities, but it certainly seems to be a result of that activity.
And. I think it’s important to remove the ambiguity and straighten this out.
Anthony: So I want to point out to listeners, this is this is how the sausage is made kind of thing. This is why we need your support. This is not incredibly sexy work. This is not incredibly fast work. It takes a long period of time, but autonomous vehicles are coming in some form or another.
And a big thing the center product safety is doing is what Fred’s just talking about is being. Ahead of the curve and getting these organizations and these companies to really, hey, define minimal risk condition basically translate to, hey, let’s make sure we when this stuff goes out large scale that it’s, not going to run you over.
It’s not going to cause any damage. Hey, if you’re ever thinking, why do I want to give this to, to this organization? It’s because of that reason, and because Fred has said the phrase piggly wiggly three times, and he’s very happy about that.
Fred: I do love that, but yeah, you said earlier software is hard politics is very hard.
And as a lunch bucket engineer, I’m not particularly well qualified to do that, so I can only rely on the logic. It’s up to people like Michael, who’s really smart about this stuff, to work the politics. Michael, beat me up. What are we doing wrong here?
Michael: I don’t think we’re doing anything wrong in this circumstance.
There’s a lack of consensus standards on autonomous vehicles and on basically everything that, the is calling level 0 to 6. There’s still a lot of issues in that level 2 to level 3 area. We’ve talked about some of the vehicles that are. Really operating as level threes, which is very limited.
It’s just Mercedes and Nevada and California. And then there’s a whole host of vehicles that are operating in what they want to call level two plus, but it’s actually, there’s a lot of level three similarities that they’re, what they’re really doing is trying to keep the liability on the consumer. They don’t want the software being liable for any bad decisions that it makes at this point.
Fred: So there’s I’m glad you brought that up. Important that you did bring that up because people have to understand that when a company says it’s level two or level two plus or level three, there is no. Governmental organization or even an industry organization that says, yeah, that’s right. So whatever the company says is what it is.
And there’s no such thing as a level 2 plus definition within the report that describes that. As Michael said, level 3, for example. Says the sustained and as defined as conditional driving automation, the sustained and ODD specific performance by an ADS of the entire DDT with the expectation that the DDT fallback ready user is receptive to ADS issued request to intervene as well as to DDT performance relevant system failures and other vehicle systems and will respond appropriately.
That sounds pretty good, but there’s nothing in there about safety critical failures that what duration qualifies as sustained. The report is silent on that. It says use is receptive to ADS issue to request to intervene. What the hell does that mean? What does, the report never defines receptive.
We, we didn’t know what it means. When somebody says, Whoa, we’re level three in conformance with J3016. All that means is nothing. It’s whatever they want it to mean.
Anthony: Hey, that’s the line I used to pick up my wife, okay? I said, look, hey, I’m level 3 compliant with J3016. And she’s oh god, why did I make these choices in my life?
Fred: That worked, huh? I thought she was a New Yorker too.
Anthony: Yeah. Okay, so listeners, SAE, if you want to buy a nut and a bolt. Way to go. Autonomous vehicles, they need a, they’re a way to go. I’m repeating myself now. Hey, how about we do some recalls? Cause we’re eating enough people’s time today. Recalls?
They sound great! We’re good at recalls. And so is Ford. Cause Ford does them quite a bit. Bit Ford Explorers 2020 to 2022. We’re talking 238,000 vehicles. They are equipped with a real rear axle bushing, part number, blah, blah, blah, and the following, powertrains, blah, blah, blah. I, blah, blah, blah. ’cause it’s just, too much.
And the the fractured rear axle bolt will allow the rear axle housing to move out of position, resulting in severe noise and vibration. It sounds like it’s going to be worse than just some rear noise and vibration, right? Doesn’t something else bad happen? Oh, that’s a separate recall. No, that’s the same recall I’m talking about.
This is the this is the fire one. No, this is not a
Michael: separate one. Oh, this is a rollaway. So essentially what’s happening here is, and I think if you. Fred might need to help us translate this a little, but it’s got a mounting bolt on the rear axle and how they describe it as power train torque through the driveline causes axle rotation of the pinion angled towards the subframe.
Which exerts a bending force on the rear axle bolt. Peak torque is normally experienced during a launch event. I think a launch event is when you floor it. The word that I’ve come to, cause that’s also, launch event is what’s described as. What the GM Hummer and the Teslas do when they have those super fast modes.
Apparently that is a launcher’s event. I said my Lamborghini launch event. And it’s after numerous peak torque events are experienced, the bolt may suffer fatigue failure. And that’s what makes the problem happen. Just basically you put the car in park and it’s not even in park anymore Because you’ve been flooring it too much is what the recall sounds like so they’re going to replace this With a part that allows you to floor it a little more
Anthony: I like to call it gun it just I hit that!
Alright, Ford’s got another one. What is Ford recalling here? This is a Ford recalling SUVs and trucks that could catch fire. Certain model 2020 to 2022 Explorers, do it a roll away. No, that’s the same thing. Wait, why does it say fire?
Michael: Huh. I don’t know what that was. You
Anthony: can, that’s just the wrong graphic on Ford
Michael: did have a recall this morning on their the mock E vehicles.
Those were, I think they were recalled last year or there was a service campaign on them to fix the battery. It was on these extended range electric Mach E vehicles and they’re having power failures. And NHTSA opened up a query into that. And it looks like just this morning, FORGE submitted a 573 on that.
That recalls the mach E’s with the extended range battery.
Anthony: What is a 573?
Michael: A 573 is a safety recall report that is filed with The D O T when a company decides to conduct a recall.
Anthony: I got it. Okay. Next up Tesla does every year that company, 50, almost 55, 000 vehicles, the 2021 to 2023 model X on certain ones, the vehicle control of the text, the sensor voltage that indicates brake fluid level may not have a sufficient threshold range at low fluid levels.
This may result in noncompliance with. FMVSS135, section 5. 5. 4. Yeah, so basically it says, Hey, our brake fluid level sensor is wrong.
Michael: And that’s non compliant with the motor vehicle safety standards and they have to fix it in this case. All you have to do is upload your latest full self driving beta software release to get it.
Anthony: That’s what they’re saying. Software release 2023. or later corrected the condition by correcting the vehicle controller set threshold range at low but anyone still listening and then the last one we got here is, Cruise. Again, we’ve already touched on Cruise and how NHTSA is investigating them.
This is prompted by, this is my favorite part of this ODI investigation. It says prompted by SGO and social media reports. The Office of Defective Investigation, I need like a superhero soundtrack behind when I say that, has received reports of incidents in which automated driving systems equipped vehicles operated by a cruise may not have exercised appropriate caution around pedestrians.
So this is just a follow up to what we talked about earlier, where’s there’s a fun video, I’ll see if I can find a link to it, where these People, they’re walking across the crosswalk, they’re doing the right thing, they stop at the crosswalk, they look both ways. It’s a couple adults and some children, and they’re walking through and the cruise is just I bet I can get them.
How much for the kids? Five bucks a kid? They just gave them dirty looks, but you’re just giving dirty looks to a robot. Robot didn’t care. Robot kept going.
And that’s okay. I guess that’s not going to end the show because the two of them were just speechless
Fred: there. Oh, I just wanted to go back to one thing you were talking about earlier when you told our listeners that I was getting bored with this. So I’m not. Okay, the statistics from NHTSA are that there is approximately one Vehicle failure that causes a death every 2 billion miles with a B, 2 billion miles, something like that.
Knowing that injuries, the ratio of injuries to deaths is about 10. This means that there is roughly one in every 200 million miles driven injury associated with the operation of the vehicle. Okay. For a standard vehicle. So now we look at the experience in cruise, which has had how many now 77 something incidents.
Of an out of control
Anthony: vehicle. It wasn’t out of control, it just was, it made a different choice.
Fred: Maybe. Some choices are bad, Anthony. But out of control vehicles into 5 million miles driven. And if you run the numbers on that, it turns out that the hazard associated with these cruise vehicles is many thousands of times higher.
than hazard associated with a conventional vehicle based upon NHTSA statistics and NHTSA records. So I
Anthony: just wanted to clarify that point.
Kyle Vaught from Cruise is on the line and he says no. He says he’s a good boy and that his stuff is right. Humans are bad drivers.
Fred: What Kyle’s trying to do is a hard thing, but it’s too bad that he’s doing such a bad job of it.
Anthony: Hey, with that, hey listeners, thanks for tuning in this week. We’ll be back next week and maybe we’ll get to about a half a dozen stories we couldn’t get to this week because we had so much fun and we had some medical minimal risk condition. So till next time, thank you.
Fred: Thank you. Bye-bye.
Thanks for listening.
For more information, visit www.auto safety.org.