It’ll all be better in the future

EV Hummers, Goodyear tire investigation, Kia/Hyundai’s can’t get insured, cities pushing back against AV’s, recall roundup and the Tao of Fred.

Subscribe using your favorite podcast service:

Transcript

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: Welcome to the hole in the back of the computer.

Fred: Ah, got it. Finally.

Anthony: Wonderful. And now Wonderful. Begin a new episode of Audio Experience. Wait. , center for Auto Safety Experience. We need a better title. Anyway. Hey folks. Thanks for joining us again. Hey, before the show even starts, click like, click subscribe, write a review on iTunes podcast. Do something like that.

Now, the. So let’s start off. There’s a report that comes out once a year and Jalopnik picked it up this week or past week from this group called Anderson Economic Group, which is far as I can tell, is just funded by Toyota in my opinion. We’re basically, they come out once a year and say that EVs are too expensive.

You can’t charge them. It’s way too expensive, gasoline’s cheaper. Their report has a bunch of assumptions in it. Number one being that the average American makes $70,000 a year, and that when you go to charge your EV at some rest, stop somewhere, the time it takes you to charge your losing wages, which is, I come on, this is just silliness here.

But they make a big assumption that you’re charging 40% of your time away from your home, whereas I don’t think that’s how people are using EVs. And they also, had a big thing in there about deadhead miles, which is what long haul truckers deal with. When, they take a load to one destination, they come back the other way, they don’t have a load.

And so this just sounds a lot like a load. The only takeaway I can take from this that I can agree with is that they basically point out that the current public charging infrastructure in the US is not that good outside of the East Coast, west Coast. It’s, I wouldn’t try and travel across the country, but, I can travel from New York to DC If I had an EV and I wouldn’t be concerned about it or New York to Maine even.

It’d be fine. So don’t believe the hype people.

Michael: There’s a I don’t know if there’s any connection between Toyota and those guys, we’ve seen recently that Toyota, that the Toyota had a change at the top, and it appears that, they wanted to make a, as a company, make a bigger push towards EVs.

It’s interesting, we’re still sitting on the sidelines here wondering what EVs are going to mean for vehicle weight and crashes and a lot of other issues. And at the moment we’re not completely sold on the idea that EVs are going to be the answer for for everything in the transportation ecosystem.

But as that article points out, there’s a lot of, there’s still a lot of things up in the air, a lot of issues to consider when you’re looking at gasoline versus battery.

Anthony: Okay.

Fred: a pregnant pause

Anthony: looking at Fred. And I was like, I know he’s got something to say , but he was just stepping back .

Fred: A cynic would say that and of course I’m not a cynic and I know your gentleman are not either, but a cynic might say that this whole idea of EVs is fluff created to avoid the really meaningful discussions about changes in traffic patterns and highway construction that are necessary to really reduce the footprint of pollutants that are entering the country.

We’re never gonna get there by putting everybody in an ev. The only way we’re going to get there is to. Reinforce and rebuild public transportation systems in a way that would allow me living out here in the sticks to, go pick up the bus to go into the big city and get my pig knuckles or, whatever it is I want to do.

That’s just absent now. And we’ve seen this over and over actually, where a competing company will promise that in the future everything will be better with this new technology they’re developing simply to delay sales of their competitors, not so good product. GE did it with jet engines several years ago.

And even though that they never did produce the engineer touting it was successful in depressing the sales of the competing jet engine. I think that we’re going through a period of a similar phenomenon. The real decisions have to be made, expensive decisions have to be made. About rebuilding the infrastructure and making it accessible to everybody to really make a contribution toward reducing the carbon footprint.

And by talking up the EVs, it’s a great way for the government and people in a position to make these decisions, to avoid making the decisions, to say that in the future, everything will be better. So we don’t need to do anything now. We’ve seen it with safety, we’ve seen it with voluntary standards that don’t work.

It’s an old ploy. And I don’t know, a cynic might say that. This is just another example of bait and switch. Of course, I’m not a cynic and I, I know you gentlemen aren’t, but I just thought I’d throw it out there.

Anthony: I think that’s the theme of today’s episode. In the future, things are gonna get better.

Don’t pay attention to now.

Fred: Don’t look up. Don’t look up.

Anthony: Personally my take on the EV is I look at it as a backdoor way to force a massive upgrade to our electoral infrastructure. But I’m not a cynic. I’m a naive optimist. And speaking and then oh,

Michael: sorry. Two weeks ago, and I don’t think we’ve covered this yet cuz we had a guest last week.

We we saw Jennifer Homendy, the chair of the NTSB, come out and say that we need to be concerned about the really heavy electric vehicles like the Hum V and a lot of the pickups that are on the roads and voice, some very well founded concerns there. I would I don’t think that environmental.

Envisioned the Hum V as what’s going to come in and save the world when they were when they were putting their support behind EVs. And so it was a little disappointing to see the president of the United States in a Hummer EV yesterday, or two days ago, whenever that was. They’re missing the plot.

I can forgive him for leaving all those classified documents next to the Corvette in his garage. But I don’t know about the Hungry V thing. ,

Anthony: he’s also a naive optimist. ,

Fred: his job is hurting cats. His job is not to understand things. He is doing the best he can with the tools.

He is got available. I think

Michael: I, I would’ve liked to see him in the Chevy.

Anthony: So a recap for audience listeners. The main problem with the Hummer EV is not that it makes you look like a fool, it’s that it weighs over 9,000 pounds. And as Fred has pointed out many times on this show, there’s a thing called physics.

9,000 pound vehicle takes a long time to stop, which is plenty of time to crush you in your Toyota Corolla. There’s no coming back from that.

Michael: And particularly if you’re a pedestrian ,

Fred: we did a little research this week and showed that instead of buying the Hummer for the same amount of mass, you can go out and buy a really nice sailboat, save some money, and have a lot more fun.

So drive back and forth to the marina in your Toyota Yas and get a nice sailboat and invite us out for a drink Some. ,

Michael: I have a theory, competing theory that if you have a Venn diagram of Hummer owners and sailboat owners, the circles don’t meet in the middle . So I’m not sure if the demographics work out on that one.

Fred: Good point. Good point.

Anthony: They probably don’t, and I’m pretty sure if I put a sale on a Hummer and put it in the river, I don’t think I would get it to. But that could just be me. So speaking of the Humer ev, and I think, I’m not sure if we’ve touched on this, but there’s a good article arguing for right now roads are paid for through gasoline taxes and with EVs, obviously don’t fuel up it with gasoline.

And so one of the things we’ve talked about is charging vehicles based on the weight they have. So basically an annual tax or something like that. And I think this works for fuel vehicles too, because a Hummer gasoline, vehicle weighs a ridiculous amount as well. So charging vehicles for that, because the heavier vehicle is the more wear and tear it puts on roads in the infrastructure.

Is anyone making any real progress on this approach?

Fred: Yeah. Every commercial truck is taxed on that basis.

Anthony: Really? So 18 wheelers they’re paying

Fred: more, they’re paying more than smaller trucks. Yeah. So it’s a standard technique. It’s already developed. People understand how to do it, how to implement it how to collect the taxes.

I’d say let it rip

Anthony: so it’s not done through their fueling filling up at the gas pump,

Fred: They pay at the gas pump too, but there’s an additional road use tax that the heavy trucks have to pay. So I, I think an additional road use tax for heavy Hummers would be a fantastic. .

Michael: I think one of the problems there is that all the policies right now on, on EVs and taxes are giving tax breaks to people for buying them in the first place.

So that’s where all the real movement has been on taxes, and there’s been very little movement towards taxing based on weight. Although I think that could be a, a great idea, particularly for vehicles that are, extremely heavy when you start getting past sedans, wagons into the large SUVs.

Maybe we set a, a set of 500 pounds and above. I’m just throwing out there 6,000 pounds or above, where you start getting an incrementally larger tax, the heavier your passenger vehicle gets.

Anthony: But as Michael pointed out for some of our larger listeners they weigh the vehicle without you.

Michael: Thank God yes, .

Anthony: Okay. Big heavy cars need what? They need. Big, heavy tires and in an ongoing situation with the Center Fraud Safety and Goodyear justice Department probes, Goodyear’s Handling of Recall Tire. The Justice in investigating Goodyear Tire for its handling of a recall tire that has been linked to eight deaths and dozen, dozens of injuries.

This is a good. .

Michael: Yeah, I think, we could probably do a whole episode or write a book on this Goodyear Tire Saga. It’s been going on since the late nineties is when the crashes started happening, maybe the early two thousands, that period. And it was around the same time as the Firestone Tire Ford and Firestone Tire as she was going through Congress.

And we had legislation passed to put in some better regulations for nitsa monitoring defects and better regulations around the recall process, some early warning reporting, and that type of stuff came into place. And there was this massive federal response to this tire defect going on with Ford and Firestone.

And meanwhile, Goodyear had a similar problem. It didn’t, they don’t, they, there were only a, there were a lot less of these tires on the roads, so Goodyear settled these cases and didn’t report the defect to Mitsa. And now, almost 20 years later, we’re seeing finally justice moves slow.

We’re seeing the possibility that Goodyear’s going to face criminal charges. The Justice Department has I believe they started a grand jury basically who is talking with Dave Kurtz, who is the lawyer who’s been at the center of this for a long time. He’s the one who discovered that Goodyear was, knew that these tires weren’t met for the vehicles they were selling them for.

And he had a client that had settled with Goodyear at some point, and then later on he found out that Goodyear was hiding information from a lot of other potential victims who were, who had cases. So he basically made it, he’s a hero here. He basically made it his mission to ensure that, that, that the story came to light.

And it has through, a lot of good reporting and we’ve done some work to try to get the documents that Goodyear, the tests and other documents that Goodyear was hiding out of Arizona court and into the public eye some of those made it into the public eye in a leak from the court to a Gillo reporter a few years ago.

So it’s a very interesting story, but the, it, what it really comes down to is Goodyear has used, in my opinion, fairly, sleazy legal tactics to hide a defect and Hopefully this grand jury is looking into the legal team that Goodyear has been using over the years. On, on, on this specific case with the G 1 59 tires because there were a lot of problems there and consumers were put at risk and NITSA should have been notified of a defect long before.

They had to come in, I believe, last year and tell Goodyear, Hey, you need to recall these. Goodyear said, they’re not really a defect, but okay, we’ll do it. I think last check they had replaced around tires total, so it’s not as though the recall is costing them a lot at this point, which was their ultimate goal to push it many years.

Anthony: So they’re arguing that things will be better in the future.

Fred: That’s, haven’t, yeah, haven’t we heard this story before? I think there was a movie about the tobacco industry a few years ago that laid this out in detail. And so for your documentary filmmakers out there, If you’d like to get filled in on the details, just drop us a note.

We’d be happy to help.

Anthony: Fred will send you his headshot. So are these tires still on the road or consumers at risk still?

Michael: Apparently at least 12 of them were, cuz those have been replaced. But yes they’re, they are still on vehicles. It’s just really hard to track tires because, they’re not registering the state.

There’s really, it’s always been difficult, even in tire recalls to track down where the tires end up.

Anthony: That, that’s interesting. So as a consumer, I buy a, a brand new car, is not registered that, hey, it came with whatever brand tires and make model

Michael: or the car might be, but these were RV tires and they weren’t always sold, with the vehicle when it was manufactured.

RVs are manufactured by, in, in a slightly different way as well than vehicles. So it, the tires could have been added at a certain point in the manufacturer where they aren’t connected with that vehicle’s van. There’s really, I don’t believe there’s a system that tracks tires based on the RV van Plus as replacements, tires.

Obviously they wouldn’t be connected to the vehicle by a event as well. So it’s just very difficult to track

Anthony: tires. Does the manufacturer at least know where their tires are? Like, cuz it’s an expensive purchase? I imagine

Michael: someone they know where they were sold possibly, but I, and they could possibly track down owners that way.

, I’m sure that’s a difficult process, here we’re talking about 20 years, 15 to 20 years after the purchase, so it becomes really difficult.

Anthony: Okay. Speaking of something that’s gonna become difficult for certain people looks like there’s a couple places that will no longer ensure certain models of Kia and Hyundais.

We’ve talked about this problem before where our TikTok fans have been stealing Kia and Hyundais because you can go on the thing called the internet. I wouldn’t go there. It’s a waste of time. And learn how to steal a Kia in a Hyundai just using a USB cable. That’s upset people like Fred cuz back in his day to Hotwire a card required, splicing lines, whereas this is just, anyone can do this.

So it looks like Progressive and State Farm are refusing the right policies in certain cities for older Hyundai and Kia models that have been deemed too easy to steal. Has this ever happened before where an insurance company is just yeah, no, we’re not insuring your cars anymore because the cars are worthless at that point?

Michael: Yeah. I don’t, I’m not aware of this happening before. I leave that possibility up. I certainly don’t know everything, but the it’s certainly unusual. . It’s interesting here, we worked on Hyundai Kia fires that were happening, I believe at last count it was well over 4,000 fires, around 5,000 fires across a small swath of model years.

And we always wondered why insurance companies didn’t refuse coverage on those vehicles because it was pretty clear that there was a very widespread defect that was destroying the cars. So in this case it’s even newer vehicles. I think these are 20 15, 20 16 to 2019. Kias and Hyundais. And these were the ones we talked about on a podcast where we were discussing the Nitsa theft prevention standards and a mobilizers.

And there’s an option for manufacturers. You can either do your, meet the Nitsa parts marketing requirements, which requires you to mark parts with, I think it’s some type of serial number that identifies them. So it prevents them stolen parts from being sold. That’s one way to supposedly prevent thefts.

The other option is the immobilizer, which these Kias and Hyundais did not have. They chose the parts marking. And so that’s what enabled this TikTok hack to to work. And so now without a, some type of really, probably very expensive retrofit to this, to the vehicle’s ignition systems, these cars are gonna be susceptible to theft.

And I believe that the insurance companies involved, I think it’s progressive in State Farm. They’re limiting. They’re not, it’s not a blanket ban on insurance. They say it’s temporary. We’re not sure, what that exactly that means. But it’s only applying to vehicles in certain locations, certain cities or states, where they’re seeing a problematic theft rates.

So that’s, it’s understandable from the insurance perspective but it’s also, unusual that this isn’t something you see every day. It, it signifies that there’s a really significant problem going on. It’s something we’ve, I think we’ve identified in the past that the rise in vehicle thefts and some of the public safety issues that’s created, it appears it’s also creating an insurance headache.

Fred: Crazy. Maybe TikTok was created by Toyota to force people out of buying Kias. What do you think?

Anthony: I think you’re onto something there. I think. . I don’t know what it is, but I think you’re onto it and you should probably have less of it in your diet. I tried in my favorite topic, my favorite recent topic.

The teenage angst of Cruz and Waymo San Francisco. We’ve talked about how Waymo vehicles and cruise, I think it’s mainly cruise vehicles in San Francisco, that they’re just come to an intersection and they’re like, I give up. I don’t wanna live anymore. And I’ve asked numerous times, how does San Francisco allow this to keep happening?

Apparently San Francisco’s been listening to our show and they realized, wait, we look like fools. And they’re not happy with crews and Waymo anymore particularly that they’re blocking their famous trolleys and trying to run over fire hoses amongst many other issues. So is San Francisco, coming to their senses.

Fred: First I’d just say to San Francisco, thank you for listening and also love the bread. Please send some by. Agreed.

Michael: I would say that of, San Francisco has been at the heart of this. They’ve had the ve autonomous vehicles being tested on their streets, I think in greater numbers and longer than any other area.

And they were the first state the first city a few months ago to really come out against some of the issues that they were seeing on their streets. So they may have even beaten us to that one, Anthony, I’m not sure. But the issue, here in is that San Francisco isn’t, doesn’t really have the choice here because they have to go and protest to California’s state autonomous vehicle regulator if they want something to take place on their streets, which is problematic.

And there’s, it’s, it points to some, what could be a pretty serious tensions in other states. Maybe not so much in California, but I, I think we’re seeing that here where San Francisco is not able to regulate the where the avs are operating on their streets. There’s a process where the AV companies go to the state and try to expand their area and there’s, it’s just it’s taken all of the decision making out of the hands of the local authorities who are frankly the closest to these problems and issues that we’re seeing with traffic delays, emergency vehicle problems, fire hoses and such.

Anthony: So in other cities we see that in Washington state where Washington State, they correct me if I’m wrong, it looks like they passed at the state level saying, Hey, we’ll let each municipality set up their own AV rules. And then Seattle came in and said, yeah, our AV rule is no, we don’t want it. No.

Michael: So no they passed at what they, what we call enabling legislation, where you, it allows the municipalities to determine their own priorities when it comes to the technology.

And so what hap what’s happening in Washington state is Seattle in November of last year, what made some requirements to what the state’s requiring. They basically said, you have to do things like submit emergency responder action plans. We need a little, we need better data. We need, we actually need some evidence that, these can operate safe on our streets.

We need data on. Traffic problems and other things. And the AV lobby went directly to, it appears, went directly to the folks in Washington State Capitol and said look at these guys in Seattle. We need to reign them in. So they have now introduced a bill. They introduced a bill last week in the Washington Senate that would preempt cities in Washington state from being able to regulate AVS on their street.

It’s, we’re not really in favor of that. Think it raises some really, but big questions about, who’s the better? Who’s more capable of regulating? Is it the Seattle d o t or the state? D o t? In some states, we’ll see, like New York City. Can you imagine New York City being, having its AV regulated by Albany, or could you imagine the folks in Richmond regulating AVS in Northern Virginia?

There’re gonna be some pretty serious political issues that arise cause of those

Anthony: tensions. A as a New Yorker, sadly, I can, since Albany controls the New York City subway system, , which is insane, but that’s a different story. Okay so this is just starting to progress in Washington. This bill was just introduced, it hasn’t been passed, and I imagine the local.

Folks of Seattle will be like, wait a second. So let’s

Michael: I it’s just, it’s a little disappointing to see, a city who’s trying to do the right thing, get the data, trying to make its decisions based on good data and based on the experience of other cities when they introduce, just a couple of additional requirements to ensure they have the information to ensure their emergency responders are protected that they’re immediately rebuffed by the folks in the state capitol.

That’s just, I don’t think that’s a very good way to go about regulation of this thing. Plus it’s, you. You’re arguably taking the people who are closest to the situation, who are dealing with the frus, some of the frustrations involved, some of the, potential safety issues involved.

You’re taking them out of the equation and making them go through another layer of bureaucracy to get positive things done. We think it’s an inefficient way to go about it as well.

Fred: Based on what we know so far about the legislation, it’s clear that the legislation is not intended to improve safety.

Yeah. It’s very clear that the legislation is intended to vitiate those actions by Seattle, which Seattle believes are in its interest to promote public safety. Uhoh froze on well, a lot of a lot of political activity to, sorry. No,

Anthony: you’re here. You just froze for a split second. . We’re all good.

Fred: Oh, all right.

Good. This is a, perhaps a tragic consequence of long-term political activity by the chambers of Commerce and their allies to make sure that they can manipulate local regulations by preemption At the state level we’ve seen it over and over again in a lot of different issues. Some having to do with gay rights, some having to do with automotive rights.

So this is one manifestation of the larger political movement that hopefully will soon be identified and countered.

Anthony: I think this is a good segue into The towel of Fred,

Adey: you’ve now entered the Dow of

Anthony: Fred because the promise with AVS has been like, Hey, we’re gonna make things safer cuz we take the human out of the equation.

It’ll all be just the magic of computers. And we all know computers are great because, according to Equifax, I was born in the 1980s. I wasn’t, and I have been married since the 1990s. I haven’t been. So computers are infallible as we know. So Fred, I’m gonna try and do the introduction here and you’re gonna correct me where I’m wrong.

I’ve had my coffee, but this is a complex one. So you’re gonna cover an. an AV reliability standard, correct?

Fred: Maybe kinda okay. Away, I just, now I also wanna point out that the fact that AVS are being developed by engineers does not mean that they’re not being developed by people.

There, there are, there is an overlap in those Venn diagrams within humans and technology, but debatable just to clear the record but getting back to stuff that matters. So there was a report put out by nitsa, our friends National Highway Traffic Safety Administration in 2015 titled Critical Reasons for Crashes Investigated in the National Motor Vehicle Crash Causation Survey.

Sure. To be a best seller at some point. But what’s important about this is that it has a statistic in here that says 96% of crashes involve humans as a critical factor. The industry took this and ran with it and said that means that 96% of excuse me, 94% of accidents, collisions, deaths are caused by human beings.

It’s not at all what the report says. It says that humans are a factor in those, but still it present. It provides some interesting data. Cuz what it says is that if you dive into it a little bit, that, and here I’m flipping through my notes, forgive me for that, that 2% of the fatal crashes are caused by events that are associated with the vehicle itself.

Now this is actually a remarkable achievement for automotive design to think that only one out of every 2 billion. passenger miles driven are associated with a, fundamental defect in the device that’s carrying the people. But still what it does is it sets a, an upper limit for how reliable the software and the other AV unique features in the vehicle have to be in order to avoid contributing to its contribution as a critical factor in highway safety deaths.

In other words, in order to do no harm, you need to be at least as reliable as the cars that are out there on the street now that are involved in accidents. So if you par at data a little bit, what you can find is that of the 6% that are not attributed to the humans, 2% are attributed to the vehicles.

Now this gets into something called, reliability resource allocation, which means that all of the reliability of the individual parts contribute to the reliability of the whole. And so if we say the 2% are associated with the vehicle itself, then we’ve gotta say some portion of that is associated with the data processing part.

And some other portion is associated with all of the mechanical parts that we’re used to, tires, axles, parts that break, all that sort of stuff. So I arbitrarily split it in half and said 1% of that reliability budget has gotta be absorbed by the software and data processing system. So if you crunch the numbers a little bit, what that means is that for the AV features, To have critical factors that are no, that are not contributing to a defect in highway safety compared with the current experience as documented by nhtsa.

The AV unique features can be no more occur, no more often than once every, about 5 billion miles. This is the tough standard. Yeah. But now, in order to turn that into terms that are compatible with software engineering, you’ve got to come up with a metric that says, failures per hour, failures per day, failures per millennium, something like that, because that’s what software engineers are used to dealing with.

So I assume that the average speed of a vehicle is 35 miles per hour. You could use any other number here that’s reasonable, but I think that’s a reasonable number. It’s halfway between parking and getting a speeding ticket. The meantime between life critical failures of the entire set of AV or a D a s automated driver system.

Unique components including data processing software and hardware, which is a standard parameter. Characterizing software driven life or mission critical failures of all types can be no higher than about one failure per 143 million hours of operation. That’s a tough standard. Have, are you with me so far?

I

Anthony: think so. So right now, no one’s obviously anywhere near this standard.

Fred: We don’t know. Nobody has ever, nobody’s ever developed an objective standard for how reliable these systems have got to be. Now, if you look at current experience, we know that the AVS are much worse. And, the principle demonstrator of this is all the Tesla accidents.

You can make all kinds of arguments to say that it’s safer. It’s not, it’s fair, satisfied, it’s not. But the fact is, it is a very high bar for software reliability to merely sustain the current unacceptable slaughter rate of 42,000 people per year. So if you want to make the roads safer, you’ve got to go even farther with the software reliability and the data processing reliability than once per 143 million hours of operation.

Now, I

Anthony: not, sorry, before you going, so that, is, that number, is that per driver or is that collectively that 143 million.

Fred: Oh, great question. That is per software installation. . So if you, okay, so right. Each vehicle has got a unique software associated with it. No, two software implementations would be the same for two different vehicles because the vehicles are different.

They’ve got different characteristics. I may be wrong on that, but it’s not an important point. So that’s per software informa. That’s per software installation, software configuration. Now, it’s important to note that if you change the software configuration, like doing an over the air update, you gotta restart the clock.

So unless you’re willing to put it out there and get the data you’re simply never going to know. Now, if you make two assumptions, and for the nerds out there, I’m looking at you, my sister if you make a stochastic and ergo assumption about the occurrence of these faults, , then you can say, okay, we’re gonna spread this out over the entire fleet of vehicles operating.

So let’s say there’s 300 million or I don’t know what I’ll rewind that a little bit. So if you look at the death rate of one death per approximately a hundred million miles you’re looking at something like, oh, I think I got that calculated here.

Anthony: And that one death per a hundred million miles, is that today’s accident rate on the hundred

Fred: that is today’s accident rate. Yeah. Something like 2% of that could be attributed to the vehicle and you’d be in the right ballpark or 1% of that. You would be looking at one failure per billion miles associated with the cars.

Now you can do that if you make those sarcastic and orgo assumptions by saying we’ll look at the whole fleet. So if the whole fleet in the United States is automatic vehicles and nav, I got the same software running in them, then it can say, okay, I’ve got a pretty big base. Now I can go ahead and calculate what the actual reliability rate is based upon the number of crashes that I’m seeing that are actually attributable to the software the software and AV specific defects.

So it’s possible to do that. It’s possible it’ll never happen, but it’s possible to do that. What it does do is presents an objective standard. For what the software engineers need to expect when they test the software. Now, a as Anthony, you’ve been in the software business. There are several ways of accelerating the life testing of software.

You can do stress testing, you can inject faults, you can run it way faster, you can put it in simulations, a lot of things you can do. So the fact that you’re looking at less than one failure per 143 million of operation. Sounds like it’s a, an impossible thing to do, but it’s not. Software engineers can do that, and these kinds of calculations are routinely done and mission critical software implementations in the military and in a commercial environment, for example operating in offshore oil rig or operating a missile system.

These are simply the kinds of calculations that the industry’s quite used to. So it could, it could be done even if you, but but let me back off. The important point here is that yes, you can set an objective standard for how reliable the AV specific operating system in the vehicle has to be in order to not contribute much less, make better, but in order to not contribute to increasing highway hazards.

Does that make sense?

Anthony: That, that absolutely makes sense. Is there anything related standard wise with the airline industry, because I imagine they clearly have a bunch of mission critical software. Do they have a failure rate per hours flown or miles flown?

Fred: Each component would have a target failure rate cuz Okay, so what happens?

in an overall system analysis, as you say I’m going to accept a failure rate of X. One per million operations, one per billion operations, some number, and then what you do is the manager of risk, as you say, I’m going to apportion that reliability down to the different systems. So part of it’s gonna go to the structure that’s holding it in place.

Part of it’s going to go to, and I’ll make up the numbers, the the flight controller and within the flight controller. Now they’ve got their own budget of, some why failures per hour. And they’ll say, okay, I’m gonna allocate that down to the sub-components. The transformer has to be this and such reliability in order to contribute to the overall.

So you can set up this tree and you see my fingers fluttering down on the screen, but you can set up this whole tree in which you distribute. Your risk of budget and your reliability budget out to the lowest level components where you can go ahead and do the test. So the individual tests on the components can be run very rapidly in many cases analytically.

You have to build that up to the top level, and then if you couple that with software validation with some very tough cases, you can make a very strong case that you’ve got the reliability that you need, you’ve demonstrated it empirically and you’re good to go. That’s typically what happens in a in a complex engineered system of any kind that has mission critical or life critical software embedded in it.

Does that make sense?

Anthony: Yeah. I think, Waymo, Cruz, Aurora, Fred’s, just put all of you on notice. Here’s the basic standard you have to meet. , and

Fred: It’s hard for me to believe that nobody else has looked at this data and come to the same conclusion as the, that it jumps right out at you once you take a look at it and get away from the myth that proves 96, 90 4% of the crashes are going to mysteriously go away.

This is a very important number. It’s well established in the industry, and I think, that’s a great starting place.

Anthony: I think other people have probably come across this number, but they’re not, they don’t want to tout it. If I work for Waymo or whatnot, and I saw this, I’d be like, Hey, oh, let’s ,

Fred: oh yeah it is gonna be a pain in the neck for them.

They have to spend money doing that. And the last thing they want to do now is delay their return on capital, because frankly, this stretch kind of thin. But hey, in the interest of public safety, somebody should do that. Now, interestingly, . One of the complaints that, that we’ve often heard is that no third party is qualified to look over the operations of a sophisticated vehicle manufacturer cuz Oh, it’s just too hard.

And we can’t have the kind of third party review that the Center for Auto Safety has long advocated for these avs cuz it’s just too hard and it’s a Wednesday. And, come on guys, give us a break. So it’s interesting to that Jalopnik reported that Volvo Group has agreed to a substantial fine for recall delays and has agreed to Volvo Group will be overseen in part by an independent third party auditor.

Looking into the group’s recalls dating packed July, 2013. Wow. Michael, isn’t that just what we’ve been advocating? .

Michael: We like third party independent auditors. Nitsa in the last few years has typically every time they’ve done a large civil penalty like that one, which is the Volvo truck and bus group, which is somewhat separate from the vehicle manufacturer.

What they were doing was just a wide array of bad reporting of recalls, defects early warning reports and some other things it looks like. And what the agency does is they then appoint an independent monitor. You’ll see in Toccata there was an independent monitor appointed over Chrysler when they had some recall issues a few years back.

Having an independent outside observer. For some of these companies is very important. Typically it happens after they’ve been naughty and been caught and been fined. But in, we think that manufacturers should be at least checked somehow before they deploy a vehicle.

They claim to be safer than a human driver onto the roads. And we talked a little bit about that process and how the insurance companies might play a role. When we spoke with Edge Case research a few weeks back about the industry using UL 4,600, which we discussed with Deb Prints last week, to at least be able to give policy makers and the public confidence that the vehicles are going to perform.

To the rate that has been set by humans, which is a one death, every one point something, 1.1 something million a hundred million miles traveled, which is a very high bar when you think about it. And it almost lends, it, it makes me think sometimes that humans are a lot better drivers than we give them credit for sometimes.

And then these machines may have a lot further to go than we think, or that we, than we’ve fought for the past decade, for sure.

Fred: Now, before we send it, dial back to its temple in the woods for another week, I have a quiz for Michael and Anthony. Oh, no. Yes. So continued use of wishful thinking instead of hard data to design in needed and document the safety critical reliability in any engineered product, including avs is if not negligent at best.

And question is what? Fill in the blank Tesla’s motto. , Michael.

Michael: It’s not negligent at best.

Fred: Irresponsible faith-based engineering . You guys

Anthony: knew that. Like I said, Tesla’s motto, they’re marketing. Anyway,

Fred: and that’s it.

Anthony: That was a good one. I like to point out though, cuz this gets lost when you hear these very large numbers.

One, one fatality per 1.1 over a hundred million miles. That sounds oh, the roads are very safe. But as we’ve pointed out many times, and Fred pointed out just a few minutes ago, that’s 42,000 deaths per year on US roads. That’s a pretty big number. That, that’s the number that we’re regularly fighting to reduce.

And we’ve pointed out a number of ways in this podcast to bring that number

Fred: down. Yeah. And let me reemphasize that, that number that metric I came up with for software reliability. Doesn’t make the roads any safer. It only enact, it would only enact a standard that says do no harm. Which I think is the minimum that any manufacturer should be able to provide us customers.

Anthony: Okay. But, so speaking of fancy software in cars, consumer reports came out and they did their rankings of a s systems. Automated Driving assistance systems. God, I should know this. So think of your adaptive cruise control and lane keeping software in your car if you have that, where you can set it and your car will more or less maintain the center lane and will stay a certain distance behind the car in front of you.

Consumer and what, basically what Tesla calls full self-driving. So Consumer Reports did a test of everybody’s system. Not almost everybody’s systems. And they came out and they liked Ford’s Blue Cruise. The best they describe it this is my favorite line from their article is, but while Blue Cruise’s capabilities are impressive and can make driving more relaxing, cars that can truly and safely drive themselves remain a long way.

These are just ways to make your life on a long road trip a little easier, but you’ve gotta be the one driving. Okay. And it’s neat about Blocker as a system where they have an internal facing camera that is eye tracking to make sure you’re actually paying attention to the road. It only works on mapped divided highways.

So they’ve got, and I think maybe they’ve expanded beyond that, but they’ve gone out there and mapped all these roads. They know basically what should be going on. And in the Blue Cruise case, something I don’t understand is if you stop paying attention, it will more the driver a number of times, and then it will eventually slow the vehicle down to five miles per hour and turn the hazards on.

But it doesn’t actually stop it or pull it off the road. Like what? I, that’s the, that’s one of the things I don’t happen. I don’t understand. Cuz if the driver has some sort of a heart attack or something like that, I think it’s great they’re taking this preemptive movement, but. I get a little

Fred: lost there.

You might think there should be a requirement that keeps people safe in case the system fail, but of course, that’s visionary, but, still seems like a good idea.

Anthony: They’re apparently doing better than others who aren’t doing that well, and Tesla was pretty far down this list in terms of their what they call full self-driving.

Consumer reports didn’t seem to be impressed with them at all. But,

Michael: I think it’s really good that Consumer Reports is doing this because it’s a real, it’s a confusing area for owners and potential car buyers, trying to figure out what the capabilities of all these different systems are.

So it’s really good that Consumer Reports is putting out ratings on these systems. To help consumers figure out, what this technology is because it’s not, it’s, these cars aren’t as simple as the ones we bought 20 years ago. So it’s, that’s important. The Blue Cruise issue, that is interesting that they don’t have some type of requirement that the vehicle stop, with its hazards on safe, in a safe place.

I, I’m not really sure what happens if the vehicle’s continuing to operate at five miles per hour and reaches a the end of the road or some other obstruction. So that’s odd. But you know what, they came out on top in the CR rating. So they’ve got the they’ve got a fairly impressive system.

It looks like,

Fred: And it’s better than Tesla’s system, which apparently in the case of a distracted driver, looks for flashing emergency lights to approach

Anthony: And accelerates as fast as it can. Time for recall Roundup? I think. Strap

Adey: in time for the recall roundup.

Anthony: So this week we’ve got a few of them.

Cause we haven’t done recall Roundup in a little while, Volvo potentially recalling 27,457 vehicles. This is a, oh boy, boy, this is a long one. It goes everywhere. So a reduction in brake support functions can in increase the risk of crash, lack of robustness, and software compatibility with a specific V C M two hardware version.

The div deviation may not occur during an ongoing brick event. If the deviation occurs, the system will always enter hydraulic. I don’t understand. Let’s just go into the English explanation. It got a little too much there. Over 27,000 vehicles. Sounds like the software issue with their brakes, Michael.

Michael: Yeah. There, there’s a break control module that apparently has some problem with some hardware that’s being install on it. Fred’s gonna translate that first, I believe.

Fred: So let me jump in here a little bit. I don’t understand this at all because it it revolves around this deviation may not occur during an ongoing break event.

I just don’t understand that. It doesn’t mean that it can possibly occur or it’s impossible to occur or somebody’s passed a law saying it shouldn’t occur. Or what the hell does that statement mean anyway? Oh,

Anthony: good, because I thought it was just me. Cause I was like, what?

Fred: I don’t, I just don’t get that at all.

Putting that aside, they talk about a specific problem, which is software incompatibility that can have many sources. You can have an API compatibility , something. Program interface. What is that? I can’t remember if that’s answer. Application Program interface. Application programs. Interface. So software, for example, the software needs to know that the pressure, the software needs to know the pressure and is being sent to stock.

Quote. Okay. So that’s a, that’s an iden, that’s an identifies a potential incompatibility. So the format, the software needs a formal invitation and is instead being sent a yellow sticky note. Now that doesn’t, that sounds a little glib, but, and usually a message is sent across the network that includes a digital word, length, the message structure of protocol, memory requirements a lot of things.

If any of those aren’t exactly right, you’re gonna have an incompatibility. . And you can also have a register in compatibility. For example, if you send a 16 bid message, but the computers is expecting a 64 bid message. This happens that’s brought down rockets in the past. A timing issue, let’s say that as data is needed within x milliseconds, but is set after y, which is greater than next milliseconds.

So that would be like if you, that would be like if you get, receive a invitation to a birthday party a week after the party happened. Doesn’t do you a lot of good thought that counts. How would issue, that’s hard to explain, but it’s like getting off on the wrong foot. And then there of course is the stupido effect, which is sometimes the software author is simply incompetent.

So all of those things can happen in this case. There was probably a sensor that was looking at the voltage of a particular wire, and it seems that it wasn’t getting the correct reading from that wire, which should have read zero volts because it was looking for a ground signal. It looks like it wasn’t getting that zero volt signal.

Okay. I could be wrong in this cuz I haven’t looked at the details, but that’s my guess based upon what we see. Is that enough for you, Michael? Or you need

Michael: more? Yeah, that’s, I think basically what Volvo owners need to know here too is that, you know this, it looks like you maintain breaking ability, but you lose your traction control, your electronic stability, control, your analog brake systems.

And all of the brake technology basically that’s been developed and installed on cars for the past 30 years. So this is something you want to get fixed as soon as possible, and in fact, it looks like it may be fixed through an over the air update. So you can just sit and sit on your couch and it’ll then let your recall perform itself.

Anthony: Ah, so for listeners, this is 2023 Volvo, S 60 s V 60 cc, V 90 cc XC 60 XC 90 s, XC 40 C 40. Who comes up with these names? This is the worst marketing department ever. Anyway, next, recall. Hey, my favorite subject, motorcycle related. Let’s see, Kamodo holding. Is recalling over 16,000 helmets because their helmets failed to comply during d o t testing, penetration test failure.

So basically they sold a bunch of motorcycle helmets that are just for fashion. I guess how does this happen? And they seem to be blaming the test lab instead of themselves,

Michael: right? They the I rarely see part five 70 threes. That’s the recall notice come in where manufacturers use phrases like we’re surprised.

An extremely surprised. By the non-compliance. But that’s what they’re claiming. They’re super surprised and they’re saying, we sent these to Tesla Labs in the United States and China and they passed with flying colors. So why are they failing the Nitza test? Here it just looks like they basically didn’t do enough testing or didn’t do proper testing to ensure they passed the penetration test.

And this is one more helmet that Anthony, you won’t have to worry about your son wearing .

Anthony: Good. Yeah. And so for listeners the brand names of these are built helmets are built. Vertex, b i l t if you have these, please replace them cuz they’re not doing you any good. And let’s see. Last one we have is a su.

They need some bigger trucks. Their bigger trucks needs testing. So this is, they’ve got a non-comp compliment, compliance determination. Ooh, this is another 5 73. Is this 5 73? Is this the the magic recalls?

Michael: The five 70 threes are the magic number for what the recall is. Basically nhtsa Nitsa has a regulation, a part 5 73 regulation that tells you what you have to put in your recall notice.

Anthony: So this looks like a recalibration is required for both the a s camera and a b s module. It’s always these

Michael: cameras. What they did was they built these diesel vehicles and what, these are commercial trucks, but they extended the length of the vehicles. You could have , you could have the frames extended when you ordered the vehicles from, I think it was from about 176 inches to an option of either 200 or 212 inches.

And they did that. But when they did, so they never recalibrated their advanced driver assistance systems to take account for the increase in distance and the different placement of sensors and other things. Which you know, is as far as recalls go, that seems like a pretty simple thing to have discovered and it took them six months.

So they were a little slow on this one.

Anthony: I don’t know if you’ve ever heard of the Boeing se, Boeing 7 37 max, cuz that sounds like the same sort of problem. Hey, let’s replace critical parts, not test them. .

Fred: The difference is that you could detect the source of the problem with the truck by taking out a tape measure.

It’s a little bit more difficult with a 7 37. . Fair

Anthony: enough. And our final item of the week, it’s not a recall, but 50 years ago is when Nitsa had the put in place the odometer disclosure requirement. And now I assume, cuz I haven’t read this, is that means everyone has to share their odometer reading?

Is that right? Or is that I can, yeah, that was odometer,

Michael: that was basically the start of regulations that prevented people from claiming their car, didn’t have the as much mileage as it did so that you’d buy it. That has played a very very strong role in preventing potential car buy used car buyers from getting scammed.

And we’ve still, still to this day, we’ve seen people who get in trouble for rolling back speedometers and doing other things. And what this what that provision did was it basically provided a record keeping function so that odometers could be tracked. So that people couldn’t falsify their odometer reading and pretend their car had a hundred thousand less miles on it than it does

Anthony: because now it’s all software based, it seems and I tried to crack open my windshield, the little screen and roll that, but it’s just a digital.

Michael: that was a bad choice. That many you’re talking about killing your car’s value while trying to raise it .

Anthony: I just get a little sharper pen and just, cross out the two or something. Anyway, folks, thank you for listening. I don’t know if you’ve noticed, but online we’ve been making some changes to the podcast where there’s, you can now add comments.

You can join a little discussion no one has so far, but you can go ahead and do that and start saying, Hey, this is what I love about this. This is what I hate about this. You guys are the best. You guys are the worst. We have transcripts up. Now granted, these are machine generated transcripts, so do not use these for attribution because they get words wrong.

Like they can’t spell Waymo correctly. Now sure, we could, spend a bunch of money and time going through this and making them perfect and correct, but then the roads and cars won’t be safer. That’s where we’d rather spend time. So these are pretty good updates. We’re back on the Instagram.

I know you love and yeah. Hey, thanks for listening. But please go out ra rate this, go to iTunes or however you get this and give it five stars and say how wonderful you find this podcast. Thank you for listening.

Michael: Thanks very much. Thank you. Thank you.

Anthony: Bye-bye.

Fred: Okay, so now I’m gonna do a little final rant. You may wanna splice in. I’m not sure. But people talk about avs and how wonderful they are, and specifically they say they never sleep and they never drink. They’re gonna be great. They’re gonna be a big improvement. I want people to note, The reason you don’t kill people when you drive a car is not because you have certain technical capabilities.

The reason you don’t kill people is because you have a conscience and the car does not. You have judgment and the car does not the car cannot appreciate life-changing consequences of their bad decisions. You as a driver can, the cars do not project the consequences of their limitations, not to unforeseen circumstances you do.

They’re very good at doing that. They don’t fix themselves. They can’t project the consequences of unrelated events onto the driving task. For example, if a human has a prescription for. drugs with some impairment and a recommendation that you don’t drive or operate heavy machinery, you’re likely to pay attention to that even though the information came from nowhere within the driving universe.

Exactly. And fundamentally they don’t care. And you do. That’s the big difference.

Michael: Cars don’t care ,

Anthony: but we do

Fred: end of rant .

 

Join the discussion