WEBVTT

00:00:01.960 --> 00:00:05.540
You're listening to There Auto Be A Law, the

00:00:05.540 --> 00:00:08.240
Center for Auto Safety podcast with Executive

00:00:08.240 --> 00:00:11.720
Director Michael Brooks, Chief Engineer Fred

00:00:11.720 --> 00:00:15.640
Perkins, and hosted by me, Anthony Cimino. For

00:00:15.640 --> 00:00:18.320
over 50 years, the Center for Auto Safety has

00:00:18.320 --> 00:00:24.620
worked to make cars safer. Hey, listeners, today

00:00:24.620 --> 00:00:28.379
is Wednesday, April 8th, and we're here to bring

00:00:28.379 --> 00:00:31.960
you the wonderful world of auto safety. And the

00:00:31.960 --> 00:00:36.259
wonderful parts and the nonsense parts. But I'm

00:00:36.259 --> 00:00:38.159
going to start off with a gaslight part. That's

00:00:38.159 --> 00:00:39.659
right. Jumping right in. I got to tell you, listeners,

00:00:39.759 --> 00:00:42.439
gaslight coming over the nominee each week is

00:00:42.439 --> 00:00:46.799
incredibly hard. It is. I don't like this part

00:00:46.799 --> 00:00:49.560
of the show. Really? Really? Yeah. Sometimes

00:00:49.560 --> 00:00:53.000
it's not. I tend to fly on the side of there's

00:00:53.000 --> 00:00:55.719
so much, you know, bullshit that's echoed around

00:00:55.719 --> 00:00:57.939
in the world of tech and auto is that it's easy

00:00:57.939 --> 00:01:00.320
to find somebody gaslighting the public. Yeah.

00:01:00.500 --> 00:01:03.859
But I'm like, OK, it's Waymo. It's Elon. It's

00:01:03.859 --> 00:01:07.560
ARK Investments. It's just I don't know. So this

00:01:07.560 --> 00:01:10.379
I broke out of my shell a little bit on this

00:01:10.379 --> 00:01:12.480
one. So I'm going to start off with an article

00:01:12.480 --> 00:01:14.359
that I think I only just sent around to you guys

00:01:14.359 --> 00:01:17.609
this morning. from the Washington Post titled,

00:01:17.689 --> 00:01:20.530
DC speed cameras are catching super violators.

00:01:21.049 --> 00:01:23.650
Most have Virginia and Maryland tags. And the

00:01:23.650 --> 00:01:27.989
article talks about how there's basically people

00:01:27.989 --> 00:01:31.469
speeding like well over 100 miles per hour throughout

00:01:31.469 --> 00:01:34.950
DC, but they have a Maryland or a Virginia plate.

00:01:35.150 --> 00:01:38.209
And so these speed cameras catch them, but because

00:01:38.209 --> 00:01:41.790
of laws, the DC government can't actually go

00:01:41.790 --> 00:01:44.269
after these people. There's no way to. Catch

00:01:44.269 --> 00:01:46.530
them or find them. They send out fines, but the

00:01:46.530 --> 00:01:49.709
people are just like, I learned an amazing thing

00:01:49.709 --> 00:01:53.930
in this that apparently one way that I'm going

00:01:53.930 --> 00:01:57.250
to quote paraphrase in the article. One way drivers

00:01:57.250 --> 00:01:59.510
evade responsibility for camera tickets was by

00:01:59.510 --> 00:02:02.090
getting tags from Virginia, which allows non

00:02:02.090 --> 00:02:05.049
-residents to register vehicles there. About

00:02:05.049 --> 00:02:08.310
14 ,000 D .C. residents have Virginia tags. That's

00:02:08.310 --> 00:02:11.689
mind blowing to me. Like what? I mean, freedom,

00:02:11.830 --> 00:02:14.449
I guess. But that's not my gaslight. My gaslight

00:02:14.449 --> 00:02:17.229
is they go after some of these people with these

00:02:17.229 --> 00:02:20.569
massive amounts of speeding tickets. There's

00:02:20.569 --> 00:02:23.430
a woman, Kylie Ann Sullivan of Fredericksburg,

00:02:23.449 --> 00:02:25.849
Virginia. She's failed to pay one hundred ninety

00:02:25.849 --> 00:02:29.590
seven citations. She was caught going one hundred

00:02:29.590 --> 00:02:32.150
and fifty one miles per hour. Oh, my God. Yeah.

00:02:32.229 --> 00:02:35.050
Onto South Capitol Street. So quoting from this

00:02:35.050 --> 00:02:38.750
like a 30 mile an hour zone. I mean, I can't

00:02:38.750 --> 00:02:41.509
imagine how you could go that fast. Whatever.

00:02:42.110 --> 00:02:44.550
So, quoting, she said in a letter to the court

00:02:44.550 --> 00:02:46.669
that her ex -boyfriend was the one behind the

00:02:46.669 --> 00:02:49.469
wheel for more than 98 % of these offenses and

00:02:49.469 --> 00:02:51.330
that she has not driven since he totaled the

00:02:51.330 --> 00:02:53.349
SUV three months after the high -speed drive.

00:02:53.590 --> 00:02:56.009
I would also like to stress that no one was ever

00:02:56.009 --> 00:02:58.689
injured or harmed as a result of any of these

00:02:58.689 --> 00:03:01.729
incidents, she added. Her case is pending. That

00:03:01.729 --> 00:03:04.590
is my gaslight, is this woman here being like,

00:03:04.669 --> 00:03:07.789
hey, it wasn't me. I wasn't the one driving.

00:03:09.000 --> 00:03:10.979
It wasn't my bag of drugs. I was just holding

00:03:10.979 --> 00:03:14.680
it for somebody else. And no one was injured

00:03:14.680 --> 00:03:18.159
because at 151 miles per hour, I couldn't see

00:03:18.159 --> 00:03:22.020
them if I did injure them. I don't know how you

00:03:22.020 --> 00:03:23.740
can do that. That street doesn't have long enough

00:03:23.740 --> 00:03:26.680
straightaway stretch to do that. That's what

00:03:26.680 --> 00:03:28.199
I was really wondering. How does someone get

00:03:28.199 --> 00:03:30.900
up to 150 miles an hour in D .C.? In a Honda.

00:03:32.639 --> 00:03:36.740
It's like, what? Yeah. So there is somebody else

00:03:36.740 --> 00:03:39.479
in the article that make the. the similar claim

00:03:39.479 --> 00:03:42.379
of like, it wasn't me driving. It wasn't, wasn't

00:03:42.379 --> 00:03:44.979
me. It was just, I share the car with other people.

00:03:44.979 --> 00:03:47.639
And this woman, that's the number one excuse.

00:03:47.879 --> 00:03:49.759
That's the number one excuse, right? For camera

00:03:49.759 --> 00:03:52.719
tickets. Because, I mean, I guess the claim is

00:03:52.719 --> 00:03:55.199
they can't place you in the car, but they can

00:03:55.199 --> 00:03:59.199
place your car. And so, you know. Why not blame

00:03:59.199 --> 00:04:01.780
someone else? It reminds me of a friend when

00:04:01.780 --> 00:04:03.340
she got out of law school. She was doing legal

00:04:03.340 --> 00:04:05.939
aid work and she said it was down in Atlanta,

00:04:06.120 --> 00:04:07.719
I think. And she said half the time she'd get

00:04:07.719 --> 00:04:10.319
these people in and they'd be like, look, it

00:04:10.319 --> 00:04:12.159
wasn't my bag of drugs. And she's like, all right,

00:04:12.180 --> 00:04:13.560
we're going to get rid of that argument because

00:04:13.560 --> 00:04:15.979
that's not going to hold up. No, I was just holding

00:04:15.979 --> 00:04:18.879
it for somebody else. Like, look, your life is

00:04:18.879 --> 00:04:21.540
not the inside of a Stevie Wonder song. OK, let

00:04:21.540 --> 00:04:25.879
it go. So that's my nominee for Gaslight of the

00:04:25.879 --> 00:04:28.959
Week. It is. uh everyone in virginia everyone

00:04:28.959 --> 00:04:32.180
in dc who has a virginia plate and specifically

00:04:32.180 --> 00:04:37.339
this uh woman kylie and sullivan i believe is

00:04:37.339 --> 00:04:41.259
what it was yes that's my that's how i'm starting

00:04:41.259 --> 00:04:44.579
off that's pretty good it sounds like it's kind

00:04:44.579 --> 00:04:47.480
of a local interest story though anthony you're

00:04:47.480 --> 00:04:50.579
usually a much more of a globalist kathy woods

00:04:50.579 --> 00:04:52.779
said some bullshit too what else do you want

00:04:52.779 --> 00:04:59.439
from me All right, so let me show you my gaslight

00:04:59.439 --> 00:05:01.519
this week, which Michael was going to get to

00:05:01.519 --> 00:05:06.019
anyway. But it was Senator Markey asked all of

00:05:06.019 --> 00:05:08.500
the AV manufacturers for some information about

00:05:08.500 --> 00:05:11.920
how they operate and what they do and response

00:05:11.920 --> 00:05:16.899
time in the event of a fault and all that sort

00:05:16.899 --> 00:05:21.379
of stuff. And so all of the companies sent a

00:05:21.379 --> 00:05:25.319
response back to Markey. basically highlighting

00:05:25.319 --> 00:05:28.680
their latency. And there's a misunderstanding

00:05:28.680 --> 00:05:32.259
that latency means the amount of time between

00:05:32.259 --> 00:05:36.579
the AV encountering a problem and the response

00:05:36.579 --> 00:05:40.879
to that problem. So how many companies were there,

00:05:40.899 --> 00:05:43.500
Michael? Six companies? Seven. Seven companies?

00:05:43.920 --> 00:05:47.180
Well, six if you're not counting Tesla, but we

00:05:47.180 --> 00:05:51.639
have to count Tesla. Sure, why not? Only one

00:05:51.639 --> 00:05:57.050
of them answered correctly. The latency is only

00:05:57.050 --> 00:05:59.009
a small part of the actual response time. The

00:05:59.009 --> 00:06:02.029
actual response time includes the latency, which

00:06:02.029 --> 00:06:05.009
is the time between issuance of a digital signal

00:06:05.009 --> 00:06:07.949
on the Internet and receipt of that digital signal

00:06:07.949 --> 00:06:13.370
by the appropriate target. The actual response

00:06:13.370 --> 00:06:16.610
time includes latency in both directions, plus

00:06:16.610 --> 00:06:18.670
the amount of time it takes for the human being

00:06:18.670 --> 00:06:26.019
to attend to the message. Many of these people

00:06:26.019 --> 00:06:29.759
have a tiered system in which the message comes

00:06:29.759 --> 00:06:32.399
in from the AV. A human looks at it and says,

00:06:32.480 --> 00:06:36.139
oh, this has to go to group X or group Y. And

00:06:36.139 --> 00:06:40.899
then group X or group Y assigns that issue or

00:06:40.899 --> 00:06:43.560
that ticket, if you will, to one of the human

00:06:43.560 --> 00:06:46.379
beings in there who then figures out what to

00:06:46.379 --> 00:06:50.720
do, sends it back. And then you've got the latency

00:06:50.720 --> 00:06:55.069
between the... transmission and the receipt.

00:06:55.310 --> 00:06:59.610
So this all takes a fair amount of time, but

00:06:59.610 --> 00:07:03.750
only one company identified to Senator Markey's

00:07:03.750 --> 00:07:06.550
office that latency is in fact only a small part

00:07:06.550 --> 00:07:10.050
of that round trip message return. So that's

00:07:10.050 --> 00:07:16.050
my, my gaslight this week is that six of the

00:07:16.050 --> 00:07:22.709
seven took advantage of the lack of sophistication

00:07:22.709 --> 00:07:24.970
of the person who wrote the request for information

00:07:24.970 --> 00:07:28.730
and misled the senator about how long it takes

00:07:28.730 --> 00:07:34.209
for the AV to respond to some problem that gets

00:07:34.209 --> 00:07:37.750
identified by the AV and then actually take action

00:07:37.750 --> 00:07:40.750
in accordance with that. And I got to give credit

00:07:40.750 --> 00:07:43.490
to my friends at Waymo who actually went out

00:07:43.490 --> 00:07:47.430
of their way to explain to the senator and their

00:07:47.430 --> 00:07:50.399
response. That, in fact, latency is only one

00:07:50.399 --> 00:07:53.199
small part of the overall issue. All of the other

00:07:53.199 --> 00:07:55.500
issues are associated, all the other delays are

00:07:55.500 --> 00:07:58.339
associated with the human factors of sorting

00:07:58.339 --> 00:08:01.040
the command, getting it to the right person,

00:08:01.120 --> 00:08:02.740
figuring out the answer, and then getting it

00:08:02.740 --> 00:08:05.399
back and answering that it takes several seconds

00:08:05.399 --> 00:08:09.579
at a minimum for the AV to respond to whatever

00:08:09.579 --> 00:08:14.899
the issue was. So my plot, it goes to my friends

00:08:14.899 --> 00:08:18.899
at Waymo and the Bullshit Award. Or Gaslight?

00:08:18.959 --> 00:08:22.660
Bullshit. Which Gaslight award? Goes to the other

00:08:22.660 --> 00:08:28.519
six manufacturers. Aurora was one of them. Tesla,

00:08:28.699 --> 00:08:32.559
Zooks. Who else, Michael? May Mobility, Nero.

00:08:33.139 --> 00:08:36.360
Did you say Motional? No. No, I did not say Motional.

00:08:36.820 --> 00:08:40.779
So congratulations to all of those other companies.

00:08:41.179 --> 00:08:43.799
You've done a great job of Gaslighting the Senator.

00:08:44.460 --> 00:08:48.139
Well done. Yeah, I'd say that all of them, yeah,

00:08:48.200 --> 00:08:50.399
all of them gaslighted in one way, since every

00:08:50.399 --> 00:08:54.080
single company refused to disclose how frequently

00:08:54.080 --> 00:08:56.899
these remote operators are intervening, which

00:08:56.899 --> 00:09:00.559
tells us a lot, right? I mean, knowing how often

00:09:00.559 --> 00:09:03.899
a remote operator has to intervene in a vehicle's

00:09:03.899 --> 00:09:06.240
operation really tells us how autonomous those

00:09:06.240 --> 00:09:09.879
vehicles truly are, right? Right. Well, also,

00:09:09.980 --> 00:09:12.860
as Phil Koopman pointed out, why do they have

00:09:12.860 --> 00:09:17.889
to intervene at all? vehicle is solely responsible

00:09:17.889 --> 00:09:21.289
for the safe operation why is there a human in

00:09:21.289 --> 00:09:23.870
the loop at all now what the senator should have

00:09:23.870 --> 00:09:26.529
asked for is how much does how much time does

00:09:26.529 --> 00:09:30.289
it take for an av to respond when there is a

00:09:30.289 --> 00:09:33.210
human in the loop to form part of the driving

00:09:33.210 --> 00:09:36.830
command response right maybe next time yeah even

00:09:36.830 --> 00:09:39.629
just simpler just yeah how long does the actual

00:09:39.629 --> 00:09:43.179
response take And it's going to be a whole lot

00:09:43.179 --> 00:09:44.840
more than the numbers there. They're focused

00:09:44.840 --> 00:09:47.080
on the report, which are, you know, like in milliseconds.

00:09:47.480 --> 00:09:50.320
Right. Michael, before we jump on your gas, can

00:09:50.320 --> 00:09:53.120
we I want to stay with this because. Yeah. So

00:09:53.120 --> 00:09:56.139
Marky, his I want to quote from his press release

00:09:56.139 --> 00:09:58.620
on this says my investigation revealed a wide

00:09:58.620 --> 00:10:00.500
range of concerning practices from employees

00:10:00.500 --> 00:10:03.600
assisting vehicles from overseas to wide variations

00:10:03.600 --> 00:10:05.700
and communication lag times between vehicles

00:10:05.700 --> 00:10:08.059
and human operators. These operations demand

00:10:08.059 --> 00:10:10.000
urgent federal regulation. I'm calling on NHTSA

00:10:10.000 --> 00:10:12.179
to investigate. And I'm working on legislation

00:10:12.179 --> 00:10:15.240
to impose strict guardrails on AV companies used

00:10:15.240 --> 00:10:19.220
to remote operators. Okay, so forgetting the

00:10:19.220 --> 00:10:23.220
side of that, doing this is needed, but in the

00:10:23.220 --> 00:10:25.080
current environment, this is probably a massive

00:10:25.080 --> 00:10:36.879
long shot. I broke myself. There was a question

00:10:36.879 --> 00:10:41.019
in there. Is that coffee shortage still striking

00:10:41.019 --> 00:10:42.980
in New York City? I've been having to wake up

00:10:42.980 --> 00:10:46.080
too early lately. It's really set off my circadian.

00:10:46.279 --> 00:10:49.659
I'm assuming you're going to ask what can be

00:10:49.659 --> 00:10:52.440
done here at the federal level to address the

00:10:52.440 --> 00:10:54.580
problem. Even more of that is like, what are

00:10:54.580 --> 00:10:58.379
the actual steps here for Markey? Yeah, he's

00:10:58.379 --> 00:11:01.899
going to call for this data. As Fred just pointed

00:11:01.899 --> 00:11:04.659
out, at least six out of seven dentists are going

00:11:04.659 --> 00:11:07.850
to lie to him. Or not lie, but obfuscate. And

00:11:07.850 --> 00:11:12.669
not really get to the truth here. This seems

00:11:12.669 --> 00:11:18.649
like a fairly simple thing again. And I want

00:11:18.649 --> 00:11:20.610
to continue from Mark. He says autonomous vehicle

00:11:20.610 --> 00:11:22.450
companies have long boasted they can eliminate

00:11:22.450 --> 00:11:25.769
road fatalities caused by human error. But the

00:11:25.769 --> 00:11:27.970
fact is they're not being honest about how much

00:11:27.970 --> 00:11:29.769
humans are actually involved in this. So you're

00:11:29.769 --> 00:11:32.769
not eliminating human error. You're just shifting

00:11:32.769 --> 00:11:34.789
it from the driver's seat to a call center in

00:11:34.789 --> 00:11:37.649
the Philippines. Yeah. And then trying to make

00:11:37.649 --> 00:11:41.690
that remote operator non -responsible for outcomes,

00:11:41.950 --> 00:11:44.269
too. Let's not forget that they're hoping to

00:11:44.269 --> 00:11:46.590
shield themselves from liability if there are

00:11:46.590 --> 00:11:49.330
mistakes made. Right. So, yeah, that's the thing.

00:11:49.429 --> 00:11:51.990
So there's wouldn't that be the simpler thing

00:11:51.990 --> 00:11:55.289
is is forgetting all of that and just focus on

00:11:55.289 --> 00:11:59.149
enforcing or assigning a liability to this? Because

00:11:59.149 --> 00:12:01.710
we know on the state level that these cars are

00:12:01.710 --> 00:12:04.750
exempt. They can murder at will and no one will

00:12:04.750 --> 00:12:07.460
get caught. He's working on this on the federal

00:12:07.460 --> 00:12:12.039
level campaign. I guess as far as political expediency

00:12:12.039 --> 00:12:14.340
goes, right, it's probably going to be much more

00:12:14.340 --> 00:12:17.840
difficult to get a bill through that assigns

00:12:17.840 --> 00:12:20.419
liability to certain parties under certain conditions

00:12:20.419 --> 00:12:23.539
than it is to maybe get a bill through that makes

00:12:23.539 --> 00:12:26.080
the industry have to be a little more transparent

00:12:26.080 --> 00:12:30.320
about their remote operations. I mean, why? Why?

00:12:30.480 --> 00:12:33.549
Because. I mean, the liability is in one there.

00:12:33.549 --> 00:12:36.809
They're giving up information and scenario two.

00:12:36.889 --> 00:12:41.950
They're giving up money. Right. I get that. But

00:12:41.950 --> 00:12:44.529
it's you know, it's not, you know, assigning

00:12:44.529 --> 00:12:47.750
liability under certain conditions. It's it doesn't

00:12:47.750 --> 00:12:49.950
matter what the conditions are when I'm driving

00:12:49.950 --> 00:12:54.360
my car. I'm always liable for any accident. If

00:12:54.360 --> 00:12:56.460
it's, if, if there's fault attributed to you,

00:12:56.539 --> 00:12:59.019
right. So there has to be some type of a fault

00:12:59.019 --> 00:13:01.059
there. You can't just be driving down the road

00:13:01.059 --> 00:13:03.279
and then you get T -boned by a dump truck and

00:13:03.279 --> 00:13:05.019
you're somehow at fault for that. Right. Well,

00:13:05.059 --> 00:13:06.659
I think in New York state, there's no fault.

00:13:06.700 --> 00:13:09.960
So that's an insurance thing. Oh, but, but you

00:13:09.960 --> 00:13:11.740
know, you know what I'm saying? So is, is if

00:13:11.740 --> 00:13:16.000
I'm yeah. That, I don't know. I just, I'm upset

00:13:16.000 --> 00:13:22.080
about this whole. Just stop. I'm upset about

00:13:22.080 --> 00:13:25.019
the liability aspect. That's it. That's my story.

00:13:26.580 --> 00:13:29.500
Anthony, you're muttering. I know. And another

00:13:29.500 --> 00:13:34.539
thing that came into play in the findings of

00:13:34.539 --> 00:13:36.899
the report was that every company, this could

00:13:36.899 --> 00:13:38.840
have been another gaslight for you guys, every

00:13:38.840 --> 00:13:41.440
company claimed its autonomous driving systems

00:13:41.440 --> 00:13:45.440
would ignore dangerous advice from... remote

00:13:45.440 --> 00:13:49.139
operators. How? I mean, literally, how do you

00:13:49.139 --> 00:13:55.080
program an autonomous vehicle to parse through

00:13:55.080 --> 00:13:58.539
the advice coming from its human remote operator

00:13:58.539 --> 00:14:02.120
and decide whether or not to ignore that advice?

00:14:02.360 --> 00:14:06.220
That sounds just like an absolute lie, right?

00:14:06.360 --> 00:14:09.200
And every company claimed that. It's science

00:14:09.200 --> 00:14:12.879
fiction. Can these vehicles really do that? Because

00:14:12.879 --> 00:14:16.000
if they can do that, then why do they need REOs

00:14:16.000 --> 00:14:19.840
in the first place? It's called magic. Yeah,

00:14:19.919 --> 00:14:23.639
it just, that sounds super sketchy. All accidents

00:14:23.639 --> 00:14:27.820
are edge cases. And so a human operator telling

00:14:27.820 --> 00:14:30.100
them to do something dangerous would be an edge

00:14:30.100 --> 00:14:33.639
case. As we've learned, you can't program around

00:14:33.639 --> 00:14:37.330
every possible edge case. And we know that's

00:14:37.330 --> 00:14:40.769
not true because we saw what happened with the

00:14:40.769 --> 00:14:42.669
Waymo and the school bus in that one instance

00:14:42.669 --> 00:14:47.029
that the NTSB looked at where the operator or

00:14:47.029 --> 00:14:49.809
the remote operator said, oh, no, that school

00:14:49.809 --> 00:14:52.429
bus is not dropping off passengers. You can go

00:14:52.429 --> 00:14:54.570
ahead and pass it. And the Waymo did, right?

00:14:55.169 --> 00:14:58.009
Yeah, well, I think it's even worse, actually,

00:14:58.029 --> 00:15:00.710
because all the AV operators are saying they've

00:15:00.710 --> 00:15:06.919
now gone to end -to -end AI, right? So in order

00:15:06.919 --> 00:15:10.980
for any input to affect the end -to -end AI,

00:15:11.440 --> 00:15:15.320
it's got to have happened many, many times and

00:15:15.320 --> 00:15:18.019
have been absorbed and processed through the

00:15:18.019 --> 00:15:21.059
genetic algorithms, genetic learning algorithms,

00:15:21.220 --> 00:15:23.159
I guess they're called, in the large language

00:15:23.159 --> 00:15:26.820
models. So how the hell does that work? You'd

00:15:26.820 --> 00:15:30.379
have to have a rule -based overlay for the end

00:15:30.379 --> 00:15:33.460
-to -end AI in order for an individual input

00:15:33.460 --> 00:15:37.269
from... a remote source to have any impact whatsoever.

00:15:37.750 --> 00:15:41.110
So they're either not using end -to -end AI or

00:15:41.110 --> 00:15:45.490
they're not able to absorb any information from

00:15:45.490 --> 00:15:50.649
a remote human user. So what the hell? There's

00:15:50.649 --> 00:15:53.190
a real disjoint here. The human -in -a -loop

00:15:53.190 --> 00:15:56.710
factor is completely inconsistent with end -to

00:15:56.710 --> 00:16:02.470
-end AI governing the vehicle motion. I must

00:16:02.470 --> 00:16:05.309
be missing something because... This makes no

00:16:05.309 --> 00:16:08.789
sense to me at all. You just need more stock

00:16:08.789 --> 00:16:12.470
options to help you understand the cognitive

00:16:12.470 --> 00:16:14.129
dissonance. Well, that's true. That's always

00:16:14.129 --> 00:16:17.549
possible. And I'm open to that. Listeners, if

00:16:17.549 --> 00:16:19.789
you want to fund Fred's stock options, go to

00:16:19.789 --> 00:16:22.889
autosafety .org and click on Donate. And with

00:16:22.889 --> 00:16:24.549
that, we'll jump to Michael Brooks' Gaslight.

00:16:26.490 --> 00:16:29.029
Hell, my Gaslight is connected, not necessarily

00:16:29.029 --> 00:16:32.490
an announcement itself, to the DOT's announcements

00:16:32.490 --> 00:16:35.850
this week that... there are as they're claiming

00:16:35.850 --> 00:16:38.909
record low traffic deaths and fatalities in 2025

00:16:38.909 --> 00:16:41.889
these are the preliminary estimates of traffic

00:16:41.889 --> 00:16:44.529
fatalities which are generally pretty close to

00:16:44.529 --> 00:16:47.649
what we see as the end number every year um so

00:16:47.649 --> 00:16:49.809
we'll see how that works out as the year moves

00:16:49.809 --> 00:16:54.049
on they finalize the numbers so They are estimating

00:16:54.049 --> 00:16:58.570
that there were 36 ,640 fatalities in 2025, which

00:16:58.570 --> 00:17:03.529
is a 6 .7 decrease from 2024. That's good. You

00:17:03.529 --> 00:17:07.450
know, and in fact, I believe, and I don't know

00:17:07.450 --> 00:17:09.349
if this number has been completely verified,

00:17:09.490 --> 00:17:13.000
but it looked to be close, if not. Correct. The

00:17:13.000 --> 00:17:15.539
second lowest traffic fatality rate in recorded

00:17:15.539 --> 00:17:20.259
history at 1 .10 fatalities per 100 million vehicle

00:17:20.259 --> 00:17:25.660
miles. 100 million vehicle miles traveled. That's

00:17:25.660 --> 00:17:29.299
good news. And, you know, it's it's you know,

00:17:29.299 --> 00:17:33.720
I think my gaslight stems from some of the political

00:17:33.720 --> 00:17:36.940
crowing here. I mean, what we're looking at here

00:17:36.940 --> 00:17:39.279
is, you know, a preliminary estimate, first of

00:17:39.279 --> 00:17:44.049
all. but beyond that you know this is you know

00:17:44.049 --> 00:17:49.190
a fatality rate at 1 .10 you know looks great

00:17:49.190 --> 00:17:52.930
in comparison to the past five years um starting

00:17:52.930 --> 00:17:56.549
with with 2020 and coven when we saw the range

00:17:56.549 --> 00:18:01.700
um of fatality rates jump up to about, I think

00:18:01.700 --> 00:18:05.359
it was 1 .38. So, there's a significant difference

00:18:05.359 --> 00:18:07.180
between the fatalities occurring on the road

00:18:07.180 --> 00:18:11.339
today versus 2020, 2021, when we saw the worst

00:18:11.339 --> 00:18:16.619
rates that we've seen in a long time. But, you

00:18:16.619 --> 00:18:22.180
know, from 2009 through 2019, that range was

00:18:22.180 --> 00:18:26.640
from 1 .08 to 1 .19 fatalities per 100 million

00:18:26.640 --> 00:18:30.319
miles. So, If you look at this from my perspective,

00:18:30.740 --> 00:18:32.839
you know, it's great news the fatality rate is

00:18:32.839 --> 00:18:36.339
declining over a small sample of years post -pandemic

00:18:36.339 --> 00:18:38.700
now, right? But it's really important to note

00:18:38.700 --> 00:18:41.380
that we haven't made any substantial reductions

00:18:41.380 --> 00:18:43.700
in the traffic fatality rate for more than 15

00:18:43.700 --> 00:18:46.599
years at this point, which if you go back and

00:18:46.599 --> 00:18:50.319
look at the fatality rates over time, I don't

00:18:50.319 --> 00:18:52.680
know that we've had that long of a period of

00:18:52.680 --> 00:18:56.920
stagnation. and the fatality rate kind of staying

00:18:56.920 --> 00:18:59.019
the same place and not improving. And that's

00:18:59.019 --> 00:19:01.619
in a time where, you know, manufacturers are

00:19:01.619 --> 00:19:04.420
promising more safety tech and all this other,

00:19:04.440 --> 00:19:06.859
you know, all this other stuff, you know, more

00:19:06.859 --> 00:19:10.819
than ever. So it really makes you, you know,

00:19:10.819 --> 00:19:14.539
question. a statement like this one from NHTSA

00:19:14.539 --> 00:19:17.200
Administrator Jonathan Morrison. Under President

00:19:17.200 --> 00:19:19.720
Trump and Secretary Duffy, American roads are

00:19:19.720 --> 00:19:22.619
safer. The Trump administration has a strong

00:19:22.619 --> 00:19:24.539
relationship with our law enforcement partners

00:19:24.539 --> 00:19:26.940
responsible for keeping our communities safe.

00:19:27.319 --> 00:19:31.079
NHTSA is doing our part by doubling down on safety

00:19:31.079 --> 00:19:33.599
strategies that reduce risky driving behaviors

00:19:33.599 --> 00:19:38.279
before they cost lives. And somewhat brushing

00:19:38.279 --> 00:19:40.759
over the political crowing there, which obviously

00:19:42.100 --> 00:19:45.059
an administration being in office for a year

00:19:45.059 --> 00:19:48.779
is not going to have a significant impact on

00:19:48.779 --> 00:19:52.279
how people are driving and the number of fatalities

00:19:52.279 --> 00:19:54.559
on the road. So taking credit there is questionable

00:19:54.559 --> 00:19:59.119
at best. But NHTSA is not doubling down on safety

00:19:59.119 --> 00:20:01.799
strategies that reduce risky driving behaviors.

00:20:02.019 --> 00:20:04.640
They're not pursuing, you know, one of the number

00:20:04.640 --> 00:20:06.680
one, their number two and number three things

00:20:06.680 --> 00:20:09.140
they could pursue right now to reduce risky driving

00:20:09.140 --> 00:20:12.130
behaviors. Getting a rule out that establishes

00:20:12.130 --> 00:20:14.730
a plan to get better driver monitoring into vehicles

00:20:14.730 --> 00:20:17.309
to prevent distracted driving. Getting a rule

00:20:17.309 --> 00:20:20.549
out to reduce impaired driving that they're mandated

00:20:20.549 --> 00:20:23.670
by Congress to do and they're late on to get

00:20:23.670 --> 00:20:26.109
technology into vehicles to prevent drunk driving.

00:20:26.430 --> 00:20:29.289
Number three is getting a rule out to require

00:20:29.289 --> 00:20:33.430
speed. prevention technology into vehicles. All

00:20:33.430 --> 00:20:35.950
those things would reduce risky driving behavior.

00:20:36.109 --> 00:20:39.529
You're not doubling down on those. It's just

00:20:39.529 --> 00:20:42.470
a fact. NHTSA is not pursuing those strategies.

00:20:42.789 --> 00:20:45.269
They're effectively ignoring them and pointing

00:20:45.269 --> 00:20:47.670
to law enforcement as a solution here, which

00:20:47.670 --> 00:20:51.630
it's not. So for that, that's my gaslight this

00:20:51.630 --> 00:20:54.930
week. That's good. But don't you think law enforcement

00:20:54.930 --> 00:20:57.410
has helped? I mean, we've seen the grappler.

00:20:57.759 --> 00:21:00.480
Law enforcement is always out there helping.

00:21:00.660 --> 00:21:03.380
I think you have to question whether you can

00:21:03.380 --> 00:21:06.380
rely on law enforcement to completely address

00:21:06.380 --> 00:21:09.079
the problems of drunk driving, speeding, and

00:21:09.079 --> 00:21:12.240
distraction. I don't think you can. I think they're

00:21:12.240 --> 00:21:14.880
always going to be there providing a support

00:21:14.880 --> 00:21:17.220
and some enforcement that's going to deter some

00:21:17.220 --> 00:21:19.980
people, but the people who are the worst speeders,

00:21:20.000 --> 00:21:22.519
the worst drunk drivers, and the worst distracted

00:21:22.519 --> 00:21:27.769
drivers don't care. And until their cars stop

00:21:27.769 --> 00:21:30.009
them from doing those behaviors, they're going

00:21:30.009 --> 00:21:32.869
to continue. You know, it's also worth noting

00:21:32.869 --> 00:21:37.210
that NISA has been doged. They've lost a lot

00:21:37.210 --> 00:21:41.190
of good people. And I frankly don't believe this

00:21:41.190 --> 00:21:44.289
estimate at all, other than some numbers somebody

00:21:44.289 --> 00:21:49.470
pulled out of their ass. It's earlier than it

00:21:49.470 --> 00:21:51.849
normally comes out. How do they do that with

00:21:51.849 --> 00:21:55.099
fewer people? This is the estimate. So it comes

00:21:55.099 --> 00:21:58.839
out about this time every year. And this is all

00:21:58.839 --> 00:22:01.000
based on fatality analysis reporting system.

00:22:01.279 --> 00:22:04.720
So, you know, and, you know, it follows the recent

00:22:04.720 --> 00:22:07.019
trend coming down from COVID. So I don't see

00:22:07.019 --> 00:22:11.039
any evidence of an issue there with the data.

00:22:11.140 --> 00:22:14.559
I mean, you know, perhaps, maybe in some world.

00:22:14.640 --> 00:22:17.059
I know that there's a lot of questions about

00:22:17.059 --> 00:22:19.759
various data operations in the federal government

00:22:19.759 --> 00:22:21.519
right now when you've got a government that.

00:22:22.089 --> 00:22:26.049
you know, seems to just want to create things

00:22:26.049 --> 00:22:28.690
out of whole cloth and then find the data to

00:22:28.690 --> 00:22:31.089
support it later. But I don't know about that

00:22:31.089 --> 00:22:33.609
take on this one because that's just something

00:22:33.609 --> 00:22:36.630
that would, you know, if NHTSA was somehow fudging

00:22:36.630 --> 00:22:41.529
traffic fatality data, that would be a gigantic

00:22:41.529 --> 00:22:46.109
failure on their part. I bet my usual nickel

00:22:46.109 --> 00:22:48.250
on this one. You can't afford it. See how it

00:22:48.250 --> 00:22:52.200
works out. What do you think is the reason for

00:22:52.200 --> 00:22:54.460
the decline here? I mean, not Morrison's whatnot

00:22:54.460 --> 00:22:56.660
saying we're going after unqualified truckers

00:22:56.660 --> 00:22:59.480
and I think we're basically the same place on

00:22:59.480 --> 00:23:01.920
safety where we were 15 years ago. And we're

00:23:01.920 --> 00:23:05.440
just after COVID raised those rates temporarily.

00:23:05.900 --> 00:23:09.539
We're returning to the norm. Oh, OK. Well, because

00:23:09.539 --> 00:23:12.420
we're right where we were, you know, around 2009

00:23:12.420 --> 00:23:16.539
at this point. Oh, yeah. OK. Well, that's good

00:23:16.539 --> 00:23:18.940
to hear. And don't know. It's not good. Well,

00:23:19.019 --> 00:23:21.819
it's that we've come down from the the covid,

00:23:21.920 --> 00:23:24.740
the post covid. Yes. Come down from whatever

00:23:24.740 --> 00:23:27.240
what what the things that were causing the covid

00:23:27.240 --> 00:23:29.619
spike. But I think it's bad news that we don't

00:23:29.619 --> 00:23:32.160
see. Plus, I mean, let's keep in mind here that

00:23:32.160 --> 00:23:35.279
the United States is far, far behind on fatality

00:23:35.279 --> 00:23:38.500
rates in vehicles compared to the rest of the

00:23:38.500 --> 00:23:40.880
developed world. Right. We're not doing the job

00:23:40.880 --> 00:23:44.259
right. And to celebrate, you know, this. kind

00:23:44.259 --> 00:23:50.519
of return to 2009 um doesn't seem to be a very

00:23:50.519 --> 00:23:53.240
you know uh it seems to be kind of tone deaf

00:23:53.240 --> 00:23:55.200
on the part of the administration but you know

00:23:55.200 --> 00:23:57.819
that's part for the course yeah that's not surprising

00:23:57.819 --> 00:24:01.740
at all uh with that i'm going to uh judge our

00:24:01.740 --> 00:24:04.119
gas light here and the winner is going to be

00:24:04.119 --> 00:24:05.980
michael just because i like the color of his

00:24:05.980 --> 00:24:09.940
uh shirt there you go that's how we do it okay

00:24:10.670 --> 00:24:12.230
I'm going to wear this shirt every week. No,

00:24:12.269 --> 00:24:13.769
you're not going to win every week. It's, you

00:24:13.769 --> 00:24:18.670
know, it feels like you need the win. So I'm

00:24:18.670 --> 00:24:20.329
giving it to you, even though, honestly, mine

00:24:20.329 --> 00:24:25.390
was the most fun. So moving on, let's go into

00:24:25.390 --> 00:24:30.150
another government related thing about from NTSB.

00:24:30.829 --> 00:24:36.609
That's right. NTSB, it's an early report that

00:24:36.609 --> 00:24:38.769
basically say, I'm going to quote from this AP

00:24:38.769 --> 00:24:41.150
article. The systems that allow drivers to take

00:24:41.150 --> 00:24:42.950
their hands off the wheel are convenient, but

00:24:42.950 --> 00:24:45.289
don't improve safety because people are often

00:24:45.289 --> 00:24:47.470
too reliant on them and end up paying more attention

00:24:47.470 --> 00:24:49.569
to their cell phones and infotainment screens

00:24:49.569 --> 00:24:52.710
than to the road, the head of the NTSB said Tuesday.

00:24:53.190 --> 00:24:55.190
And this came out of, there was a hearing around

00:24:55.190 --> 00:24:58.869
Ford's Blue Cruise where they had two fatal crashes.

00:24:59.049 --> 00:25:00.950
The hearing was about two fatal crashes. Unfortunately,

00:25:01.369 --> 00:25:04.670
I think they've had a few more than that. Whereas...

00:25:05.000 --> 00:25:07.980
Hey, you're getting eyes off, hands off, the

00:25:07.980 --> 00:25:11.279
car will drive itself. And as we were just talking

00:25:11.279 --> 00:25:14.500
with autonomous vehicles, that's not a thing.

00:25:15.359 --> 00:25:19.140
And this has been the thing sold to the consumers

00:25:19.140 --> 00:25:21.500
saying, hey, I have Blue Cruise or I have Super

00:25:21.500 --> 00:25:26.599
Cruise or I have Muska Cruise, whatever. And

00:25:26.599 --> 00:25:28.240
people stopped paying attention to the road.

00:25:29.059 --> 00:25:33.299
Further down in this article, we have a... Continuing

00:25:33.299 --> 00:25:35.819
again, quoting, NTSB members expressed concerns

00:25:35.819 --> 00:25:37.900
Tuesday that drivers who are under the influence

00:25:37.900 --> 00:25:40.359
of alcohol or drugs seem to believe that using

00:25:40.359 --> 00:25:43.019
a driver assistance system can help them drive

00:25:43.019 --> 00:25:45.259
while impaired. It's obvious to anyone paying

00:25:45.259 --> 00:25:47.440
attention that people are buying this technology

00:25:47.440 --> 00:25:49.920
with the plan of using it to help them drive

00:25:49.920 --> 00:25:54.000
home drunk, Bill Kopman said. Which, I mean,

00:25:54.140 --> 00:25:58.319
brilliant advertising. Hey, look, we know you're

00:25:58.319 --> 00:26:01.039
going to stop at the bar and have a couple. So

00:26:01.039 --> 00:26:04.400
let us drive you home. Some restrictions may

00:26:04.400 --> 00:26:06.400
apply. This is for entertainment purposes only.

00:26:06.519 --> 00:26:08.940
I mean, effectively, both of these, I think it

00:26:08.940 --> 00:26:11.299
helps to look at these two crashes they're looking

00:26:11.299 --> 00:26:14.950
at, the similarities in them. they occurred at

00:26:14.950 --> 00:26:18.890
night they occurred because i guess the initial

00:26:18.890 --> 00:26:21.569
cause of it is you know there were stationary

00:26:21.569 --> 00:26:25.250
vehicles parked in in in the travel way right

00:26:25.250 --> 00:26:29.230
at night time at highway speeds and these these

00:26:29.230 --> 00:26:33.750
ford vehicles came along um using blue cruise

00:26:33.750 --> 00:26:37.390
and were not able to detect those vehicles in

00:26:37.390 --> 00:26:43.630
time to avoid the collisions um you know there's

00:26:43.630 --> 00:26:45.549
a question here you know they talk about driver

00:26:45.549 --> 00:26:47.529
monitoring that they recommend you know better

00:26:47.529 --> 00:26:51.349
driver monitoring systems i think there's a question

00:26:51.349 --> 00:26:53.769
in the in these type of situations as to whether

00:26:53.769 --> 00:26:58.250
driver monitoring you know is is or or whether

00:26:58.250 --> 00:27:01.710
driver monitoring is going to be able to prevent

00:27:01.710 --> 00:27:04.410
these kind of crashes when the vehicle isn't

00:27:04.410 --> 00:27:08.500
even detecting a problem Um, until the last minute,

00:27:08.519 --> 00:27:10.759
seeing these vehicles, does the human involved

00:27:10.759 --> 00:27:13.460
is going to have an opportunity to take over

00:27:13.460 --> 00:27:16.619
in these, you know, bang, bang type crash situations.

00:27:16.759 --> 00:27:19.279
I don't know. Driver monitoring could help the

00:27:19.279 --> 00:27:22.920
human stay, um, focused on the driving task.

00:27:23.079 --> 00:27:25.460
And maybe that human then picks up this car,

00:27:25.660 --> 00:27:27.200
stopped in the middle of the road in front of

00:27:27.200 --> 00:27:29.319
them and avoids a collision. So yes, driver monitoring

00:27:29.319 --> 00:27:31.920
will work there. Um, but these are, you know,

00:27:32.319 --> 00:27:34.700
These are the types of crashes that scare us

00:27:34.700 --> 00:27:39.420
with level two and above systems. Nighttime,

00:27:39.619 --> 00:27:41.200
you know, the Blue Crew system. I'm not sure

00:27:41.200 --> 00:27:43.180
if they have, you know, a LiDAR or if they're

00:27:43.180 --> 00:27:46.299
just using cameras like Tesla, but either way,

00:27:46.480 --> 00:27:48.920
they do. Just cameras. There's no LiDAR. There's

00:27:48.920 --> 00:27:50.740
no LiDAR on Blue Crews? They don't have radar?

00:27:50.900 --> 00:27:54.079
They have radar. Yeah, yeah. A lot of the, even

00:27:54.079 --> 00:27:56.039
Tesla used to have radar before they eliminated

00:27:56.039 --> 00:27:59.539
that in favor of the visual thing. But there's

00:27:59.539 --> 00:28:08.069
a problem here. particularly with that situation,

00:28:08.369 --> 00:28:10.710
you know, that dark situation when we're worried

00:28:10.710 --> 00:28:14.349
that cameras aren't going to work properly. you

00:28:14.349 --> 00:28:17.269
know automatic emergency braking even non -level

00:28:17.269 --> 00:28:20.009
two right uh should be working in these situations

00:28:20.009 --> 00:28:23.670
to prevent crashes so this is an issue that you

00:28:23.670 --> 00:28:26.190
know you know involves a lot of different you

00:28:26.190 --> 00:28:29.069
know human factors considerations is the human

00:28:29.069 --> 00:28:31.250
doing the right thing is the car doing the right

00:28:31.250 --> 00:28:33.569
thing is the car monitoring the human to make

00:28:33.569 --> 00:28:36.369
sure the human's doing the right thing so they

00:28:36.369 --> 00:28:39.130
issued a lot of recommendations here um to the

00:28:39.130 --> 00:28:42.589
dot you know they basically said hey dot you

00:28:42.589 --> 00:28:48.220
know we've you need to integrate And make sure

00:28:48.220 --> 00:28:50.240
a lot of this stuff works. You need to make sure

00:28:50.240 --> 00:28:53.099
automatic emergency braking is working. There

00:28:53.099 --> 00:28:54.740
was some speeding involved in one of these crashes.

00:28:54.900 --> 00:28:56.980
You need to make sure intelligent speed assistance

00:28:56.980 --> 00:28:59.660
is working and driver monitoring is working properly.

00:29:00.660 --> 00:29:03.779
You need to promote better driver engagement

00:29:03.779 --> 00:29:06.039
in the driving task. And you need to prevent

00:29:06.039 --> 00:29:10.119
misuse and automation complacency. And here's

00:29:10.119 --> 00:29:12.200
what I like. You need to set maximum operational

00:29:12.200 --> 00:29:14.880
speeds for automation systems based on their

00:29:14.880 --> 00:29:17.890
capabilities. If your vehicle is not capable

00:29:17.890 --> 00:29:21.430
of detecting a stationary vehicle in the road

00:29:21.430 --> 00:29:24.789
at highway speeds, then why are you allowing

00:29:24.789 --> 00:29:29.589
it to operate at that speed? There's some other

00:29:29.589 --> 00:29:31.589
interesting recommendations around, you know,

00:29:31.589 --> 00:29:34.569
in this case, there were there were some alcohol

00:29:34.569 --> 00:29:39.269
or drug involvement. So NTSB reiterated their

00:29:39.269 --> 00:29:41.849
recommendation that it's a, you know, get on

00:29:41.849 --> 00:29:44.009
the ball when it comes to impairment prevention

00:29:44.009 --> 00:29:47.390
technology. um which is something that you know

00:29:47.390 --> 00:29:49.630
what we just talked about that's that that's

00:29:49.630 --> 00:29:52.130
something that nitza was required to get some

00:29:52.130 --> 00:29:55.869
work done on prior to now and simply hasn't they've

00:29:55.869 --> 00:29:57.750
kind of kicked the can down the road on that

00:29:57.750 --> 00:29:59.569
we've talked about that with the dad systems

00:29:59.569 --> 00:30:02.490
and other things um they also sent some recommendation

00:30:02.490 --> 00:30:07.009
to to ford um and you know one of them was around

00:30:07.009 --> 00:30:10.450
making the driver monitoring better um in this

00:30:10.450 --> 00:30:13.630
situation it looked like they thought that that

00:30:13.630 --> 00:30:17.569
a driver could be looking at their phone and

00:30:17.569 --> 00:30:21.069
the system would interpret it as though the driver

00:30:21.069 --> 00:30:23.750
was engaged, even though the driver clearly wasn't.

00:30:24.069 --> 00:30:28.809
So they told Ford to do that. And probably one

00:30:28.809 --> 00:30:31.190
of the more interesting things in here to me

00:30:31.190 --> 00:30:33.869
is that they've also recommended to the U .S.

00:30:33.869 --> 00:30:39.569
DOT and NHTSA to... require manufacturers who

00:30:39.569 --> 00:30:42.250
are producing vehicles with this data, I mean,

00:30:42.250 --> 00:30:47.630
with these capabilities to collect data that

00:30:47.630 --> 00:30:50.329
would allow those manufacturers to properly report

00:30:50.329 --> 00:30:55.039
problems in these systems. effectively kind of

00:30:55.039 --> 00:30:57.519
holding the rest of the industry to a Tesla standard

00:30:57.519 --> 00:31:00.339
of data collection. One of the reasons that Tesla

00:31:00.339 --> 00:31:03.059
shows up so much more frequently in the standing

00:31:03.059 --> 00:31:05.400
general order data than other manufacturers,

00:31:05.799 --> 00:31:08.279
you know, they obviously have a lot more connected

00:31:08.279 --> 00:31:10.279
vehicles on the road. That's part of it. But

00:31:10.279 --> 00:31:12.240
the other part of it is when a Tesla vehicle

00:31:12.240 --> 00:31:15.900
is in a crash. where the airbags are deployed,

00:31:16.200 --> 00:31:18.099
you know, that's automatically reported back

00:31:18.099 --> 00:31:20.160
to Tesla, which then has a duty to report that

00:31:20.160 --> 00:31:22.900
on to NHTSA as part of the standing general order.

00:31:23.059 --> 00:31:27.640
Not all manufacturers have that. ability to collect

00:31:27.640 --> 00:31:31.500
and send the data back to home base and that

00:31:31.500 --> 00:31:34.039
then obligates them to report it to NHTSA. So

00:31:34.039 --> 00:31:37.240
the NTSB here is recommending that, you know,

00:31:37.240 --> 00:31:39.599
like Anthony said, maybe all these crashes aren't

00:31:39.599 --> 00:31:42.740
being captured. They're not, apparently. And

00:31:42.740 --> 00:31:45.920
something that NHTSA needs to do is to enable

00:31:45.920 --> 00:31:47.700
something that we've always called kind of like

00:31:47.700 --> 00:31:52.609
an advanced. EDR or a, you know, it's an advanced

00:31:52.609 --> 00:31:55.509
black box system. It's a black box system in

00:31:55.509 --> 00:31:59.230
the vehicle that collects. Similar to what Tesla

00:31:59.230 --> 00:32:01.069
is doing, collecting, you know, a million lines

00:32:01.069 --> 00:32:05.609
of data every minute or so, collects a lot of

00:32:05.609 --> 00:32:08.089
ongoing data on the vehicle operation and sends

00:32:08.089 --> 00:32:10.450
it back to the manufacturer so that they can

00:32:10.450 --> 00:32:13.569
then use that for developing safer vehicles and

00:32:13.569 --> 00:32:17.190
also report safety incidents to the federal government

00:32:17.190 --> 00:32:19.230
so that we can continue to make these types of

00:32:19.230 --> 00:32:22.809
systems safer. I like what you said there is

00:32:22.809 --> 00:32:26.799
why. Why sell these products if these vehicles

00:32:26.799 --> 00:32:29.059
can't operate under these conditions? Why allow

00:32:29.059 --> 00:32:31.859
it to go at this speed if it can't handle this?

00:32:31.940 --> 00:32:33.779
And it makes me think of automatic emergency

00:32:33.779 --> 00:32:35.920
braking, whereas every manufacturer is saying,

00:32:36.019 --> 00:32:38.460
we can't get it to work at highway speed. That's

00:32:38.460 --> 00:32:40.799
insane. But hey, we're going to sell you a system

00:32:40.799 --> 00:32:43.740
that runs at highway speed and above, but we

00:32:43.740 --> 00:32:46.720
can't make AEB work. So yeah, that's insane.

00:32:47.759 --> 00:32:50.140
That's very similar to the same nonsense. We're

00:32:50.140 --> 00:32:54.359
constantly pushing back on the AV industry. Well,

00:32:54.359 --> 00:32:56.039
when the AV industry is out there saying we can

00:32:56.039 --> 00:32:58.980
do everything. And then the automobile industry

00:32:58.980 --> 00:33:01.920
is at the same time out of the side of its mouth

00:33:01.920 --> 00:33:05.700
saying we can't, we can't. meet this AEB standard.

00:33:06.000 --> 00:33:08.339
And I'm just sitting there rolling my eyes going,

00:33:08.539 --> 00:33:12.059
how can you break autonomous vehicles appropriately

00:33:12.059 --> 00:33:15.220
and claim they're safer than humans and all that

00:33:15.220 --> 00:33:18.059
nonsense that you continuously trumpet and at

00:33:18.059 --> 00:33:20.960
the same time say that you can't make automatic

00:33:20.960 --> 00:33:24.700
emergency braking that detects bicyclists, motorcyclists,

00:33:24.700 --> 00:33:26.779
and can't operate at higher speeds and blah,

00:33:26.880 --> 00:33:29.079
blah, blah, blah, blah. It doesn't make sense.

00:33:30.480 --> 00:33:33.299
Yeah, from a consumer's point of view, because

00:33:33.299 --> 00:33:35.960
when both GM and Ford, they came out with their

00:33:35.960 --> 00:33:38.140
hands -off systems, their Blue Cruise, Super

00:33:38.140 --> 00:33:40.519
Cruise, I was like, oh, this seems like a better

00:33:40.519 --> 00:33:44.099
approach. They're using HD maps, it's geo -fenced,

00:33:44.119 --> 00:33:46.900
whatnot, but then realizing, oh, we're going

00:33:46.900 --> 00:33:49.079
to let your cars go at 70 miles per hour, but

00:33:49.079 --> 00:33:51.859
we can't brake. I think consumers have to have

00:33:51.859 --> 00:33:55.440
this in mind that... what they're selling to

00:33:55.440 --> 00:33:58.759
you is, here's a loaded gun. Well, I mean, I

00:33:58.759 --> 00:34:00.619
think it would be great. You know, I think that's

00:34:00.619 --> 00:34:02.180
the number one of the number one issues here

00:34:02.180 --> 00:34:05.900
is that some of us are, you know, highly skeptical

00:34:05.900 --> 00:34:10.039
of claims made by corporate America. And some

00:34:10.039 --> 00:34:12.840
people just believe it, right? And they hop in

00:34:12.840 --> 00:34:14.880
these cars and they believe a lot of the hype,

00:34:14.980 --> 00:34:17.679
even if the manufacturers caution them in one

00:34:17.679 --> 00:34:20.599
way or the other. And so they get in trouble.

00:34:21.289 --> 00:34:23.289
Yeah. And I don't blame the consumer at all.

00:34:23.510 --> 00:34:26.050
It's it's this is how it's. No, I do not. Because

00:34:26.050 --> 00:34:28.949
come on, you're a busy day and you're told repeatedly

00:34:28.949 --> 00:34:30.869
these cars drive themselves. We're living in

00:34:30.869 --> 00:34:32.889
the future. Everything is great. People hear

00:34:32.889 --> 00:34:34.889
artificial intelligence. They skip the word artificial

00:34:34.889 --> 00:34:38.429
and just focus on intelligence. And yeah, you're

00:34:38.429 --> 00:34:41.530
a busy person. I don't believe it. I think that's

00:34:41.530 --> 00:34:43.809
where NHTSA. I kind of put a dual responsibility

00:34:43.809 --> 00:34:45.809
there. See, I don't want to blame the drivers.

00:34:46.130 --> 00:34:49.409
I think as consumers, we have a responsibility

00:34:49.409 --> 00:34:53.340
to. to try to operate the vehicle safely and

00:34:53.340 --> 00:34:56.019
apprise ourselves of what we're actually operating,

00:34:56.260 --> 00:35:00.800
right? Yes. There's certainly a duty there. But

00:35:00.800 --> 00:35:03.219
at the same time, I think manufacturers are failing

00:35:03.219 --> 00:35:05.679
in their duty to properly describe these systems

00:35:05.679 --> 00:35:08.719
and their capabilities. And they're so much more

00:35:08.719 --> 00:35:11.000
intent on flooding vehicles with all these convenience

00:35:11.000 --> 00:35:14.260
features and monetizing it, essentially, than

00:35:14.260 --> 00:35:16.480
they are at ensuring that they operate safely.

00:35:16.940 --> 00:35:22.280
Yes. So consumer beware. Uh, yes, that's what

00:35:22.280 --> 00:35:24.840
we're going to go there. Uh, let's jump to a,

00:35:24.940 --> 00:35:28.539
uh, also related to speeding. This is, this is

00:35:28.539 --> 00:35:29.960
interesting. I think it was 10th grade physics

00:35:29.960 --> 00:35:32.699
class. I invented this system and now Cal Colorado

00:35:32.699 --> 00:35:35.000
has released it. Whereas I was driving in the

00:35:35.000 --> 00:35:37.519
New York state thruway and realizing the New

00:35:37.519 --> 00:35:39.199
York state thruway is a, it's, you know, it's

00:35:39.199 --> 00:35:42.019
your toll is based on the distance you go. So,

00:35:42.019 --> 00:35:44.519
and I was like, well, Hey, I know I'm getting

00:35:44.519 --> 00:35:46.820
on in the Harriman entrance and I'm leaving the

00:35:46.820 --> 00:35:49.710
New Paltz one. Speed limit's 55, but I'm doing

00:35:49.710 --> 00:35:53.369
70. I was 16. Leave me alone. And I realized

00:35:53.369 --> 00:35:56.690
it's a set fix. It's a fixed distance. They should

00:35:56.690 --> 00:35:59.289
be able to know that I arrived way too early

00:35:59.289 --> 00:36:02.449
at this point. And so now the state of Colorado,

00:36:02.610 --> 00:36:06.590
they're putting Waze and Google Maps and Apple

00:36:06.590 --> 00:36:09.269
Maps, all these systems that alert you to speed

00:36:09.269 --> 00:36:11.929
cameras up ahead, essentially out of business.

00:36:12.110 --> 00:36:13.869
I'm going to quote from an article from MotorOne

00:36:13.869 --> 00:36:17.519
.com. The state's new automated vehicle identification

00:36:17.519 --> 00:36:20.699
systems uses several cameras to calculate your

00:36:20.699 --> 00:36:23.380
average speed between them. And if it is 10 miles

00:36:23.380 --> 00:36:25.639
per hour or more over the limit, you get a ticket.

00:36:25.880 --> 00:36:28.280
No longer be able to slow down as you approach

00:36:28.280 --> 00:36:30.760
a camera and speed back up after passing it.

00:36:30.840 --> 00:36:32.639
Not that you should be speeding on public roads

00:36:32.639 --> 00:36:35.780
in the first place. This is, I think, a very

00:36:35.780 --> 00:36:40.320
simple solution to a major problem. I mean, yeah.

00:36:40.940 --> 00:36:42.960
And it gets at the worst of the speeders too,

00:36:43.059 --> 00:36:44.780
right? I mean, you're not going to be able to,

00:36:44.920 --> 00:36:48.760
you know, it's, I don't, the average person getting

00:36:48.760 --> 00:36:51.440
called into this section is not going to be just

00:36:51.440 --> 00:36:53.480
going 10 miles per hour of the limit because

00:36:53.480 --> 00:36:56.619
you're going to have slowdowns during your travels.

00:36:56.699 --> 00:36:58.780
So, you know, if you get a ticket here, you're

00:36:58.780 --> 00:37:02.340
probably exceeding the speed limit by significantly

00:37:02.340 --> 00:37:06.320
more than just 10 miles per hour at certain times

00:37:06.320 --> 00:37:09.800
during your trip. I think it's a great. A great

00:37:09.800 --> 00:37:11.639
way to approach things. There's no way to dodge

00:37:11.639 --> 00:37:15.639
this system unless you have a time machine. Yeah,

00:37:15.719 --> 00:37:17.500
the the Colorado Department of Transportation

00:37:17.500 --> 00:37:19.699
installed the cameras along a construction zone.

00:37:19.780 --> 00:37:21.880
The finance, seventy five dollars and zero points

00:37:21.880 --> 00:37:24.059
for exceeding the speed limit. And the police

00:37:24.059 --> 00:37:26.000
issue it to the vehicle's owner, regardless of

00:37:26.000 --> 00:37:28.320
who is driving. So you can always go back to

00:37:28.320 --> 00:37:31.739
the case of it wasn't me, man, wasn't me. But

00:37:31.739 --> 00:37:35.619
I think that's maybe maybe this type of situation

00:37:35.619 --> 00:37:39.360
will allow. law enforcement to do a better job

00:37:39.360 --> 00:37:42.659
attaching tickets to cars and autonomous vehicles.

00:37:42.940 --> 00:37:46.079
Yeah. Since the person seems to be trying to,

00:37:46.119 --> 00:37:49.659
you know, squeeze out from under the punishment

00:37:49.659 --> 00:37:51.659
in so many situations by claiming they weren't

00:37:51.659 --> 00:37:53.940
in the vehicle. That, and I think all 16 -year

00:37:53.940 --> 00:37:55.980
-old drivers should get a speeding ticket automated

00:37:55.980 --> 00:37:57.940
to be like, oh, shit, I can't do this anymore.

00:37:58.739 --> 00:38:01.460
Yeah, I got one from a real cop when I was 15,

00:38:01.579 --> 00:38:04.139
and that did it for me. Oh, I only got one from

00:38:04.139 --> 00:38:06.969
a real cop when I was in my 30s. He pulled me

00:38:06.969 --> 00:38:08.110
and said, do you know how fast you're going?

00:38:08.289 --> 00:38:10.090
And I'm not going to continue the rest of this

00:38:10.090 --> 00:38:13.369
because we're on the auto safety podcast. Anyway,

00:38:14.050 --> 00:38:16.489
go to autosafety .org, click on donate. I'm not

00:38:16.489 --> 00:38:19.250
going to reveal my life as a Fred Perkins type

00:38:19.250 --> 00:38:23.190
character. Let's jump. Please, I'm very respectful

00:38:23.190 --> 00:38:26.889
of officers who stopped me. I was very respectful.

00:38:26.889 --> 00:38:29.110
I just wanted to giggle when he told me how fast

00:38:29.110 --> 00:38:31.050
he thought I was going. It's like this car has

00:38:31.050 --> 00:38:34.210
a turbo, man. There's no one on the road. Yeah,

00:38:34.210 --> 00:38:36.449
I was respectful. I didn't even cry in front

00:38:36.449 --> 00:38:38.190
of him. I saved that from when I got home to

00:38:38.190 --> 00:38:41.670
my parents. It was, look, the New York State

00:38:41.670 --> 00:38:43.849
throw, there was nobody on it. Perfect straight

00:38:43.849 --> 00:38:48.550
shot. Had to try it. Anyway, speaking of Fred

00:38:48.550 --> 00:38:50.590
Perkins, let's go to his favorite topic, Waymo.

00:38:52.309 --> 00:38:58.309
So San Antonio News, it's KSAT 12. I'm actually

00:38:58.309 --> 00:38:59.650
slightly confused about this because there's

00:38:59.650 --> 00:39:03.659
also KAST. Well, it's. The Internet's confusing.

00:39:04.380 --> 00:39:08.340
Where is somebody caught on camera, a Waymo going

00:39:08.340 --> 00:39:13.380
the wrong way down a school drop off area? They

00:39:13.380 --> 00:39:16.340
do this everywhere. And this is not only fascinating

00:39:16.340 --> 00:39:18.000
that they're going the wrong way down a one way

00:39:18.000 --> 00:39:21.440
street, but they blew past a police officer saying,

00:39:21.500 --> 00:39:23.719
you can't go this way. And the Waymo is just

00:39:23.719 --> 00:39:24.980
like, screw you. Who are you going to arrest?

00:39:25.099 --> 00:39:29.940
No one. Sucker. So there's another. frightening

00:39:29.940 --> 00:39:32.519
thing and i mean at this point you have to wonder

00:39:32.519 --> 00:39:36.300
if waymo should just suspend operations and in

00:39:36.300 --> 00:39:39.239
school zones and school zones are around school

00:39:39.239 --> 00:39:42.699
buses they're clearly not not doing a good job

00:39:42.699 --> 00:39:45.420
here right oh and so apparently waymo is now

00:39:45.420 --> 00:39:48.039
operating in san antonio i didn't realize that

00:39:48.039 --> 00:39:50.679
yeah so now they're just like hey austin school

00:39:50.679 --> 00:39:54.590
districts you're not special Yeah, and I don't

00:39:54.590 --> 00:39:58.150
know if they're operating there fully autonomous.

00:39:58.289 --> 00:40:00.829
I guess they might be. I think that started a

00:40:00.829 --> 00:40:02.650
couple months ago, actually. So there's not a

00:40:02.650 --> 00:40:06.909
safety driver or a training driver that's in

00:40:06.909 --> 00:40:08.570
there in cities where they haven't fully deployed

00:40:08.570 --> 00:40:10.969
yet. It looks like they're fully deployed in

00:40:10.969 --> 00:40:14.989
San Antonio. Another Waymo story. There's a story.

00:40:15.070 --> 00:40:19.070
This is in, I believe, Los Angeles. A woman ordered

00:40:19.070 --> 00:40:23.889
a Waymo. It's picking her up at a fast food restaurant

00:40:23.889 --> 00:40:26.849
and it goes the wrong way down the drive -thru.

00:40:27.369 --> 00:40:30.409
The video is hilarious as her son tries to be

00:40:30.409 --> 00:40:32.510
like, Waymo, you can't go here. Way, way, back

00:40:32.510 --> 00:40:35.309
up, back up. It's funny. People in the other

00:40:35.309 --> 00:40:36.530
part of the drive -thru are like, what's going

00:40:36.530 --> 00:40:40.150
on? And the poor son is like, hey, sorry. But

00:40:40.150 --> 00:40:42.670
the weirdest part of the video is not that the

00:40:42.670 --> 00:40:45.829
Waymo does the wrong thing. It's a Waymo. The

00:40:45.829 --> 00:40:48.449
weirdest part is the guy lets his mother get

00:40:48.449 --> 00:40:51.489
into the backseat of the Waymo. Well, it's headed

00:40:51.489 --> 00:40:54.110
the wrong way. I don't know how this video ended.

00:40:54.150 --> 00:40:56.389
Are you sure it wasn't his mother -in -law? Well,

00:40:56.829 --> 00:41:00.110
you know, look, he let another human get into

00:41:00.110 --> 00:41:02.230
the Waymo while it's pointed the wrong direction.

00:41:03.110 --> 00:41:07.289
Yeah, these are two funny little light videos

00:41:07.289 --> 00:41:10.869
you can see giggling about how Waymos are perfect.

00:41:10.969 --> 00:41:13.030
They never drive drunk. They never drive angry.

00:41:13.269 --> 00:41:16.590
They never drive confused. Yeah, well, remember,

00:41:16.710 --> 00:41:20.670
this is end -to -end AI, right? The Waymos are

00:41:20.670 --> 00:41:22.829
going to have to do this hundreds of times before

00:41:22.829 --> 00:41:28.190
it affects the end -to -end AI logic to tell

00:41:28.190 --> 00:41:31.510
them not to do it. So, you know, this is something

00:41:31.510 --> 00:41:35.630
that's going to be a recurring phenomenon with

00:41:35.630 --> 00:41:37.909
the Waymos. And everybody else is using end -to

00:41:37.909 --> 00:41:41.809
-end AI just because they can't interpret stuff

00:41:41.809 --> 00:41:44.869
that doesn't happen over and over again. It's

00:41:44.869 --> 00:41:47.230
just not possible with their end -to -end AI.

00:41:47.960 --> 00:41:50.579
It's just maddening to me that we're in an era

00:41:50.579 --> 00:41:53.900
where we're willing to accept the fact that if

00:41:53.900 --> 00:42:00.380
an end -to -end AI -enabled AV screws up royally,

00:42:00.400 --> 00:42:04.300
that we're supposed to say, yeah, but keep training

00:42:04.300 --> 00:42:06.599
and letting it make this horrible mistake over

00:42:06.599 --> 00:42:10.559
and over again until it's fixed versus fix it,

00:42:10.599 --> 00:42:13.670
you know? It's kind of a whole new world, you

00:42:13.670 --> 00:42:15.710
know, for automobiles. You know, when there was

00:42:15.710 --> 00:42:18.849
a defect in the past like that, it's, you know,

00:42:18.849 --> 00:42:21.329
something you have to fix now. You got to fix

00:42:21.329 --> 00:42:24.809
it. But the way that these companies are describing

00:42:24.809 --> 00:42:26.969
their training models and things, it's like,

00:42:27.050 --> 00:42:32.000
yeah, well, it'll happen at some point. And we're

00:42:32.000 --> 00:42:33.840
not going to stop operating around, you know,

00:42:33.840 --> 00:42:36.619
school buses in the meantime or drive -thrus.

00:42:36.679 --> 00:42:39.260
We can't just fix it. We've got to let the vehicle

00:42:39.260 --> 00:42:43.039
kind of learn its way. Yeah. I mean, when you're

00:42:43.039 --> 00:42:46.920
a terrible driver as a child, going back to my

00:42:46.920 --> 00:42:49.440
speeding ticket story in which I was grounded

00:42:49.440 --> 00:42:52.159
from driving for at least two months, I think.

00:42:52.239 --> 00:42:55.199
Wow, the police are aggressive there. You know,

00:42:55.260 --> 00:42:58.079
that's how we address, you know, human drivers,

00:42:58.159 --> 00:43:00.039
at least, you know, those with parents who are

00:43:00.039 --> 00:43:02.099
looking out for them and trying to make sure

00:43:02.099 --> 00:43:04.739
they learn their lesson and get better at driving.

00:43:05.219 --> 00:43:08.679
They're, you know, removed from action. Whereas

00:43:08.679 --> 00:43:10.900
with these autonomous vehicles, we're saying,

00:43:11.019 --> 00:43:13.400
no, get back out there, kid. You need to learn

00:43:13.400 --> 00:43:16.360
some more. Well, that's the key difference between

00:43:16.360 --> 00:43:19.920
humans and AI. You know, in your case, what you

00:43:19.920 --> 00:43:22.820
did is you had a lesson and you were able to

00:43:22.820 --> 00:43:25.809
expand that out. to all of the circumstances

00:43:25.809 --> 00:43:28.909
where it might apply. The end -to -end AI is

00:43:28.909 --> 00:43:33.489
exactly the opposite. What it has to do is accumulate

00:43:33.489 --> 00:43:37.110
lots and lots of experiences with questionable

00:43:37.110 --> 00:43:40.030
results, and then somehow resolve that into a

00:43:40.030 --> 00:43:44.789
rule that's applicable to those situations. So

00:43:44.789 --> 00:43:48.090
the human logic is just exactly the opposite

00:43:48.090 --> 00:43:52.500
of what the end -to -end AI is doing. attempt

00:43:52.500 --> 00:43:57.360
to equate those two is just dead wrong but as

00:43:57.360 --> 00:44:00.059
friends of the show missy cummings pointed out

00:44:00.059 --> 00:44:02.360
you can keep training these things forever and

00:44:02.360 --> 00:44:05.739
ever and it will eventually hallucinate because

00:44:05.739 --> 00:44:09.179
that's how the math works out it's not uh it's

00:44:09.179 --> 00:44:12.719
not an error in the math it's just the math it's

00:44:12.719 --> 00:44:15.920
gonna hallucinate so maybe the uemo was operating

00:44:15.920 --> 00:44:19.900
properly it was just like no man If I was just

00:44:19.900 --> 00:44:22.019
like, I want to lose weight. So if I go through

00:44:22.019 --> 00:44:24.320
the drive through in reverse, I didn't eat those

00:44:24.320 --> 00:44:27.099
fries and that shake. Who knows? Maybe they thought

00:44:27.099 --> 00:44:29.260
they were in the UK and the driver's side was

00:44:29.260 --> 00:44:31.300
on the right. We just don't know. Maybe. Yeah.

00:44:31.320 --> 00:44:33.820
It was like, I want to I want to learn reverse

00:44:33.820 --> 00:44:36.599
psychology, drive past the bus in the wrong direction.

00:44:38.059 --> 00:44:40.400
But at any rate, we're told we just have to wait

00:44:40.400 --> 00:44:45.380
and be patient while. the AI learns versus we're

00:44:45.380 --> 00:44:47.579
going to fix that. Yeah. In the future, everything

00:44:47.579 --> 00:44:50.960
will be better. It won't be. It's like bringing

00:44:50.960 --> 00:44:54.079
fake it till you make it into the area of automotive

00:44:54.079 --> 00:44:57.599
defects. This is not the way Piggly Wiggly operates,

00:44:57.780 --> 00:45:00.599
by the way. They're very deterministic. Joe's

00:45:00.599 --> 00:45:03.280
over. Well played, Mr. Pert. We haven't heard

00:45:03.280 --> 00:45:06.079
Piggly Wiggly in weeks. I know it. It's been

00:45:06.079 --> 00:45:08.059
a drought. I know. And because of that, you get

00:45:08.059 --> 00:45:11.769
to go into your towel. Oh, thank you. All right,

00:45:11.789 --> 00:45:13.650
well, today I'm kind of going back to my roots,

00:45:13.690 --> 00:45:18.469
talking about nuts and bolts. Screwballs. We've

00:45:18.469 --> 00:45:21.130
spent altogether too much time talking about

00:45:21.130 --> 00:45:26.130
AI and stuff that we can't control. So this will

00:45:26.130 --> 00:45:30.489
be just a very quick tutorial on how computers

00:45:30.489 --> 00:45:33.469
can actually control the car. So there's a component

00:45:33.469 --> 00:45:36.150
called the ball screw, and what it does... is

00:45:36.150 --> 00:45:39.610
it translates rotational motion into translational

00:45:39.610 --> 00:45:43.909
motion. So if you rotate the shaft, a device

00:45:43.909 --> 00:45:47.230
goes back and forth on the shaft. So in electric

00:45:47.230 --> 00:45:49.329
vehicles in particular, this is very important

00:45:49.329 --> 00:45:51.949
because you may not have a hydraulic system.

00:45:52.190 --> 00:45:55.809
So if you want the computer to control where

00:45:55.809 --> 00:45:58.869
the car is going, it's got to have some way of

00:45:58.869 --> 00:46:01.130
talking to the front wheels because the front

00:46:01.130 --> 00:46:04.090
wheels are changing direction and thereby...

00:46:04.809 --> 00:46:06.590
Direct in the car, one direction or another.

00:46:06.789 --> 00:46:10.610
So ball screws are one of several ways to do

00:46:10.610 --> 00:46:15.929
this. They're ubiquitous in cars. There's lots

00:46:15.929 --> 00:46:19.090
of them. And they're really cool little devices.

00:46:19.389 --> 00:46:22.710
So they do it very efficiently. There's a helically

00:46:22.710 --> 00:46:26.230
grooved shaft, and there's a nut that is fitted

00:46:26.230 --> 00:46:28.449
to the shaft. It has recirculating ball bearings

00:46:28.449 --> 00:46:32.690
inside. So you twirl the shaft around, and the...

00:46:33.159 --> 00:46:37.880
and the nut slides back and forth. There's various

00:46:37.880 --> 00:46:41.440
means of controlling this, but in every case,

00:46:41.519 --> 00:46:45.460
there's an electric motor in electric vehicles,

00:46:45.460 --> 00:46:48.960
or a hydraulic motor is an option, that turns

00:46:48.960 --> 00:46:52.199
the shaft precisely a certain amount, and then

00:46:52.199 --> 00:46:56.059
the nut moves a certain amount, and so your car

00:46:56.059 --> 00:46:59.880
will go the required amount left or right, and

00:46:59.880 --> 00:47:03.239
thereby steering the car. There's a lot of ways

00:47:03.239 --> 00:47:07.119
of implementing this. There's various kinds of

00:47:07.119 --> 00:47:09.840
electric motors you can use. You can use a standard

00:47:09.840 --> 00:47:13.820
DC motor. You can use something called a stepping

00:47:13.820 --> 00:47:18.199
motor, which you've all experienced in your printer

00:47:18.199 --> 00:47:21.019
at home. When it sings, it's really operating

00:47:21.019 --> 00:47:24.820
through the stepping motor. Then there's open

00:47:24.820 --> 00:47:27.679
loop control, which is the computer sends a signal,

00:47:27.739 --> 00:47:30.780
which could be either analog or digital, to the

00:47:30.780 --> 00:47:33.619
electric motor. which then rotates the shaft

00:47:33.619 --> 00:47:37.880
the desired amount. So it can be something called

00:47:37.880 --> 00:47:41.280
a voltage signal. A higher voltage might mean

00:47:41.280 --> 00:47:45.360
you're moving more. Low voltage could mean you're

00:47:45.360 --> 00:47:48.300
moving back to the original position. Something

00:47:48.300 --> 00:47:50.739
called a pulse width modulated digital signal,

00:47:50.820 --> 00:47:54.820
which has a certain number of pings in it. And

00:47:54.820 --> 00:47:57.920
the number of pings is related to how far the

00:47:57.920 --> 00:48:02.610
motor rotates. conventionally linked with the

00:48:02.610 --> 00:48:07.349
stepping motor. And then you've got open loop

00:48:07.349 --> 00:48:10.809
control, which is that the computer sends a command

00:48:10.809 --> 00:48:14.210
to the motor, the motor turns, but there's no

00:48:14.210 --> 00:48:17.949
feedback from the motor to the computer that

00:48:17.949 --> 00:48:20.409
says, I've actually accomplished what you wanted

00:48:20.409 --> 00:48:23.570
me to do. Now in an EV where you have a human

00:48:23.570 --> 00:48:28.750
being sitting behind the steering wheel, if it's

00:48:28.750 --> 00:48:31.480
a little bit off, The human will make an adjustment.

00:48:31.800 --> 00:48:34.280
So open loop control can work perfectly well

00:48:34.280 --> 00:48:38.679
on those. But in an AV, you've really got to

00:48:38.679 --> 00:48:40.559
control it more carefully. So you've got to have

00:48:40.559 --> 00:48:44.260
some kind of feedback mechanism that says, you

00:48:44.260 --> 00:48:46.800
know, I wanted to turn the wheel or I wanted

00:48:46.800 --> 00:48:49.900
to turn the front wheels x degrees. So I'm going

00:48:49.900 --> 00:48:52.940
to measure how much they actually turn. That

00:48:52.940 --> 00:48:56.380
gets feedback. It's fed back as a closed loop

00:48:56.380 --> 00:48:58.840
to the computer. Computer makes appropriate.

00:49:00.019 --> 00:49:04.059
adjustments and so on. So that's a closed loop

00:49:04.059 --> 00:49:08.099
control. Stepping motors are often used because

00:49:08.099 --> 00:49:11.139
it's well suited to the pulse with modulated

00:49:11.139 --> 00:49:14.980
signal that comes out of the computer. If it's

00:49:14.980 --> 00:49:17.300
a two -wire control system, if you've got a multi

00:49:17.300 --> 00:49:20.480
-wire cable, you can have each of the different

00:49:20.480 --> 00:49:24.840
channels associated with a unique voltage that

00:49:24.840 --> 00:49:26.760
goes up or down depending on what your digital

00:49:26.760 --> 00:49:29.650
signal is. And blah, blah, blah. I know you're

00:49:29.650 --> 00:49:32.690
all blanking out by now. But the point is, there's

00:49:32.690 --> 00:49:36.309
different ways of doing this. I was trying to

00:49:36.309 --> 00:49:38.730
hide it. I was trying to hide it. Yeah, not very

00:49:38.730 --> 00:49:44.690
well. Another thing to think about is that if

00:49:44.690 --> 00:49:47.369
it's a digital control system, the resolution

00:49:47.369 --> 00:49:51.070
of the digital signals is inversely proportional

00:49:51.070 --> 00:49:54.710
to the speed it takes to convert the analog information

00:49:54.710 --> 00:49:58.070
to the digital information. In a very demanding

00:49:58.070 --> 00:50:01.570
situation, this can be a factor in how well the

00:50:01.570 --> 00:50:06.670
computer is driving the vehicle and how well

00:50:06.670 --> 00:50:08.650
the vehicle is responding to the commuter commands.

00:50:09.329 --> 00:50:13.309
So here's the point of all this. All these considerations

00:50:13.309 --> 00:50:16.250
impact safety because they all affect how the

00:50:16.250 --> 00:50:19.989
actual massive machine called the vehicle is

00:50:19.989 --> 00:50:22.909
operating, where it's going and what it's doing.

00:50:23.550 --> 00:50:25.670
And this is throughout the vehicle. It can be

00:50:25.670 --> 00:50:28.690
brakes. It can be steering. It can be the fan.

00:50:29.150 --> 00:50:32.849
All of these digital things happen. So for any

00:50:32.849 --> 00:50:36.349
company that says, well, my vehicle has operated

00:50:36.349 --> 00:50:40.750
for so many millions of miles and we've only

00:50:40.750 --> 00:50:43.610
had X number of people who are killed and it's

00:50:43.610 --> 00:50:46.710
our fault. Therefore, it's safe to go ahead.

00:50:48.369 --> 00:50:51.449
It's complete nonsense because all of these tiny

00:50:51.449 --> 00:50:54.949
little factors. The details matter, will affect

00:50:54.949 --> 00:50:57.010
the overall safety of an individual vehicle.

00:50:57.670 --> 00:51:01.809
And unless your base of statistics is made up

00:51:01.809 --> 00:51:04.530
of identical vehicles that are identical all

00:51:04.530 --> 00:51:07.650
the way down to the smallest part of the vehicle

00:51:07.650 --> 00:51:12.210
that's affecting its operational safety, to extend

00:51:12.210 --> 00:51:17.789
any individual vehicle's performance into an

00:51:17.789 --> 00:51:20.170
aggregate to say that, well, my fleet is perfectly

00:51:20.170 --> 00:51:23.380
fine because... We've had all these different

00:51:23.380 --> 00:51:26.019
vehicles, all these different types. Everything's

00:51:26.019 --> 00:51:28.480
been safe. We haven't killed too many people

00:51:28.480 --> 00:51:30.820
yet. Therefore, we're going to be safe in the

00:51:30.820 --> 00:51:35.000
future. It's just nonsense. In particular, if

00:51:35.000 --> 00:51:37.960
you look at the evolution of the Waymo vehicle,

00:51:38.199 --> 00:51:41.780
for example, it's gone through different types

00:51:41.780 --> 00:51:44.780
of mechanical vehicles, many different loads

00:51:44.780 --> 00:51:47.679
of software. In fact, the software stack now

00:51:47.679 --> 00:51:50.320
is completely different from the software stack

00:51:50.320 --> 00:51:54.079
they used. when they accumulated the statistics

00:51:54.079 --> 00:51:56.360
that they're using now to try to justify their

00:51:56.360 --> 00:51:59.940
safety. So it's very difficult. The devil's in

00:51:59.940 --> 00:52:04.400
the details, even down to the level of what kind

00:52:04.400 --> 00:52:08.679
of motor are you using to drive the steering

00:52:08.679 --> 00:52:12.519
that's in the vehicle. So that's the point of

00:52:12.519 --> 00:52:15.519
this, right? The details are important. The details

00:52:15.519 --> 00:52:17.679
matter. And there's a lot of details that are

00:52:17.679 --> 00:52:20.519
not held constant. from one vehicle to another,

00:52:20.639 --> 00:52:24.079
or even within a particular vehicle model. So

00:52:24.079 --> 00:52:28.719
be careful when somebody says, we've got a lot

00:52:28.719 --> 00:52:31.179
of experience, therefore the future is going

00:52:31.179 --> 00:52:34.860
to be great. There's really no basis for that

00:52:34.860 --> 00:52:37.980
unless they've done the homework to look down

00:52:37.980 --> 00:52:42.139
into the guts of the vehicle and how all these

00:52:42.139 --> 00:52:46.559
different details have affected the accumulation

00:52:46.559 --> 00:52:49.099
of those statistics. Does that make any sense?

00:52:49.400 --> 00:52:52.460
You're still awake, though. I can see that. I

00:52:52.460 --> 00:52:55.079
mean, the end part, I mean, I'm still trying

00:52:55.079 --> 00:52:57.380
to figure out exactly how a ball screw works.

00:52:57.519 --> 00:53:00.800
I've pulled up some diagrams and figuring all

00:53:00.800 --> 00:53:03.599
that out. But, you know, you ended on an upward

00:53:03.599 --> 00:53:07.199
note as far as details go. I want to know, as

00:53:07.199 --> 00:53:11.099
a consumer, I go to buy a car. What should I

00:53:11.099 --> 00:53:13.360
take away from what you said? Do I worry about

00:53:13.360 --> 00:53:18.079
that or do I not worry about that? Worrying about

00:53:18.079 --> 00:53:20.059
it is unproductive because you'll never get the

00:53:20.059 --> 00:53:22.940
answer. Okay. Okay. So this is something that

00:53:22.940 --> 00:53:26.739
should be considered at the level of the government

00:53:26.739 --> 00:53:30.300
regulators who are giving people license to go

00:53:30.300 --> 00:53:32.659
ahead and operate these vehicles on public streets.

00:53:33.179 --> 00:53:37.699
They should be asking, what's the heritage? You're

00:53:37.699 --> 00:53:40.440
saying that you've gone, you know, 100 million

00:53:40.440 --> 00:53:43.519
miles, no problems. Well, how many different

00:53:43.519 --> 00:53:48.980
models? of cars are you doing that with? Is the

00:53:48.980 --> 00:53:52.019
model that you're using now and the software

00:53:52.019 --> 00:53:55.460
stack that you're using now a significant part

00:53:55.460 --> 00:53:57.280
of these statistics or are you starting over?

00:53:57.420 --> 00:54:00.880
Are you ringing the bell now with a new design

00:54:00.880 --> 00:54:03.739
and you're going to tell me in a year how well

00:54:03.739 --> 00:54:06.219
it's worked? So there's a difference between

00:54:06.219 --> 00:54:11.539
trailing metrics, which are like that. We've

00:54:11.539 --> 00:54:13.079
gone 100 million miles, haven't killed anybody

00:54:13.079 --> 00:54:15.800
yet. That's a trailing metric. And leading metrics,

00:54:15.920 --> 00:54:19.219
which say that based upon this new information,

00:54:19.559 --> 00:54:23.079
this is how the vehicle should be responding.

00:54:23.380 --> 00:54:26.380
So a leading metric, for example, would be number

00:54:26.380 --> 00:54:29.780
of factory orders to produce cars. Well, you

00:54:29.780 --> 00:54:33.019
know that if the number of orders to produce

00:54:33.019 --> 00:54:36.719
cars goes up, it's very likely to have more cars

00:54:36.719 --> 00:54:39.070
on the street. In a few years, right? Because

00:54:39.070 --> 00:54:40.809
you've got the orders and you're doing that.

00:54:40.829 --> 00:54:45.250
It's a leading metric. If you say, well, over

00:54:45.250 --> 00:54:47.769
the last five years, X number of people have

00:54:47.769 --> 00:54:50.809
died. That's a trillion metric. And the fact

00:54:50.809 --> 00:54:52.889
that X number of people died over the last five

00:54:52.889 --> 00:54:56.949
years is only distantly related to the number

00:54:56.949 --> 00:54:59.429
of people who will die in the next five years.

00:55:00.130 --> 00:55:02.429
So do you see the difference there? I gotcha.

00:55:02.510 --> 00:55:05.710
I gotcha. Some of the metric system. Yeah. And

00:55:05.710 --> 00:55:10.469
so. If you were to say, okay, I've had this particular

00:55:10.469 --> 00:55:13.469
driving hazard, it's because of a lousy ball

00:55:13.469 --> 00:55:17.230
screw. I've now replaced the ball screw with

00:55:17.230 --> 00:55:20.949
a superior design. And here's the statistics

00:55:20.949 --> 00:55:23.690
from the engineering results that say why I'll

00:55:23.690 --> 00:55:26.630
solve this problem. So therefore, we shouldn't

00:55:26.630 --> 00:55:29.510
have this problem in the future. That's a different

00:55:29.510 --> 00:55:32.289
approach. That's using leading statistics or

00:55:32.289 --> 00:55:34.690
leading indicators to say that in the future.

00:55:35.260 --> 00:55:37.400
It shouldn't be a problem. That's the whole heart

00:55:37.400 --> 00:55:41.860
of the safety recalls, right? You've got evidence

00:55:41.860 --> 00:55:45.019
that a particular problem has been identified

00:55:45.019 --> 00:55:48.500
and you've got evidence that you fixed that particular

00:55:48.500 --> 00:55:51.559
problem. Therefore, in the future operation of

00:55:51.559 --> 00:55:53.639
the vehicles, once they've had the safety recall

00:55:53.639 --> 00:55:56.440
in place, you shouldn't have that particular

00:55:56.440 --> 00:56:01.179
problem for the future operation. Unless you're

00:56:01.179 --> 00:56:05.250
talking about a rear view camera. You still shouldn't

00:56:05.250 --> 00:56:08.190
have it if they fix the problem, but still. So

00:56:08.190 --> 00:56:10.730
I guess what I'm trying to highlight here is

00:56:10.730 --> 00:56:14.670
the difference between illusions based upon past

00:56:14.670 --> 00:56:21.869
behavior of an idea and the reality of the operations

00:56:21.869 --> 00:56:26.010
in the future based upon the actual understanding

00:56:26.010 --> 00:56:30.510
of what's happening in the vehicle. End of rant.

00:56:30.909 --> 00:56:34.639
End of rant. Yes. And with that, we'll jump into

00:56:34.639 --> 00:56:41.980
recalls. First up, Toyota, 73 ,528 vehicles.

00:56:42.039 --> 00:56:46.480
The 2023 to 2025 Toyota Corolla Cross Hybrid.

00:56:47.079 --> 00:56:50.460
The minimum, oh, this is under certain conditions,

00:56:50.679 --> 00:56:53.659
pedestrians may have reduced ability to detect

00:56:53.659 --> 00:56:56.639
the AVAS sound, increasing risk of crash with

00:56:56.639 --> 00:56:59.940
a pedestrian during reversing maneuvers. Now,

00:56:59.980 --> 00:57:02.599
this is the... Like the, I don't know, what's

00:57:02.599 --> 00:57:06.019
the, I don't know, the whistler. Yeah, the electric

00:57:06.019 --> 00:57:09.539
vehicles that don't produce enough noise to alert

00:57:09.539 --> 00:57:11.440
people around them until they reach a certain

00:57:11.440 --> 00:57:14.599
speed. So under that speed, they have to play

00:57:14.599 --> 00:57:19.260
a sound at a certain, let's see, I think they

00:57:19.260 --> 00:57:24.239
call it a certain sound pressure level requirement

00:57:24.239 --> 00:57:29.789
that's specified in FMVSS. These Toyotas, when

00:57:29.789 --> 00:57:32.309
they're in reverse, are not producing a loud

00:57:32.309 --> 00:57:37.389
enough sound. So they're going to modify your

00:57:37.389 --> 00:57:41.130
software to make the reverse sound louder. Owners

00:57:41.130 --> 00:57:43.889
should hear about that in late May. I think it's

00:57:43.889 --> 00:57:46.510
just going to be beep, beep, beep, beep, like

00:57:46.510 --> 00:57:48.769
a big truck. It will confuse people. They won't

00:57:48.769 --> 00:57:51.329
know where it's coming from. Anyway, next up,

00:57:51.429 --> 00:57:56.340
Mercedes 932 vehicles. The 2026 Mercedes Maybach

00:57:56.340 --> 00:58:00.460
GLS 604 Matic. I think that's what Fred, Fred,

00:58:00.500 --> 00:58:03.960
you have the Maybach, right? Some other GLS 450

00:58:03.960 --> 00:58:08.320
4Matic. I've got one, but it's wet, so I can't

00:58:08.320 --> 00:58:11.000
use it. There are all these 2026 Mercedes that

00:58:11.000 --> 00:58:14.559
I'm sure cost well over $100 ,000. And a side

00:58:14.559 --> 00:58:16.139
note, I have a friend who grew up in Europe and

00:58:16.139 --> 00:58:17.880
he's like, I'm always like, oh, some Mercedes

00:58:17.880 --> 00:58:20.179
really nice. And he's like, I would never buy

00:58:20.179 --> 00:58:22.360
a Mercedes. And I was like, why? He said, because

00:58:22.360 --> 00:58:24.500
we use them as taxi cabs in Europe. I was like,

00:58:24.559 --> 00:58:29.300
oh, fair point. This is an event of a crash requiring

00:58:29.300 --> 00:58:32.000
passenger airbag activation that passenger airbag

00:58:32.000 --> 00:58:35.519
might not deploy as intended, which could impair

00:58:35.519 --> 00:58:37.360
the restraining effect of the airbag since it

00:58:37.360 --> 00:58:41.980
didn't deploy. Yeah, what's causing that all

00:58:41.980 --> 00:58:44.920
here appears to be that they, or is, that they

00:58:44.920 --> 00:58:48.719
have used, I think they used the wrong fabric

00:58:48.719 --> 00:58:52.340
in manufacturing the airbags. They call it a...

00:58:52.900 --> 00:58:57.199
Let's see. They call it a deviant or a deviation

00:58:57.199 --> 00:59:03.940
from the airbag fabric material that they typically

00:59:03.940 --> 00:59:06.519
use. Look, if you can afford a Maybach, you're

00:59:06.519 --> 00:59:09.980
some sort of deviant. Yeah, they're – you know,

00:59:10.039 --> 00:59:12.010
they're – It looks like they're going to get

00:59:12.010 --> 00:59:15.130
you a new airbag. I'm assuming the replacement

00:59:15.130 --> 00:59:18.309
here has got to be replacing the airbag module.

00:59:18.429 --> 00:59:20.610
So they're replacing the airbag module. And it's

00:59:20.610 --> 00:59:23.369
only going to take them a month or so to get

00:59:23.369 --> 00:59:25.309
this one ready. It looks like mid -May, later

00:59:25.309 --> 00:59:27.510
May is when owners will be hearing about this.

00:59:27.809 --> 00:59:30.809
Well, there aren't very many Maybachs. There's

00:59:30.809 --> 00:59:33.530
only 900 vehicles that recall anyway. And while

00:59:33.530 --> 00:59:36.949
you're getting your Maybach repaired, while you're

00:59:36.949 --> 00:59:38.769
waiting for them to do that, just go to autosafety

00:59:38.769 --> 00:59:40.900
.org and click on Donate. I mean, it's a rounding

00:59:40.900 --> 00:59:43.960
error for you. Yeah, well, there were only 13

00:59:43.960 --> 00:59:46.239
Maybachs involved in this recall. Well, those

00:59:46.239 --> 00:59:50.260
13 people got extra money to spare is what I'm

00:59:50.260 --> 00:59:53.840
trying to say. Hey, and our last recall, shock

00:59:53.840 --> 00:59:57.739
of all shocks, the thunder from down under, the

00:59:57.739 --> 01:00:02.860
world reigning champion, the Cisco kid, I don't

01:00:02.860 --> 01:00:07.960
know, Ford Motor Company, 422 ,613 vehicles.

01:00:08.539 --> 01:00:10.579
Have they ever recalled less than 100 ,000 vehicles

01:00:10.579 --> 01:00:15.079
at a time? Anyway. Yes. The 2021 to 2023 Lincoln

01:00:15.079 --> 01:00:18.199
Navigator, the same for the Expedition, the 2022

01:00:18.199 --> 01:00:22.400
to 2023 Ford Super Duty. Interesting. A windshield

01:00:22.400 --> 01:00:25.519
wiper arm may operate erratically, becoming inoperable

01:00:25.519 --> 01:00:28.300
and detached from the vehicle. That's not great.

01:00:28.639 --> 01:00:34.699
We should get that one fixed. I like how it connects

01:00:34.699 --> 01:00:39.889
into the Ford Super Duty, though. I don't even

01:00:39.889 --> 01:00:44.190
know what that means, but essentially you're

01:00:44.190 --> 01:00:48.010
going to have a windshield wiper arm that doesn't

01:00:48.010 --> 01:00:50.789
work. And if you're in a storm, obviously that

01:00:50.789 --> 01:00:54.010
is not just a safety defect in this case when

01:00:54.010 --> 01:00:56.650
it breaks, but it's also a violation of the Federal

01:00:56.650 --> 01:00:58.869
Motor Vehicle Safety Standards because you wouldn't

01:00:58.869 --> 01:01:00.809
have an operable wiper system at that point.

01:01:02.869 --> 01:01:04.929
It looks like they're going to just replace your

01:01:04.929 --> 01:01:10.539
wiper system. But, but they don't seem to have

01:01:10.539 --> 01:01:14.519
a final date for this action. Owners are going

01:01:14.519 --> 01:01:16.300
to hear about it in the next couple of weeks,

01:01:16.440 --> 01:01:20.840
but it's not clear when the actual remedy will

01:01:20.840 --> 01:01:23.039
be available. Well, Ford's just hoping that a

01:01:23.039 --> 01:01:25.539
Chinese company will buy them soon enough. No.

01:01:25.639 --> 01:01:28.480
Oh, okay. So I lied. We do have one more, but

01:01:28.480 --> 01:01:30.739
this one didn't make it into NHTSA's database,

01:01:31.059 --> 01:01:34.280
if that's true. I get it. Yeah, I haven't seen.

01:01:34.539 --> 01:01:37.300
Let me see if I've seen that one yet. I didn't

01:01:37.300 --> 01:01:40.420
check this morning. We have a link to Motor Trend.

01:01:41.000 --> 01:01:45.039
This is a stop sale for the 2026 Chevrolet Corvette

01:01:45.039 --> 01:01:47.980
due to an issue with the rear turn signals, the

01:01:47.980 --> 01:01:52.539
2025 and 2026 Corvette, which is a, I don't know

01:01:52.539 --> 01:01:55.940
how I feel about the way it looks. I don't even

01:01:55.940 --> 01:01:57.760
know what it looks like. I mean, it's got a picture

01:01:57.760 --> 01:02:02.639
on this site. Yeah, the turn signal's not good.

01:02:02.940 --> 01:02:05.400
GM says they're going to fix it over the air.

01:02:05.820 --> 01:02:12.000
which is cool. Yeah, when we'll... Yeah, we don't

01:02:12.000 --> 01:02:13.639
really know anything about this one yet because

01:02:13.639 --> 01:02:15.880
it showed up in the news, but for whatever reason,

01:02:15.940 --> 01:02:18.199
it hasn't showed up on NHTSA's website, so we

01:02:18.199 --> 01:02:20.980
don't have the defect submission. Yeah, so if

01:02:20.980 --> 01:02:22.780
you have this vehicle, tell us, what do you think

01:02:22.780 --> 01:02:24.860
about the way it looks? And if you don't have

01:02:24.860 --> 01:02:26.460
this vehicle and you've seen it, what do you

01:02:26.460 --> 01:02:31.000
think about the way it looks? Anyway, with that,

01:02:31.019 --> 01:02:34.059
that's our show. We'll be back next week. Fingers

01:02:34.059 --> 01:02:40.619
crossed. Bye -bye. Bye -bye. For more information,

01:02:40.780 --> 01:02:43.719
visit www .autosafety .org.