WEBVTT

00:00:01.960 --> 00:00:05.559
You're listening to There Auto Be A Law, the

00:00:05.559 --> 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 --> 00:00:18.300
over 50 years, the Center for Auto Safety has

00:00:18.300 --> 00:00:25.820
worked to make cars safer. Hey, listeners, welcome

00:00:25.820 --> 00:00:30.820
to another episode of the Waymo Zone. Sure. I

00:00:30.820 --> 00:00:35.619
don't know. No. No, not the way my... Okay. Anyway,

00:00:35.880 --> 00:00:39.619
today is Wednesday, July 22nd. The weather has

00:00:39.619 --> 00:00:41.799
cooled down dramatically where I am. I don't

00:00:41.799 --> 00:00:44.439
know if it is for you. Oh, for Michael it's cooled

00:00:44.439 --> 00:00:48.740
down and Fred thinks it's 1972 and he's going

00:00:48.740 --> 00:00:50.939
to go play tennis because he's wearing a headband.

00:00:51.179 --> 00:00:55.289
It's very disturbing. For those of you... I was

00:00:55.289 --> 00:00:58.369
thinking more like Dr. J. Yeah, it's definitely

00:00:58.369 --> 00:01:02.469
the 1970s. Whatever's happening there, it's not

00:01:02.469 --> 00:01:04.569
right. But he's back, so we're happy about that.

00:01:04.629 --> 00:01:06.909
Fred, last week we let listeners know that you

00:01:06.909 --> 00:01:08.769
were crash -testing Waymo's. How did that go?

00:01:10.030 --> 00:01:12.569
Went pretty well, yeah. We ran into a couple

00:01:12.569 --> 00:01:15.930
of Zookses, but, you know, kept it all in the

00:01:15.930 --> 00:01:18.049
family. All right, well, with that, let's get

00:01:18.049 --> 00:01:21.719
into some real news. Hard -heading stuff. And

00:01:21.719 --> 00:01:23.719
we'll talk more about Waymo, but this is more

00:01:23.719 --> 00:01:27.079
about robo -taxis in general. From the Austin

00:01:27.079 --> 00:01:30.239
American Statesman is an article talking about

00:01:30.239 --> 00:01:34.459
how first responders and autonomous robo -taxis

00:01:34.459 --> 00:01:37.840
don't get along very well. And I've got a couple

00:01:37.840 --> 00:01:41.500
good quotes I've highlighted here. This is from

00:01:41.500 --> 00:01:45.269
Austin, Travis County. And the quote is, but

00:01:45.269 --> 00:01:47.629
robo taxis either blocked or ignored the direction

00:01:47.629 --> 00:01:51.250
of Austin police and 25 percent of reported incidences

00:01:51.250 --> 00:01:54.930
from July 2023 to June 2026, according to the

00:01:54.930 --> 00:01:56.950
city of Austin's autonomous vehicle database.

00:01:57.189 --> 00:02:00.510
Data collected by the city since 2023 shows interference

00:02:00.510 --> 00:02:03.209
with law enforcement and other traffic is among

00:02:03.209 --> 00:02:05.629
the top safety concerns surrounding autonomous

00:02:05.629 --> 00:02:08.569
vehicles in the city. Further down, we've got

00:02:08.569 --> 00:02:11.270
a quote says. I feel like whatever we say makes

00:02:11.270 --> 00:02:13.569
no difference at this point. One Austin police

00:02:13.569 --> 00:02:16.270
officer wrote in a report to the city. What do

00:02:16.270 --> 00:02:18.830
we need to do to make sure this does not happen

00:02:18.830 --> 00:02:21.389
again and for Waymo to understand how serious

00:02:21.389 --> 00:02:25.009
this is? Waymo, of course, replies saying, safety

00:02:25.009 --> 00:02:27.270
is our number one priority. We work with the

00:02:27.270 --> 00:02:30.490
cities and everything closely. We're listening

00:02:30.490 --> 00:02:33.349
to your concerns. As soon as the next available

00:02:33.349 --> 00:02:37.849
representatives will take your call. But it brings

00:02:37.849 --> 00:02:40.150
up the issue that I bring up repeatedly. Waymos

00:02:40.150 --> 00:02:42.889
and robo taxis and Zooks, they can do whatever

00:02:42.889 --> 00:02:45.930
the hell they want and there's no consequences.

00:02:46.770 --> 00:02:49.590
Am I right? Well, there's consequences for the

00:02:49.590 --> 00:02:52.629
city who is having to pay everyone who's responding

00:02:52.629 --> 00:02:55.550
to these scenes, right? There just aren't any

00:02:55.550 --> 00:02:58.229
consequences for the company that stands to gain

00:02:58.229 --> 00:03:02.550
from all this. Yes. And part of Waymo's response

00:03:02.550 --> 00:03:04.590
is saying, hey, we have all of this training.

00:03:05.259 --> 00:03:08.080
For first responders. So, hey, first responders,

00:03:08.240 --> 00:03:10.719
we're making your job harder, but we've offered

00:03:10.719 --> 00:03:14.259
you training. And so I took their training. We're

00:03:14.259 --> 00:03:16.379
going to make your job harder and make you go

00:03:16.379 --> 00:03:18.819
back to school. So, yeah. So I'm going right

00:03:18.819 --> 00:03:21.500
into my gaslight because I did their online training

00:03:21.500 --> 00:03:25.800
because it's available to anybody. And it's amazing.

00:03:26.000 --> 00:03:30.099
So I'm going to quote from their. their training

00:03:30.099 --> 00:03:32.599
here. This is a direct quote from their online

00:03:32.599 --> 00:03:35.400
training because it's just done online. Waymo

00:03:35.400 --> 00:03:38.639
vehicles can only be operated without a human

00:03:38.639 --> 00:03:41.719
driver in autonomous mode within specific geofenced

00:03:41.719 --> 00:03:44.159
areas where they have been tested and validated

00:03:44.159 --> 00:03:47.060
to safely perform all the dynamic tasks of driving.

00:03:48.020 --> 00:03:52.580
And we know that is not true because they have

00:03:52.580 --> 00:03:55.039
not been validated to safely perform all the

00:03:55.039 --> 00:03:59.360
dynamic tasks of driving. Yeah, that's that's

00:03:59.360 --> 00:04:02.060
nonsense. Well, not only have they not been,

00:04:02.099 --> 00:04:05.500
they cannot be. It's just not enough money and

00:04:05.500 --> 00:04:09.000
resources in the world to validate all of the

00:04:09.000 --> 00:04:12.580
potential critical safety failures that they

00:04:12.580 --> 00:04:15.039
might encounter. It's amazing. Further down.

00:04:15.120 --> 00:04:20.560
So they have a this is how Waymo's fully autonomous

00:04:20.560 --> 00:04:22.819
vehicles respond to police and emergency vehicles.

00:04:23.560 --> 00:04:25.699
The Waymo vehicle uses its sensors to detect,

00:04:25.779 --> 00:04:27.779
identify, and respond to active emergency vehicles

00:04:27.779 --> 00:04:30.000
and first responders. Sirens, flashing lights,

00:04:30.120 --> 00:04:35.819
emergency vehicles. That fails repeatedly. If

00:04:35.819 --> 00:04:38.259
the Waymo autonomous vehicle detects that a police

00:04:38.259 --> 00:04:40.019
or emergency vehicle is behind it and flashing

00:04:40.019 --> 00:04:42.519
its light, the Waymo vehicle is designed to pull

00:04:42.519 --> 00:04:44.600
over and stop when it finds a safe place to do

00:04:44.600 --> 00:04:47.800
so. Once stopped for a traffic stop, the vehicle

00:04:47.800 --> 00:04:49.860
remains stationary for the duration of the stop.

00:04:50.569 --> 00:04:53.250
None of this is this is this is aspirational

00:04:53.250 --> 00:04:55.870
at best. And so they have as part of this, as

00:04:55.870 --> 00:04:58.110
you as a first responder can earn a certificate

00:04:58.110 --> 00:05:00.529
saying that you understand how autonomous vehicles

00:05:00.529 --> 00:05:04.769
work. So but but you don't. Right. Because they're

00:05:04.769 --> 00:05:07.910
not giving you any insight into the unpredictability

00:05:07.910 --> 00:05:11.410
by making statements that are conclusory and

00:05:11.410 --> 00:05:14.470
essentially pretend that they don't have any

00:05:14.470 --> 00:05:17.689
of these safety issues that we've seen. So my

00:05:17.689 --> 00:05:21.439
my wife or some of her. clients they she um i

00:05:21.439 --> 00:05:24.040
will not reveal who they are but some of them

00:05:24.040 --> 00:05:25.959
are these large tech companies and she does nothing

00:05:25.959 --> 00:05:28.439
with tech she's a horticulturalist but these

00:05:28.439 --> 00:05:30.319
large tech companies because they're under consent

00:05:30.319 --> 00:05:33.160
decrees from the government for releasing privacy

00:05:33.160 --> 00:05:36.860
she has to take these online safety courses every

00:05:36.860 --> 00:05:39.939
year and it's a very in -depth process of how

00:05:39.939 --> 00:05:41.899
you're not going to take people's data and blah

00:05:41.899 --> 00:05:44.160
blah blah and she has no access to any of this

00:05:44.160 --> 00:05:46.339
stuff anyway she can't cause any harm whatsoever

00:05:46.339 --> 00:05:50.160
but this waymo training for first responders

00:05:50.160 --> 00:05:53.100
i clicked through all the slides didn't answer

00:05:53.100 --> 00:05:55.500
a single question and it brings me to the quiz

00:05:55.500 --> 00:05:57.759
it doesn't stop me says that i got anything wrong

00:05:57.759 --> 00:06:01.100
and go ahead and so for me as a first responder

00:06:01.100 --> 00:06:04.220
where i'm in a life -saving situation how many

00:06:04.220 --> 00:06:07.240
questions do you think are required for me to

00:06:07.240 --> 00:06:12.000
earn a certificate from waymo any guesses zero

00:06:12.000 --> 00:06:18.040
close oh it is eight questions Eight questions.

00:06:18.220 --> 00:06:21.839
And one of them is, they're all multiple guests.

00:06:22.459 --> 00:06:26.759
And one of the questions I highlighted is, what

00:06:26.759 --> 00:06:31.040
is the lowest level on the SAE scale to not require

00:06:31.040 --> 00:06:34.259
an AV passenger to hold a valid driver's license?

00:06:34.839 --> 00:06:38.060
Level three, level four, level five, none of

00:06:38.060 --> 00:06:41.639
the above. This is a dumb question for a number

00:06:41.639 --> 00:06:44.980
of reasons. One, it implies that the SAE scale

00:06:44.980 --> 00:06:53.529
levels. mean anything, which... It also kind

00:06:53.529 --> 00:06:58.290
of conflates the state issuing driver's license

00:06:58.290 --> 00:07:02.470
with SAE. There's no such thing yet, right? Exactly.

00:07:03.269 --> 00:07:06.230
Presumably, an SAE level 3 vehicle requires a

00:07:06.230 --> 00:07:08.370
human takeover at some point, so you would expect

00:07:08.370 --> 00:07:10.990
that human to be required to have a driver's

00:07:10.990 --> 00:07:14.079
license, but there is no... hard and fast requirement

00:07:14.079 --> 00:07:16.480
like that in law. Right. But as somebody who's

00:07:16.480 --> 00:07:18.720
taken a number of standardized tests, Michael,

00:07:18.779 --> 00:07:20.540
like I'm sure you've taken the LSAT to go to

00:07:20.540 --> 00:07:22.819
law school. I love standardized tests. So this

00:07:22.819 --> 00:07:24.819
question, this is my favorite question ever.

00:07:25.000 --> 00:07:27.279
Again, I'll repeat it. What is the lowest level

00:07:27.279 --> 00:07:30.720
on the SAE scale to not require an AV passenger

00:07:30.720 --> 00:07:34.579
to hold a valid driver's license? You're the

00:07:34.579 --> 00:07:37.060
passenger. Why do you need a driver's license

00:07:37.060 --> 00:07:41.949
ever? That's my takeaway. You're the passenger.

00:07:42.129 --> 00:07:44.689
You're the customer. You're not responsible for

00:07:44.689 --> 00:07:47.170
anything. When I get into a cab, I don't need

00:07:47.170 --> 00:07:52.610
a driver's license. That's mine. This is my gaslight.

00:07:52.689 --> 00:07:57.750
Is Waymo wasting first responders' time? They

00:07:57.750 --> 00:07:59.889
do have a diagram in there showing these are

00:07:59.889 --> 00:08:02.009
the parts of the vehicle you can't cut if you

00:08:02.009 --> 00:08:04.170
have to get in there and disable battery and

00:08:04.170 --> 00:08:06.750
electronics. And it's pretty much the entire

00:08:06.750 --> 00:08:11.350
vehicle. which is mind blowing to me. It has

00:08:11.350 --> 00:08:13.189
in depth in there how to disable their batteries.

00:08:14.149 --> 00:08:17.189
That's not part of this quiz at all. None of

00:08:17.189 --> 00:08:20.730
that. There's no questions in there at all about

00:08:20.730 --> 00:08:22.769
that. They're all nonsense questions. The closest

00:08:22.769 --> 00:08:25.810
they get to it is what color are the high voltage

00:08:25.810 --> 00:08:30.250
lines? That's important. Yeah, they're orange.

00:08:30.410 --> 00:08:34.750
That's it. That's the answer. So my gaslight

00:08:34.750 --> 00:08:37.409
right off the bat is Waymo trying to pretend

00:08:37.409 --> 00:08:40.250
that they're offering valid certificates to first

00:08:40.250 --> 00:08:42.169
responders. And instead, they're just wasting

00:08:42.169 --> 00:08:45.250
more and more of first responders times with

00:08:45.250 --> 00:08:50.649
nonsense. I want to reinforce the nonsense. Oh,

00:08:50.669 --> 00:08:53.750
good. One of the statements you said was that

00:08:53.750 --> 00:08:57.850
Waymo asserts that the vehicles are designed

00:08:57.850 --> 00:09:01.009
to stop for all emergency vehicles with using

00:09:01.009 --> 00:09:04.159
their sensors and all that sort of stuff. I know

00:09:04.159 --> 00:09:08.100
this sounds like nitpicking, but there's a difference

00:09:08.100 --> 00:09:11.820
between something being designed to perform an

00:09:11.820 --> 00:09:14.620
action and something actually being proven to

00:09:14.620 --> 00:09:18.279
perform an action. So that's a critical sleight

00:09:18.279 --> 00:09:20.840
of hand that they put in this, and getting people

00:09:20.840 --> 00:09:27.440
to equate the design versus the proof is pretty

00:09:27.440 --> 00:09:31.259
sneaky. It's pretty good. But also we've seen

00:09:31.259 --> 00:09:34.580
tons of examples of Waymo's ignoring the police

00:09:34.580 --> 00:09:40.379
using its sensors. Ignoring fire trucks, ignoring

00:09:40.379 --> 00:09:45.419
ambulances, all using their sensors. Ignoring

00:09:45.419 --> 00:09:47.860
may not be the right way to put it since they

00:09:47.860 --> 00:09:50.379
don't even know they're there. Oh, that's good.

00:09:51.259 --> 00:09:53.679
Ignoring implies some intelligence. You're right.

00:09:53.740 --> 00:09:57.779
There's no intelligence. So that's my guess.

00:09:58.519 --> 00:10:01.919
light of the week, starting out strong. Who wants

00:10:01.919 --> 00:10:03.960
to try and beat me this week? What do you got?

00:10:06.000 --> 00:10:08.500
Oh, nothing. You're all hesitating. Oh, now you

00:10:08.500 --> 00:10:10.059
want to go? I mean, I've got one that's related.

00:10:10.919 --> 00:10:13.480
Fire away. From, you know, there was another,

00:10:13.639 --> 00:10:16.019
an article that came out, I think it was right

00:10:16.019 --> 00:10:17.860
before our podcast last week, but we didn't get

00:10:17.860 --> 00:10:20.299
time to cover it, but it was from Bloomberg on

00:10:20.299 --> 00:10:24.740
people falling asleep and having other types

00:10:24.740 --> 00:10:28.110
of. unconsciousness happen while they're in Waymo's.

00:10:30.690 --> 00:10:36.389
And essentially what's happening there is instead

00:10:36.389 --> 00:10:41.509
of in a normal taxi cab where, you know, or a

00:10:41.509 --> 00:10:44.610
ride share where if someone's drunk and passes

00:10:44.610 --> 00:10:46.870
out, and I'm sure this is what happens, you know,

00:10:46.870 --> 00:10:49.429
probably 90 % of the time or more of these cases,

00:10:49.509 --> 00:10:52.480
instead of someone passing out in the back and

00:10:52.480 --> 00:10:55.019
the driver being able to shake them, wake them

00:10:55.019 --> 00:10:58.139
up, grab them by the collar and drag them out

00:10:58.139 --> 00:11:01.240
of the street, out of the car. What Waymos have

00:11:01.240 --> 00:11:05.559
to do is call the police who automatically have

00:11:05.559 --> 00:11:08.120
to bring an EMS unit because they've got an unconscious

00:11:08.120 --> 00:11:10.659
human in a car and they don't know what type

00:11:10.659 --> 00:11:14.190
of response they needed. And this is happening

00:11:14.190 --> 00:11:17.629
over and over and over and over again in cities

00:11:17.629 --> 00:11:21.370
where this is not just, you know, I'm assuming

00:11:21.370 --> 00:11:23.529
this is not just a Waymo issue. This is going

00:11:23.529 --> 00:11:25.850
to be something that the entire autonomous industry

00:11:25.850 --> 00:11:31.950
has to face. But in the process of, you know,

00:11:31.970 --> 00:11:35.570
the article towards the end, there's a quote

00:11:35.570 --> 00:11:38.590
from a director of product management at Waymo,

00:11:38.710 --> 00:11:41.049
David Marginis, I guess is how you pronounce

00:11:41.049 --> 00:11:45.519
it. He says Waymo's vehicles can offer a safer

00:11:45.519 --> 00:11:48.279
solution in emergencies than a traditional taxi

00:11:48.279 --> 00:11:51.679
or rideshare whose driver may not know how to

00:11:51.679 --> 00:11:59.629
respond. I mean, pretty clearly, if there is

00:11:59.629 --> 00:12:02.269
a medical emergency taking place in a cab or

00:12:02.269 --> 00:12:03.929
Uber, the driver is going to know to call, you

00:12:03.929 --> 00:12:07.309
know, to call emergency services just like a

00:12:07.309 --> 00:12:10.649
Waymo would. OK, there is no advantage to being

00:12:10.649 --> 00:12:13.690
in a Waymo in a medical emergency. This is nonsense.

00:12:14.269 --> 00:12:19.250
I don't know. You know, I don't. It's. Mind -boggling

00:12:19.250 --> 00:12:21.450
how people can say things like this when they're

00:12:21.450 --> 00:12:24.970
trying to defend whatever they think Waymo is

00:12:24.970 --> 00:12:29.389
offering here. But yes, taxi drivers and Uber

00:12:29.389 --> 00:12:32.730
drivers know how to call 911 when there's a problem

00:12:32.730 --> 00:12:35.289
in their vehicle. The best part about it is,

00:12:35.330 --> 00:12:38.490
though, in those 97 % of cases where someone's

00:12:38.490 --> 00:12:40.909
drunk or falling asleep, they can actually do

00:12:40.909 --> 00:12:43.330
something without involving emergency responders

00:12:43.330 --> 00:12:47.330
and, you know, once again requiring... the local

00:12:47.330 --> 00:12:51.769
governments to babysit these cars so um there's

00:12:51.769 --> 00:12:54.029
some there's even more nonsense in the article

00:12:54.029 --> 00:12:58.710
from a guy who actually fell asleep well in quotes

00:12:58.710 --> 00:13:00.629
you didn't get into what really happened to the

00:13:00.629 --> 00:13:03.870
guy but um his quote i'll give him a gas light

00:13:03.870 --> 00:13:07.509
too what the hell he says it feels safer in a

00:13:07.509 --> 00:13:09.629
way with humans you don't know what's going to

00:13:09.629 --> 00:13:14.789
happen but with waymo you do know Oh, you do

00:13:14.789 --> 00:13:18.190
know your video of you sleeping drunk is going

00:13:18.190 --> 00:13:21.610
to be posted on the Internet. Oh, I mean, but

00:13:21.610 --> 00:13:25.389
you don't know. I mean, I don't know. But anyway,

00:13:25.590 --> 00:13:28.690
nonsense, more nonsense from Waymo. Everything's

00:13:28.690 --> 00:13:30.169
going to be better in the future. Right, Fred?

00:13:30.610 --> 00:13:33.250
You betcha. That's what the future is all about.

00:13:33.330 --> 00:13:35.049
I like this. Before you jump in, Fred, I want

00:13:35.049 --> 00:13:36.490
to quote from this art. It's from the Chicago

00:13:36.490 --> 00:13:41.120
Tribune. where you're saying that the guy from

00:13:41.120 --> 00:13:43.139
Waymo said that, oh, it's better and whatnot,

00:13:43.299 --> 00:13:46.440
and that the people know how to, you know, a

00:13:46.440 --> 00:13:49.179
human driver wouldn't know how to call 911. Well,

00:13:49.259 --> 00:13:51.259
a Waymo doesn't know that either, because the

00:13:51.259 --> 00:13:54.360
opening story about this person, ditto Cassander,

00:13:54.559 --> 00:13:56.960
he's apparently, I'm going to quote, his six

00:13:56.960 --> 00:13:59.360
-minute trip had ended nearly an hour before.

00:13:59.639 --> 00:14:03.240
A remote Waymo assistant dialing in through the

00:14:03.240 --> 00:14:05.799
car speakers repeatedly tried to rouse him and

00:14:05.799 --> 00:14:08.850
finally called 911 when he wouldn't stir. So

00:14:08.850 --> 00:14:11.610
the guy's passed out in the car for 54 minutes

00:14:11.610 --> 00:14:15.549
at his destination. And there's no human driver

00:14:15.549 --> 00:14:19.350
who's going to wait that long. I'll call 911

00:14:19.350 --> 00:14:23.330
later. I guess if you didn't have the Waymo Platinum

00:14:23.330 --> 00:14:26.230
service, that's how long it takes for an assistant

00:14:26.230 --> 00:14:29.710
to get on the line and serve you better. Excellent.

00:14:29.789 --> 00:14:32.429
The definition of efficiency. Yeah, yeah. All

00:14:32.429 --> 00:14:36.169
right. Mr. Perkins, back from his journey around

00:14:36.169 --> 00:14:39.840
the world. What do you got for us? I've got,

00:14:39.919 --> 00:14:44.159
this is kind of unique, a repeat contender for

00:14:44.159 --> 00:14:49.379
Gaslight. Jonathan Slotkin, who is a decision.

00:14:50.179 --> 00:14:53.580
Bravo, bravo. Wrote an article called Why We

00:14:53.580 --> 00:14:57.500
Demand Perfect Machines Yet Tolerate Human Carnage.

00:14:58.539 --> 00:15:00.899
In that article, he attempts to make a false

00:15:00.899 --> 00:15:04.220
equivalence between medical technology advancements

00:15:04.220 --> 00:15:07.720
and AV proliferation. So he's got this idea that

00:15:07.720 --> 00:15:10.539
every AV on the road is saving lots of lives.

00:15:11.259 --> 00:15:14.600
He sees resistance to the AV proliferation as

00:15:14.600 --> 00:15:18.539
a perception problem. So if people only understood,

00:15:18.799 --> 00:15:21.240
everything would be great. But the fact is the

00:15:21.240 --> 00:15:26.879
resistance to AVs has far deeper roots. So one

00:15:26.879 --> 00:15:30.879
critical difference is that medicine and advancements

00:15:30.879 --> 00:15:34.059
in medicine have significant third -party requirements

00:15:34.059 --> 00:15:37.480
and certifications. that are completely lacking

00:15:37.480 --> 00:15:40.740
in AV history. You've got the FDA, you've got,

00:15:40.740 --> 00:15:43.559
you know, various organizations, studies, independent

00:15:43.559 --> 00:15:48.159
studies. You've got a requirement for peer review,

00:15:48.360 --> 00:15:51.259
real peer review, not peer review by people that

00:15:51.259 --> 00:15:57.860
you pay. And none of that is applicable to the

00:15:57.860 --> 00:16:01.500
AV proliferation. It's wild west for AVs. They

00:16:01.500 --> 00:16:04.019
can do whatever the hell they want and there's

00:16:04.019 --> 00:16:07.669
no repercussions for them. The medical practitioners

00:16:07.669 --> 00:16:10.570
are licensed, they're certified, they're periodically

00:16:10.570 --> 00:16:13.950
reviewed by independent third parties for competency,

00:16:14.330 --> 00:16:18.470
for currency, and efficacy. None of that takes

00:16:18.470 --> 00:16:22.309
place in the AV industry. The proof of safety

00:16:22.309 --> 00:16:24.669
is a necessary prerequisite for introduction

00:16:24.669 --> 00:16:28.629
of new medical technology. Not so in AVs. Nothing.

00:16:28.830 --> 00:16:30.850
You can throw whatever you want onto the roads.

00:16:32.450 --> 00:16:35.669
Medtechs and professionals must conform to ethical...

00:16:35.960 --> 00:16:39.220
and legal standards and can be sued or legally

00:16:39.220 --> 00:16:42.379
sanctioned for non -conformance. Not so in the

00:16:42.379 --> 00:16:46.200
AV industry. How do you sue an algorithm, even

00:16:46.200 --> 00:16:47.840
if you haven't abandoned your rights? And the

00:16:47.840 --> 00:16:51.500
fact is that to use the AVs, you must contractually

00:16:51.500 --> 00:16:54.059
give up all of your rights to sue them, either

00:16:54.059 --> 00:16:56.440
individually or as part of a class action. That's

00:16:56.440 --> 00:16:59.799
in the terms of service. When I had my knee operated

00:16:59.799 --> 00:17:03.700
on, I had to sign about 120 forms that said I

00:17:03.700 --> 00:17:06.230
knew what was going to happen. I knew why it

00:17:06.230 --> 00:17:08.890
was going to happen. I knew what could happen

00:17:08.890 --> 00:17:13.430
if it went wrong. I had my name and ID checked

00:17:13.430 --> 00:17:16.809
at every point in the process. I had to say five

00:17:16.809 --> 00:17:19.869
times, yes, you're going to operate on my right

00:17:19.869 --> 00:17:23.710
knee. There's a lot of checks and balances that

00:17:23.710 --> 00:17:25.769
are occurring in medical technology that are

00:17:25.769 --> 00:17:30.049
completely absent in AVs. Did you have to sign

00:17:30.049 --> 00:17:34.069
an arbitration clause in there or a clause that

00:17:34.069 --> 00:17:39.980
said? That they could change the contract while

00:17:39.980 --> 00:17:43.000
you were in the midst of surgery. While I was

00:17:43.000 --> 00:17:46.019
under anesthesia? No, they didn't do that. And

00:17:46.019 --> 00:17:49.720
to my understanding, there are even laws that

00:17:49.720 --> 00:17:52.940
apply to application of medical technology to

00:17:52.940 --> 00:17:58.220
the replacement of my knee. So there's a lot

00:17:58.220 --> 00:18:00.220
of fundamental differences here. And I'm surprised

00:18:00.220 --> 00:18:04.460
that this physician who was... smart enough to

00:18:04.460 --> 00:18:06.799
go through medical school and stuff like that,

00:18:06.920 --> 00:18:10.779
simply to hit a dead end with this research that

00:18:10.779 --> 00:18:12.920
says everything's okay. I want to give you an

00:18:12.920 --> 00:18:21.119
example. Recently, there was an instance where

00:18:21.119 --> 00:18:26.319
a Waymo tried to turn down a street where there

00:18:26.319 --> 00:18:29.480
was an exit from a school and a fire drill going

00:18:29.480 --> 00:18:32.299
on. And so there were hundreds of... Kids in

00:18:32.299 --> 00:18:34.920
the street. There were teachers in the street.

00:18:35.140 --> 00:18:37.519
And the vice principal had to actually stand

00:18:37.519 --> 00:18:40.839
in front of the Waymo to keep it from turning

00:18:40.839 --> 00:18:44.640
into this crowd of people, young kids, who were

00:18:44.640 --> 00:18:49.440
being disgorged into the street. So, you know,

00:18:49.460 --> 00:18:54.299
how do you train a system to do that? I'll get

00:18:54.299 --> 00:18:56.240
into more of this later, so I'll let that go.

00:18:56.500 --> 00:19:01.390
But the fundamental difference is. Fundamental

00:19:01.390 --> 00:19:06.809
differences. So, Slotkin also relies on industry

00:19:06.809 --> 00:19:09.809
-provided safety data that's not been reviewed

00:19:09.809 --> 00:19:13.049
by third parties. And Waymo, at the time he wrote

00:19:13.049 --> 00:19:16.630
this, apparently was still asserting that 94

00:19:16.630 --> 00:19:20.950
% fewer accidents by Waymo than humans. But they've

00:19:20.950 --> 00:19:23.869
recently gone to 75 % fewer mishaps. So, this

00:19:23.869 --> 00:19:32.109
is a moving goalpost, I guess. And Dr. Slotkin

00:19:32.109 --> 00:19:35.809
may not have been aware of just how fluid these

00:19:35.809 --> 00:19:39.910
goals are. So the bottom line is that even if

00:19:39.910 --> 00:19:43.029
AVs eventually can prove comparative safety benefits,

00:19:43.710 --> 00:19:46.710
many, many more lives will be saved by migrating

00:19:46.710 --> 00:19:50.170
that proven life -saving technology into conventional

00:19:50.170 --> 00:19:54.329
cars as quickly as possible. Preserving the tech

00:19:54.329 --> 00:19:56.869
for AVs costs hundreds of thousands of needless

00:19:56.869 --> 00:20:00.769
deaths. over the next 25 years. So what he's

00:20:00.769 --> 00:20:03.170
asserting as a way to save lives is actually

00:20:03.170 --> 00:20:09.390
costing lots of lives in the future. It's just

00:20:09.390 --> 00:20:13.109
a real misapplication of capital and technology

00:20:13.109 --> 00:20:17.230
if your goal is saving lives. So for misdirecting

00:20:17.230 --> 00:20:19.809
the appropriate focus for a life -saving automotive

00:20:19.809 --> 00:20:23.170
technology, Jonathan Slotkin earns my Gaslight

00:20:23.170 --> 00:20:28.069
nomination this week. And he's a repeat. Nominee,

00:20:28.069 --> 00:20:32.049
one of the few that we've got. That's an excellent

00:20:32.049 --> 00:20:35.309
one. I would like to point out that we were first

00:20:35.309 --> 00:20:38.549
introduced to Dr. Slotkin from an op -ed he wrote

00:20:38.549 --> 00:20:41.670
in the New York Times. After that, he got a lot

00:20:41.670 --> 00:20:45.910
of backlash, feedback, questions about what nonsense

00:20:45.910 --> 00:20:48.670
are you writing? I had a question for him that

00:20:48.670 --> 00:20:50.950
he promptly responded. My question was, what

00:20:50.950 --> 00:20:53.130
kind of funding do you receive from Alphabet,

00:20:53.309 --> 00:20:55.630
Google, Waymo? And he said immediately, I receive

00:20:55.630 --> 00:20:57.920
nothing from them. So in this article, which

00:20:57.920 --> 00:21:00.019
I will point out is not in the New York Times,

00:21:00.059 --> 00:21:06.000
it is in some online blog called Noema, which

00:21:06.000 --> 00:21:08.960
prior to this, I mean, I'm not familiar with

00:21:08.960 --> 00:21:12.259
all publications, but all of us are like, huh?

00:21:12.700 --> 00:21:18.150
What? So how far he's fallen. Possibly due to

00:21:18.150 --> 00:21:20.430
feedback and questions. Who is original? No.

00:21:20.569 --> 00:21:25.650
Op -ed piece. No, no. And this one, he says again

00:21:25.650 --> 00:21:28.549
that, you know, I received no funding from Alphabet

00:21:28.549 --> 00:21:32.390
whatsoever. But I do spend a lot of time talking

00:21:32.390 --> 00:21:36.150
to Waymo's Trent Victor. And he and I'm basically

00:21:36.150 --> 00:21:38.410
whatever he tells me, I'm writing down verbatim.

00:21:38.650 --> 00:21:41.009
I believe the Russians have a term for this kind

00:21:41.009 --> 00:21:45.309
of behavior. It's called useful idiot. But. This

00:21:45.309 --> 00:21:47.509
is just my humble opinion. And again, this is

00:21:47.509 --> 00:21:50.470
a doctor who wants to have a machine take over

00:21:50.470 --> 00:21:54.309
human operations. This is a doctor who wants

00:21:54.309 --> 00:22:01.069
to remove the human from the loop and have the

00:22:01.069 --> 00:22:06.109
computer take over the operation. Isn't he busy

00:22:06.109 --> 00:22:08.920
enough? I mean, really? I don't know. There's

00:22:08.920 --> 00:22:10.519
a lot of people out there who need medicine.

00:22:10.640 --> 00:22:13.799
He's clearly handed over his entire job to an

00:22:13.799 --> 00:22:16.720
algorithm because some doctors make mistakes.

00:22:17.519 --> 00:22:21.019
It's all a wonderful life, that poor pharmacist.

00:22:22.299 --> 00:22:24.599
So what happens when a doctor falls asleep in

00:22:24.599 --> 00:22:27.559
the middle of surgery? Do they have to call 911

00:22:27.559 --> 00:22:31.440
or do they call Waymo or what do they do? Please

00:22:31.440 --> 00:22:34.589
stay on the line. Yeah, I think the first call

00:22:34.589 --> 00:22:36.569
should be to their attorney. Your safety is our

00:22:36.569 --> 00:22:39.210
highest priority. I've had a couple of surgeries

00:22:39.210 --> 00:22:41.329
in my day and it's generally not, I've never

00:22:41.329 --> 00:22:43.250
had it where there's one doctor in the room.

00:22:43.950 --> 00:22:45.849
There's, you know, and these are not complex

00:22:45.849 --> 00:22:48.849
things. There's, you know, the doctor, the doctor's

00:22:48.849 --> 00:22:51.089
assistant, the anesthetologist, a couple nurses.

00:22:52.809 --> 00:22:55.950
But hey, in Jonathan's world, we just need the

00:22:55.950 --> 00:22:58.930
algorithm, man. I am not paid by Alphabet. I'm

00:22:58.930 --> 00:23:01.430
not. Maybe that's what he's maybe this is his

00:23:01.430 --> 00:23:03.509
thing is I'm not paid by Alphabet. Alphabet,

00:23:03.569 --> 00:23:06.309
please pay me. Come on. Come on. Just pay me

00:23:06.309 --> 00:23:09.009
already. Look what I'm doing for you. Well, I

00:23:09.009 --> 00:23:11.089
hope he hears this and elects to come on the

00:23:11.089 --> 00:23:14.230
show because I'm sure he's a much better and

00:23:14.230 --> 00:23:18.069
more involved person than the caricature that

00:23:18.069 --> 00:23:24.109
we're painting here would. Yeah, I love a lot

00:23:24.109 --> 00:23:26.960
of the things that, you know. He says in articles,

00:23:27.140 --> 00:23:29.079
because he's not saying just throw Waymo's on

00:23:29.079 --> 00:23:32.140
the street. He's saying we need data. We need,

00:23:32.140 --> 00:23:34.119
you know, a lot of the things we line up with

00:23:34.119 --> 00:23:38.880
Dr. Slotkin perfectly on. It's just this aspirational

00:23:38.880 --> 00:23:42.319
thinking that it's inevitable that autonomous

00:23:42.319 --> 00:23:45.559
vehicles are going to be safer and save a bunch

00:23:45.559 --> 00:23:49.559
of lives when we know that there's other technologies

00:23:49.559 --> 00:23:51.980
that need to be going in cars now that the auto

00:23:51.980 --> 00:23:56.750
industry is totally delaying. that could save

00:23:56.750 --> 00:23:59.869
a lot more lives before autonomous vehicles are

00:23:59.869 --> 00:24:02.549
a practical reality for most Americans. Well,

00:24:02.609 --> 00:24:04.549
a lot more. And beyond that, there's the issue

00:24:04.549 --> 00:24:07.569
of getting people out of traditional cars. People

00:24:07.569 --> 00:24:09.809
aren't just going to happily jump out of their

00:24:09.809 --> 00:24:12.849
self -driven car. I mean, the car they drive

00:24:12.849 --> 00:24:15.430
into an autonomous vehicle overnight. There's

00:24:15.430 --> 00:24:19.140
going to be a lot of resistance to that. are

00:24:19.140 --> 00:24:21.900
it can be problematic when you try to take some

00:24:21.900 --> 00:24:24.140
of their freedoms away from them and a lot of

00:24:24.140 --> 00:24:27.460
people consider driving uh a freedom well i look

00:24:27.460 --> 00:24:30.319
forward to his uh next op -ed in the safeway

00:24:30.319 --> 00:24:33.519
shopper uh because that's the trajectory get

00:24:33.519 --> 00:24:36.039
up it's your piggly wiggly hey i don't know if

00:24:36.039 --> 00:24:39.339
he's piggly wiggly ready but this week uh scoring

00:24:39.339 --> 00:24:42.660
of the gaslight nominees uh we've done the tabulation

00:24:42.660 --> 00:24:48.980
and uh i won it's uh awesome That's what happens

00:24:48.980 --> 00:24:52.420
when you trust an algorithm. What happens over

00:24:52.420 --> 00:24:55.380
time when Anthony has just lost all credibility

00:24:55.380 --> 00:24:57.880
as a judge, I just accept what he says and we

00:24:57.880 --> 00:25:00.119
move on. Hey, no, I did extra homework. That's

00:25:00.119 --> 00:25:03.519
why I won. I did the Waymo first responder certification

00:25:03.519 --> 00:25:07.380
course. Yeah, okay. Come on. Michael, you were

00:25:07.380 --> 00:25:09.640
close second because we laughed out loud. If

00:25:09.640 --> 00:25:11.660
there was a trophy or something here, I might

00:25:11.660 --> 00:25:14.759
get more concerned about it, but there's no prize.

00:25:15.200 --> 00:25:19.180
Wait, what? There's no prize? Oh, hey, listeners,

00:25:19.259 --> 00:25:21.039
there is a prize. If you go to autosafety .org

00:25:21.039 --> 00:25:24.640
and click on donate, the prize is your happiness

00:25:24.640 --> 00:25:27.380
in the world. And you feel good as a person.

00:25:27.380 --> 00:25:30.720
Does this mean you're not certifiable? I am clearly

00:25:30.720 --> 00:25:33.019
certifiable. Are you talking to my doctor, Jonathan

00:25:33.019 --> 00:25:39.319
Slotkin? Anyway, let's move on. Damn it, Tesla.

00:25:39.940 --> 00:25:43.140
I know. I'm sorry. I'm sorry, everybody. But

00:25:43.140 --> 00:25:46.319
we have to Tesla. They still exist. So we still

00:25:46.319 --> 00:25:49.519
have to do it. This is from Electrek. Tesla remote

00:25:49.519 --> 00:25:52.759
operator crashed a robo -taxi in Houston, NHTSA

00:25:52.759 --> 00:25:55.599
data shows. So in the past, for those who have

00:25:55.599 --> 00:25:58.759
not been paying attention, Tesla has to file

00:25:58.759 --> 00:26:01.359
all these reports with NHTSA, but all of them,

00:26:01.380 --> 00:26:04.099
they're redacted for proprietary reasons, and

00:26:04.099 --> 00:26:07.339
so they don't look super embarrassed. And so

00:26:07.339 --> 00:26:09.960
we've not really had access to anything of what's

00:26:09.960 --> 00:26:12.619
in these reports, but for some reason now...

00:26:12.619 --> 00:26:15.900
Recently, they started... Just on the robo -taxis,

00:26:15.900 --> 00:26:19.720
though. Not on all their ADASV crashes with full

00:26:19.720 --> 00:26:21.839
self -driving autopilot type crashes. They're

00:26:21.839 --> 00:26:24.059
still hiding that stuff. Yeah, but somehow the

00:26:24.059 --> 00:26:28.440
robo -taxis, that's coming out. I think it's

00:26:28.440 --> 00:26:32.200
because they want to look like they're being

00:26:32.200 --> 00:26:35.599
as transparent as the other robo -taxi operators.

00:26:35.940 --> 00:26:38.240
Right. So I'm going to quote from this article.

00:26:38.460 --> 00:26:42.920
This is from... Tesla's own narrative. While

00:26:42.920 --> 00:26:45.799
the ADS was being supported remotely out of the

00:26:45.799 --> 00:26:48.460
dead end road, the vehicle entered a grassy slope.

00:26:48.720 --> 00:26:51.319
As the remote assistant operator continued to

00:26:51.319 --> 00:26:53.420
recover the vehicle, they made contact with a

00:26:53.420 --> 00:26:57.759
hidden tree stump. Those damn hidden tree stumps

00:26:57.759 --> 00:27:02.859
always hiding out everywhere. This is remote

00:27:02.859 --> 00:27:06.980
operators. We've discussed this with Missy Cummings.

00:27:06.980 --> 00:27:10.690
We've discussed this with Phil Kopman. This is

00:27:10.690 --> 00:27:14.269
a horrible idea. Like, this is a horrible, horrible

00:27:14.269 --> 00:27:18.109
idea. What? Wait, what? Wait, what? Remote operators

00:27:18.109 --> 00:27:21.950
driving, taking control of the vehicles. I mean,

00:27:21.970 --> 00:27:24.710
to move it out of the way. Am I wrong? I mean,

00:27:24.710 --> 00:27:26.930
we discussed last week that remote drivers are

00:27:26.930 --> 00:27:30.190
almost necessary at this point. They are needed.

00:27:30.210 --> 00:27:34.390
The remote supervisors. Right. Remote supervisors.

00:27:34.470 --> 00:27:37.940
That's different. The remote driver is. Because

00:27:37.940 --> 00:27:41.880
of latency issues and because of those nasty

00:27:41.880 --> 00:27:46.559
hidden tree stumps. Yeah, and this is just evidence

00:27:46.559 --> 00:27:51.119
that when the vehicle isn't perfect, the fallback

00:27:51.119 --> 00:27:54.559
remote driver, you can't totally rely on them

00:27:54.559 --> 00:27:58.220
either. And this points out that maybe Tesla's

00:27:58.220 --> 00:28:01.359
camera -only system, if it can't see a tree stump,

00:28:01.579 --> 00:28:04.480
maybe they need a LiDAR. Because I think a LiDAR

00:28:04.480 --> 00:28:08.200
would see that nasty hidden tree stump. But Michael,

00:28:08.220 --> 00:28:12.759
I mean, leaving leaving a road is kind of an

00:28:12.759 --> 00:28:18.279
issue there is a grassy slope. OK, but so, Michael,

00:28:18.299 --> 00:28:22.079
is there any more insight on why Tesla has started

00:28:22.079 --> 00:28:23.880
releasing this information? Are we just guessing

00:28:23.880 --> 00:28:26.460
that they want to appear transparent? I just

00:28:26.460 --> 00:28:29.000
I just that's the only thing I can think of.

00:28:29.039 --> 00:28:32.039
But, you know, I think also when you dive into

00:28:32.039 --> 00:28:36.730
the data and I think that. Tesla is having significantly

00:28:36.730 --> 00:28:40.809
more incidents that they're having to report

00:28:40.809 --> 00:28:44.089
per mile travel than Waymo and some of the other

00:28:44.089 --> 00:28:49.180
autonomous vehicle operators. Okay. Because we

00:28:49.180 --> 00:28:52.119
know that they're only operating like, what is

00:28:52.119 --> 00:28:55.119
it, 50 vehicles or something. Right. Whereas

00:28:55.119 --> 00:28:58.759
Wayne was operating 3000 and Tesla is reporting,

00:28:59.019 --> 00:29:01.099
you know, I don't know how many crashes they

00:29:01.099 --> 00:29:03.920
reported around 20 or something. That's a lot

00:29:03.920 --> 00:29:07.079
in a small fleet like that. Right. In a small

00:29:07.079 --> 00:29:11.700
area, too. Yeah, I mean, they're having problems.

00:29:12.019 --> 00:29:14.579
And, you know, I don't think they're until they

00:29:14.579 --> 00:29:17.519
change their design and their sensor system and

00:29:17.519 --> 00:29:20.980
just a lot of things. I just don't think I just

00:29:20.980 --> 00:29:24.319
still don't see Tesla as a viable competitor

00:29:24.319 --> 00:29:28.099
with Waymo or really any of the companies who

00:29:28.099 --> 00:29:33.460
are in this space at the moment. Right. So since

00:29:33.460 --> 00:29:35.099
it's been a few minutes, we've talked about this

00:29:35.099 --> 00:29:40.009
company. Let's go talk about Waymo. I know. So

00:29:40.009 --> 00:29:42.910
Waymo is really trying to get out into D .C.

00:29:43.269 --> 00:29:46.329
and they really want in. And a few years ago,

00:29:46.410 --> 00:29:48.250
I guess the D .C. government was like, yeah,

00:29:48.289 --> 00:29:50.829
go ahead. Let your test cars out on the street.

00:29:50.950 --> 00:29:53.069
And now Waymo is like, we want to offer commercial

00:29:53.069 --> 00:29:56.869
service. And that's getting a lot of pushback.

00:29:57.269 --> 00:29:58.910
Washington Post has an article and I'm going

00:29:58.910 --> 00:30:01.289
to quote from it. Missy Cummings, a George Mason

00:30:01.289 --> 00:30:04.170
University professor, researches autonomous systems,

00:30:04.230 --> 00:30:06.309
has raised concern about D .C. roadway challenges

00:30:06.309 --> 00:30:09.329
such as frequent motorcades. One of the things

00:30:09.329 --> 00:30:11.349
that we see repeatedly, these cars really struggle

00:30:11.349 --> 00:30:14.250
around rotating first responder lights. That's

00:30:14.250 --> 00:30:16.950
a defining characteristics of these motorcades.

00:30:17.609 --> 00:30:20.829
That's an excellent point we've discussed. Further

00:30:20.829 --> 00:30:24.670
down, this is where Waymo really leans into their

00:30:24.670 --> 00:30:26.750
stuff. Business leaders in Washington have expressed

00:30:26.750 --> 00:30:29.450
support for driverless cars. So too have some

00:30:29.450 --> 00:30:31.769
members of the blind community. Claire Stanley

00:30:31.769 --> 00:30:33.490
said in an interview that she has lost track

00:30:33.490 --> 00:30:35.589
of how many times a driver has declined to give

00:30:35.589 --> 00:30:37.569
her a ride because of her guide dog, even though

00:30:37.569 --> 00:30:39.369
federal law and ride handling policy is supposed

00:30:39.369 --> 00:30:41.809
to prevent such refusals. The ability to be in

00:30:41.809 --> 00:30:43.710
a vehicle where there's no human being to fight

00:30:43.710 --> 00:30:46.670
that battle with is amazing, said Stanley, director

00:30:46.670 --> 00:30:48.230
of advocacy and governmental affairs for the

00:30:48.230 --> 00:30:50.670
American Council on the Blind. Now, having lived

00:30:50.670 --> 00:30:54.549
in D .C. and having been refused repeated service

00:30:54.549 --> 00:30:58.410
in taxi cabs and not even have a dog, I understand.

00:30:59.210 --> 00:31:00.730
They just didn't want to drive me to Southeast

00:31:00.730 --> 00:31:05.970
where I lived. So this, I mean, this is a legit

00:31:05.970 --> 00:31:08.990
problem, but the Waymo is not the solution, is

00:31:08.990 --> 00:31:11.589
it? It seems like a pretty heavy handed response,

00:31:11.789 --> 00:31:16.150
but like, hey, isn't just basic enforcement better?

00:31:17.089 --> 00:31:21.029
Am I naive? Well, I mean, I don't know. I don't

00:31:21.029 --> 00:31:22.750
know what the enforcement of that's like in D

00:31:22.750 --> 00:31:26.450
.C. I know that. I mean, I would get refused

00:31:26.450 --> 00:31:29.650
a lot when I was trying to. get a cab out to

00:31:29.650 --> 00:31:32.789
Virginia from, from taxi cabs. And even though

00:31:32.789 --> 00:31:35.509
it's illegal or they can be getting a lot of

00:31:35.509 --> 00:31:37.950
trouble from the taxi commission for doing that,

00:31:38.009 --> 00:31:40.250
they just don't care because it's not enforced.

00:31:40.509 --> 00:31:44.089
So, I mean, it is a solution. And I think that's

00:31:44.089 --> 00:31:45.750
something we've heard. I think I heard that from

00:31:45.750 --> 00:31:51.049
some folks with the Chicago transportation authority.

00:31:51.109 --> 00:31:54.119
And they were saying that. they really, really

00:31:54.119 --> 00:31:57.180
are looking forward to autonomous vehicles because

00:31:57.180 --> 00:31:59.460
it's going to solve some of the equity issues

00:31:59.460 --> 00:32:02.799
around taxis and people not getting rides in

00:32:02.799 --> 00:32:05.619
the poor neighborhoods in town. And I can certainly

00:32:05.619 --> 00:32:08.579
see that as an advantage. The only issue is,

00:32:08.680 --> 00:32:11.079
you know, obviously, we don't think they're safe

00:32:11.079 --> 00:32:14.279
enough to be doing that yet. Right. Well, the

00:32:14.279 --> 00:32:17.059
other thing is, we've not heard from the blind

00:32:17.059 --> 00:32:22.680
community in any area where the Robotaxis are

00:32:22.680 --> 00:32:26.099
currently operating, that, yeah, it's working

00:32:26.099 --> 00:32:32.180
out great. You know, if people compare the aspirations

00:32:32.180 --> 00:32:35.380
to what they want to happen in the future, sure,

00:32:35.480 --> 00:32:37.819
everything will be great. I mean, Trump got elected

00:32:37.819 --> 00:32:45.039
on that basis, right? And there's no reality

00:32:45.039 --> 00:32:49.059
behind the idea that the blind community is really

00:32:49.059 --> 00:32:56.200
going to benefit. From these way most. And even

00:32:56.200 --> 00:32:59.559
I think say they say they can, I think that we've

00:32:59.559 --> 00:33:01.140
heard from a lot of folks in the blind community

00:33:01.140 --> 00:33:03.200
that it really won't that independence won't

00:33:03.200 --> 00:33:06.099
this, even if that comes to fruition, should

00:33:06.099 --> 00:33:08.539
we be putting other people's flies at risk on

00:33:08.539 --> 00:33:10.539
the road to make that a reality for the blind

00:33:10.539 --> 00:33:14.460
community today? I say, no, I say, I mean, I

00:33:14.460 --> 00:33:18.380
mean, once these, once if. If these vehicles

00:33:18.380 --> 00:33:21.900
become safe enough that we can trust them on

00:33:21.900 --> 00:33:24.799
our roads, then yeah, I think they could certainly

00:33:24.799 --> 00:33:28.160
be a benefit there. But at this point, we still

00:33:28.160 --> 00:33:32.119
have to do a cost -benefit analysis. And the

00:33:32.119 --> 00:33:34.859
tradeoff of independence for the blind versus

00:33:34.859 --> 00:33:37.740
people being killed by autonomous vehicles because

00:33:37.740 --> 00:33:39.819
they've been on the roads too early is not acceptable.

00:33:40.359 --> 00:33:42.279
I want to push back slightly on one thing you

00:33:42.279 --> 00:33:43.940
said. You said we've heard a lot from people

00:33:43.940 --> 00:33:45.779
in the blind community. We haven't heard from

00:33:45.779 --> 00:33:47.180
a lot of people in the blind community. We've

00:33:47.180 --> 00:33:52.180
heard a lot from people who are part of the American

00:33:52.180 --> 00:33:57.220
Council of the Blind. And even more from the

00:33:57.220 --> 00:33:58.940
Blind Federation. The Blind Federation of America.

00:33:59.140 --> 00:34:00.920
We've heard a lot from those two nonprofits.

00:34:01.220 --> 00:34:04.319
And unlike Jonathan Slotkin, they are very open

00:34:04.319 --> 00:34:07.980
that they receive money from Alphabet. Really?

00:34:08.280 --> 00:34:10.780
Yeah, they don't. I mean, I always I always assumed

00:34:10.780 --> 00:34:12.860
that they did because they were showing up at

00:34:12.860 --> 00:34:15.460
NHTSA advocating for years for autonomous vehicles.

00:34:15.619 --> 00:34:17.659
But I've never seen that. Yes. Correct me if

00:34:17.659 --> 00:34:19.440
I'm wrong, but I'm pretty sure that is public

00:34:19.440 --> 00:34:23.119
knowledge that that that's wide open. So, you

00:34:23.119 --> 00:34:27.619
know, we've we've heard from from that. I don't

00:34:27.619 --> 00:34:31.360
know. I think I think Fred's got a legit point.

00:34:31.460 --> 00:34:33.400
We need more data on this. And Michael, I think

00:34:33.400 --> 00:34:36.119
you have a legit point that what I could be saying

00:34:36.119 --> 00:34:38.940
could be wrong. I think there's a tremendous

00:34:38.940 --> 00:34:41.659
vulnerability. I know they partner with them,

00:34:41.739 --> 00:34:46.440
but I don't have any direct proof that they're

00:34:46.440 --> 00:34:48.559
funding them, but I would assume that's going

00:34:48.559 --> 00:34:52.500
on in some respect. I think there's a lot of

00:34:52.500 --> 00:34:56.179
vulnerability that has never been discussed associated

00:34:56.179 --> 00:34:59.920
with blind community using Waymos. There's a

00:34:59.920 --> 00:35:03.000
lot of people who are sporting fake Ubers and

00:35:03.000 --> 00:35:05.630
who are picking up people. under the pretense

00:35:05.630 --> 00:35:07.869
that they're actually legitimate Uber drivers,

00:35:08.090 --> 00:35:10.369
and then, you know, extorting money and doing

00:35:10.369 --> 00:35:13.050
terrible things to them. There's no reason to

00:35:13.050 --> 00:35:17.309
expect any different with Waymos. If people are

00:35:17.309 --> 00:35:20.329
unable to see the vehicle that they're being

00:35:20.329 --> 00:35:25.309
steered to, why is there a belief that they will

00:35:25.309 --> 00:35:28.369
never be taken advantage of and steered into

00:35:28.369 --> 00:35:31.889
a vehicle under false premises? You know, if

00:35:31.889 --> 00:35:36.559
you call the... You call a Waymo. If you're a

00:35:36.559 --> 00:35:40.519
bad person, you see a vehicle pull up. The blind

00:35:40.519 --> 00:35:45.159
person walks towards the vehicle. And you could

00:35:45.159 --> 00:35:47.239
be mistaken for a good Samaritan who says, no,

00:35:47.360 --> 00:35:49.840
no, no, that's the wrong vehicle. Go into this

00:35:49.840 --> 00:35:52.940
other vehicle. So I think there's as many vulnerabilities

00:35:52.940 --> 00:35:56.460
associated with the blind community using Waymos

00:35:56.460 --> 00:35:59.719
as there are potential solutions. I don't think

00:35:59.719 --> 00:36:02.079
anybody's addressed that in any kind of detail.

00:36:02.920 --> 00:36:06.079
And just to do a sanity check, I did pull up

00:36:06.079 --> 00:36:08.940
a document from Google itself saying these are

00:36:08.940 --> 00:36:11.400
the list of politically engaged trade associations,

00:36:11.659 --> 00:36:13.239
independent third parties and other tax them

00:36:13.239 --> 00:36:15.300
groups that receive the most substantial contribution

00:36:15.300 --> 00:36:17.519
from Google's U .S. government affairs and public

00:36:17.519 --> 00:36:20.119
policy team. And both the National Federation

00:36:20.119 --> 00:36:21.639
of the Blind and the American Council of the

00:36:21.639 --> 00:36:24.380
Blind are listed as recipients. Who else is on

00:36:24.380 --> 00:36:27.079
there? You got a ton of people, but you know

00:36:27.079 --> 00:36:29.179
who's not on there? The Center for Auto Safety.

00:36:29.769 --> 00:36:31.670
That's right, because we don't take those dirty

00:36:31.670 --> 00:36:34.369
dollars. But we'll take them from you, because

00:36:34.369 --> 00:36:36.269
they're not dirty. Go to autosafety .org, click

00:36:36.269 --> 00:36:39.769
on Donate. Wow. Yeah. Before we jump into the

00:36:39.769 --> 00:36:42.050
town, we're going to continue with this somewhat

00:36:42.050 --> 00:36:45.690
related. So Waymo, they really want the Waymos

00:36:45.690 --> 00:36:48.789
out in D .C., running Ford. And now Waymo works

00:36:48.789 --> 00:36:52.429
with Uber, or Uber is like, hey, let's get Waymo

00:36:52.429 --> 00:36:54.550
vehicles. But people use our app to do it, so

00:36:54.550 --> 00:36:59.289
we get a cut. But Waymo is against. Sorry, Uber.

00:36:59.429 --> 00:37:02.730
It's confusing. Uber is against Waymo's proposed

00:37:02.730 --> 00:37:05.010
bill in D .C. saying, yeah, let this come out.

00:37:05.070 --> 00:37:08.070
Instead, Uber, quoting from an article we have

00:37:08.070 --> 00:37:10.710
linked from MSN, Uber, which opposes the bill,

00:37:10.769 --> 00:37:12.710
argues that the proposed rule would displace

00:37:12.710 --> 00:37:15.570
four higher human drivers and hand Waymo a de

00:37:15.570 --> 00:37:18.210
facto monopoly. It is lobbied instead for a system

00:37:18.210 --> 00:37:21.110
that require robo -taxis to operate on a ride

00:37:21.110 --> 00:37:23.849
-hailing network that also uses human drivers.

00:37:24.920 --> 00:37:26.800
According to public records viewed by TechCrunch

00:37:26.800 --> 00:37:28.159
interviews with industry and company sources.

00:37:28.659 --> 00:37:33.099
So, yeah, Michael, you're right. This could have

00:37:33.099 --> 00:37:35.119
been a gaslighting. I mean, anybody can see what's

00:37:35.119 --> 00:37:38.940
going on. Uber doesn't want Waymo taking away

00:37:38.940 --> 00:37:42.019
a segment of its market, period. So what I imagine

00:37:42.019 --> 00:37:44.940
is. And they've partnered, like they've partnered

00:37:44.940 --> 00:37:47.639
for a while. I think, I don't even, I mean, is

00:37:47.639 --> 00:37:51.440
our Waymo still available in the Uber app in

00:37:51.440 --> 00:37:53.159
certain places? I thought they were. I think

00:37:53.159 --> 00:37:56.460
they are. And here they are. Here they are at

00:37:56.460 --> 00:38:00.900
loggerheads over over these issues. So that's

00:38:00.900 --> 00:38:03.260
interesting. Hate me cynical, but I think Uber

00:38:03.260 --> 00:38:04.980
is just going to be like, look, we're going to

00:38:04.980 --> 00:38:08.320
have one human driver. That's all we need. And

00:38:08.320 --> 00:38:10.019
then you have to funnel it all through this and

00:38:10.019 --> 00:38:12.139
we'll give Waymo a little bit more percentage.

00:38:13.039 --> 00:38:15.679
You know, I hate to point this out, but Uber

00:38:15.679 --> 00:38:19.880
had one human driver in the only fatal, directly

00:38:19.880 --> 00:38:22.659
fatal autonomous collision there has been. That's

00:38:22.659 --> 00:38:26.469
true. It's hard to believe anything that comes

00:38:26.469 --> 00:38:28.809
out of the mouth of Uber, frankly. I mean, I

00:38:28.809 --> 00:38:33.250
know we kill Waymo a lot here, but Uber is...

00:38:33.250 --> 00:38:39.869
A German word. How do you say? Hey, and with

00:38:39.869 --> 00:38:42.530
that... It's hard to trust them. How do you like

00:38:42.530 --> 00:38:44.590
this for a transition into the towel, Fred? Is

00:38:44.590 --> 00:38:47.690
that a nice, soft transition? Or does this remind

00:38:47.690 --> 00:38:50.530
you of Casey Kasem years ago transitioning to

00:38:50.530 --> 00:38:52.750
a birthday dedication or a dead dog dedication?

00:38:53.360 --> 00:38:55.599
You know, I never heard Casey Kasem. I lived

00:38:55.599 --> 00:38:59.639
out in the sticks. So I'm a little handicapped

00:38:59.639 --> 00:39:02.079
with that analogy. Well, someone just people

00:39:02.079 --> 00:39:05.840
go online and Google search for Casey Kasem dead

00:39:05.840 --> 00:39:09.139
dog dedication. It's fascinating. That's an oldie

00:39:09.139 --> 00:39:12.119
from the Internet. Anyway, go on, Fred. All right.

00:39:12.360 --> 00:39:21.409
So in my own. Fred missed last week on our discussion

00:39:21.409 --> 00:39:23.750
on edge cases. So he's going to go back and tell

00:39:23.750 --> 00:39:25.750
us where we went wrong. Oh, right. So, Fred,

00:39:25.829 --> 00:39:27.710
keep in mind, these are big shoes to fill because

00:39:27.710 --> 00:39:31.329
we had, you know, Professor Amaretis. I don't

00:39:31.329 --> 00:39:33.690
know what his title is anymore, but from Carnegie

00:39:33.690 --> 00:39:36.369
Mellon, Phil Koopman. Yeah, you had all the heavy

00:39:36.369 --> 00:39:38.809
hitters on. So, you know. You're a man wearing

00:39:38.809 --> 00:39:43.929
a headband. Yeah. It's the 60s all over again.

00:39:44.789 --> 00:39:48.329
So. I have a problem with the term edge cases.

00:39:48.869 --> 00:39:53.510
And I know that my distinguished colleagues and

00:39:53.510 --> 00:39:57.250
those people who would not deign to call me a

00:39:57.250 --> 00:40:00.190
colleague as well use the term, loosely defined

00:40:00.190 --> 00:40:03.530
as a negative or dangerous AV operational event

00:40:03.530 --> 00:40:06.989
for which the AV machine learning algorithms

00:40:06.989 --> 00:40:09.889
have not been trained. The term has crept into

00:40:09.889 --> 00:40:13.750
the language without really addressing what the

00:40:13.750 --> 00:40:17.329
consequences of that are. The problem that I

00:40:17.329 --> 00:40:20.170
have, one problem that I have is that the definition

00:40:20.170 --> 00:40:22.989
really comes with the underlying assumption that

00:40:22.989 --> 00:40:25.889
the control algorithms could be trained for that

00:40:25.889 --> 00:40:30.329
edge case event. But that's just not so. There

00:40:30.329 --> 00:40:34.090
are so many edge case, what you could call edge

00:40:34.090 --> 00:40:36.869
case events or potentially edge case events,

00:40:37.070 --> 00:40:40.989
that you can, there's not enough time and money

00:40:40.989 --> 00:40:44.010
in the world to validate your response to all

00:40:44.010 --> 00:40:51.679
of those. unhappy events. There are so many.

00:40:51.940 --> 00:40:55.500
And you only know that edge cases are going to

00:40:55.500 --> 00:40:59.199
be a problem once they occur. So it becomes podological,

00:40:59.360 --> 00:41:02.139
right? What's an edge case? Well, an edge case

00:41:02.139 --> 00:41:04.980
is bad. How do you know it's bad? Well, because

00:41:04.980 --> 00:41:07.539
it occurred. Yeah, I think that's one of the

00:41:07.539 --> 00:41:10.280
major problems with trying to build safety standards

00:41:10.280 --> 00:41:13.449
that can... You know, if you say I'm going to

00:41:13.449 --> 00:41:15.309
build a safety standard that eliminates edge

00:41:15.309 --> 00:41:18.869
cases, that's almost nonsense. I think because

00:41:18.869 --> 00:41:22.849
how do you do that? Is there a way to see the

00:41:22.849 --> 00:41:25.389
unseen or to know what's unknowable that we don't

00:41:25.389 --> 00:41:28.349
know yet? It's it's it seems like a really, really

00:41:28.349 --> 00:41:30.909
tough thing to do. Well, yeah. And, you know,

00:41:30.929 --> 00:41:33.809
I've done some number crunching from time to

00:41:33.809 --> 00:41:36.349
time that shows that there are literally millions

00:41:36.349 --> 00:41:39.710
or billions of combinations of things that could

00:41:39.710 --> 00:41:44.340
occur. using only a small set of safety -critical

00:41:44.340 --> 00:41:50.000
phenomena that are possible. And if you've got

00:41:50.000 --> 00:41:52.840
things that are likely to occur in the millions,

00:41:53.119 --> 00:41:56.400
how do you establish all of those and how do

00:41:56.400 --> 00:42:01.960
you train for them? There's a couple of fundamental

00:42:01.960 --> 00:42:05.780
problems. One problem is that we tend to anthropomorphize

00:42:05.780 --> 00:42:08.780
the computer, thinking it thinks like us and

00:42:08.780 --> 00:42:11.750
does things like us. But that's not how it works.

00:42:12.489 --> 00:42:14.670
They're sort of the inverse of each other because

00:42:14.670 --> 00:42:18.010
a human will rapidly generalize from specific

00:42:18.010 --> 00:42:21.909
examples. You learned how to do long division.

00:42:22.210 --> 00:42:25.130
You never had to look at every possible long

00:42:25.130 --> 00:42:28.670
division example to know how to do that. But

00:42:28.670 --> 00:42:33.230
the AI works exactly the opposite. AI looks at

00:42:33.230 --> 00:42:37.170
a lot of different phenomena in general and then

00:42:37.170 --> 00:42:41.469
tries to revert to the specifics. to apply that

00:42:41.469 --> 00:42:45.670
logic to a specific circumstance, like in traffic,

00:42:45.809 --> 00:42:47.610
right? What the hell does it do? It looks back

00:42:47.610 --> 00:42:51.190
in the database and looks at a lot of events,

00:42:51.309 --> 00:42:53.489
none of which are identical to what it's seeing,

00:42:53.670 --> 00:42:58.210
and tries to apply that general information to

00:42:58.210 --> 00:43:02.269
the specific circumstance. It's just the inverse

00:43:02.269 --> 00:43:06.809
of the way humans think. There's even another

00:43:06.809 --> 00:43:10.869
deeper problem. No human being can unambiguously

00:43:10.869 --> 00:43:16.449
relate the cause and effect for an AI engine.

00:43:17.030 --> 00:43:20.650
It's by definition, it's something that the machine

00:43:20.650 --> 00:43:25.909
does that a human can't do. So in a traditional

00:43:25.909 --> 00:43:29.929
development program, you find a problem, you

00:43:29.929 --> 00:43:35.409
trace it back through the software to the, whatever

00:43:35.409 --> 00:43:38.579
the instigating circumstances, right? And you

00:43:38.579 --> 00:43:42.019
fix the bug and then you're good to go. That

00:43:42.019 --> 00:43:44.420
cannot happen in a machine learning environment.

00:43:44.900 --> 00:43:49.400
The only way to relate cause and effect in a

00:43:49.400 --> 00:43:51.719
machine learning environment is to just run the

00:43:51.719 --> 00:43:54.820
test thousands and thousands of times and develop

00:43:54.820 --> 00:43:58.840
a statistical relationship that says in X number

00:43:58.840 --> 00:44:02.860
of circumstances, this Y is going to happen.

00:44:03.440 --> 00:44:05.579
But you've also got a number where it's not going

00:44:05.579 --> 00:44:08.260
to happen, right? So the statistics are very

00:44:08.260 --> 00:44:11.280
different than the deterministic approach that

00:44:11.280 --> 00:44:15.539
is conventional. And no human being can understand

00:44:15.539 --> 00:44:18.880
that in prospect. Things happen. Like Phil talked

00:44:18.880 --> 00:44:23.780
last week about people or the machine learning

00:44:23.780 --> 00:44:27.730
mistaking people with the yellow vest on. Or

00:44:27.730 --> 00:44:30.110
a traffic sign, right? Because traffic signs

00:44:30.110 --> 00:44:32.889
are always yellow, so anything yellow is a traffic

00:44:32.889 --> 00:44:35.730
sign. This is something no human being would

00:44:35.730 --> 00:44:40.469
do, but it was not anticipated either that the

00:44:40.469 --> 00:44:42.309
machine learning algorithm would do that because

00:44:42.309 --> 00:44:46.489
people just don't think that way. With an AI

00:44:46.489 --> 00:44:48.449
engine, there's always uncertainty that the right

00:44:48.449 --> 00:44:51.730
causes were selected and that the probability

00:44:51.730 --> 00:44:54.150
of effect is high enough to declare safety with

00:44:54.150 --> 00:44:56.940
sufficient confidence. So you really never know.

00:44:57.079 --> 00:45:00.920
Replicating once or twice is not enough. In a

00:45:00.920 --> 00:45:04.079
recent event, a public school vice principal

00:45:04.079 --> 00:45:06.780
stood in front of a Waymo to prevent it from

00:45:06.780 --> 00:45:08.619
turning down a street where students were emerging

00:45:08.619 --> 00:45:11.579
from a school during a fire drill. There was

00:45:11.579 --> 00:45:14.699
no audible or visual clue that that would have

00:45:14.699 --> 00:45:18.980
been a trainable cue. So there's no klaxon going

00:45:18.980 --> 00:45:22.059
off that the Waymo could detect to say, well,

00:45:22.139 --> 00:45:23.860
there's a problem here. We can't go down this

00:45:23.860 --> 00:45:27.579
way. No, the only way to protect the students

00:45:27.579 --> 00:45:30.000
was for the vice principal to stand in front

00:45:30.000 --> 00:45:36.480
of the Waymo. So how do you train that? In order

00:45:36.480 --> 00:45:39.619
to train the system, you'd have to replicate

00:45:39.619 --> 00:45:43.019
it hundreds of times, minimum tens of times,

00:45:43.099 --> 00:45:45.239
hundreds of times. How do you do that? How do

00:45:45.239 --> 00:45:47.619
you replicate that situation enough to train

00:45:47.619 --> 00:45:50.340
the machine? Because if you don't have lots and

00:45:50.340 --> 00:45:52.679
lots of events in the database, the machine's

00:45:52.679 --> 00:45:55.420
not going to know what to do. There's not enough

00:45:55.420 --> 00:45:59.219
time and money in the world to replicate that

00:45:59.219 --> 00:46:01.880
over and over again. And even if you do train

00:46:01.880 --> 00:46:06.500
it for one particular school, how do you know

00:46:06.500 --> 00:46:08.480
that that's going to be good for all the schools

00:46:08.480 --> 00:46:11.239
that are going to happen? So the only thing you

00:46:11.239 --> 00:46:12.739
can really do is throw up your hands and say,

00:46:12.820 --> 00:46:16.039
well, you know, this is just not something we

00:46:16.039 --> 00:46:20.300
can cover. But calling it an edge case... gives

00:46:20.300 --> 00:46:23.639
you a perceptual cover for it to say, well, it's

00:46:23.639 --> 00:46:25.420
just an edge case. You know, it's not something

00:46:25.420 --> 00:46:28.639
that we have to worry about. Well, who says it's

00:46:28.639 --> 00:46:32.440
not something you have to worry about? Can remote

00:46:32.440 --> 00:46:35.119
assistance solve the problem? Well, remote assistance

00:46:35.119 --> 00:46:38.119
can solve some problems, but only if they're

00:46:38.119 --> 00:46:40.840
happening very, very slowly over the stationary

00:46:40.840 --> 00:46:46.980
car. If you are in an AV at speed going down

00:46:46.980 --> 00:46:49.510
the highway, And all of a sudden the control

00:46:49.510 --> 00:46:52.429
system fails, right? Should have turned left

00:46:52.429 --> 00:46:55.949
instead of turned right or whatever. If it happens

00:46:55.949 --> 00:46:58.110
at speed, there's no way in the world remote

00:46:58.110 --> 00:47:00.309
assistance is going to help you solve that problem.

00:47:01.230 --> 00:47:04.590
There's no reason to believe that all of the

00:47:04.590 --> 00:47:07.130
control system failures that you could call an

00:47:07.130 --> 00:47:11.170
edge case are going to happen at low speed. There's

00:47:11.170 --> 00:47:12.969
every reason to believe that a lot of them are

00:47:12.969 --> 00:47:15.349
going to happen at high speed. We've seen this

00:47:15.349 --> 00:47:21.280
many times, right? We've seen phantom braking

00:47:21.280 --> 00:47:24.480
occur. It caused accidents. It caused collisions

00:47:24.480 --> 00:47:27.119
to occur. There's no way that a remote assistance

00:47:27.119 --> 00:47:31.659
can sort through all of the phenomenon and the

00:47:31.659 --> 00:47:35.320
control structure, the command structure, all

00:47:35.320 --> 00:47:38.219
that stuff that happens when it's being pulled

00:47:38.219 --> 00:47:42.800
for a remote assistance. And there's no way of

00:47:42.800 --> 00:47:46.099
knowing that the car will even know. that it

00:47:46.099 --> 00:47:50.920
needs to call remote assistance. So I think here's

00:47:50.920 --> 00:47:53.460
my bottom line. Calling an unsafe AV operation

00:47:53.460 --> 00:47:57.440
an edge case is misleading and gives developers

00:47:57.440 --> 00:47:59.679
an easy out for what should be characterized

00:47:59.679 --> 00:48:04.019
as an unsafe control failure. Language is important,

00:48:04.219 --> 00:48:07.420
and these ponderous, dangerous machines should

00:48:07.420 --> 00:48:10.219
not be misconstrued as benign, except for the

00:48:10.219 --> 00:48:13.099
occasional whoopsie. Their hazardous industrial

00:48:13.099 --> 00:48:16.639
machinery being deployed in public without the

00:48:16.639 --> 00:48:19.320
customary safeguards. They're robbing people

00:48:19.320 --> 00:48:22.280
of their rights by the terms of service while

00:48:22.280 --> 00:48:24.139
endangering the public without their knowledge

00:48:24.139 --> 00:48:29.000
or consent. So there you go. I'm a little opinionated

00:48:29.000 --> 00:48:31.659
about it, but that's my discussion for this week.

00:48:33.239 --> 00:48:36.880
Excellent. We appreciate you. That's what I'm

00:48:36.880 --> 00:48:39.539
supposed to say, right? Is that it? Yeah. Okay,

00:48:39.559 --> 00:48:43.050
good. Are you supposed to say thank you for sharing,

00:48:43.150 --> 00:48:45.969
right? Oh, thank you for sharing. I hear you

00:48:45.969 --> 00:48:49.570
and I understand you, but I don't understand

00:48:49.570 --> 00:48:54.510
you. Wait, what? No, okay. I don't know. Before

00:48:54.510 --> 00:48:57.730
we go into recalls, let's jump into an article

00:48:57.730 --> 00:49:01.090
from Smart Cities Dive. It's a horrible name.

00:49:01.610 --> 00:49:05.389
Bike lane speed cameras removed from FHWA safety

00:49:05.389 --> 00:49:07.969
measures list. The Federal Highway Administration.

00:49:08.780 --> 00:49:12.840
remove these things oh my god and didn't explain

00:49:12.840 --> 00:49:16.300
it i'm gonna quote from they they cut they kind

00:49:16.300 --> 00:49:18.780
of did actually well you're quoted in this article

00:49:18.780 --> 00:49:20.159
so we'll get to you in a second you know this

00:49:20.159 --> 00:49:24.079
happened right after right after i think sean

00:49:24.079 --> 00:49:27.239
duffy tweeted that that bike lanes are essentially

00:49:27.239 --> 00:49:33.139
dei and since they're eliminating dei and all

00:49:33.139 --> 00:49:37.159
that nonsense from the government I guess they

00:49:37.159 --> 00:49:38.920
consider bike lanes part of that. So they've

00:49:38.920 --> 00:49:44.699
suspended allowing the bike lanes and speed enforcement

00:49:44.699 --> 00:49:48.079
cameras from the list of safety countermeasures

00:49:48.079 --> 00:49:50.940
that are on the Federal Highway Administration's

00:49:50.940 --> 00:49:53.940
website. Those are the approved countermeasures

00:49:53.940 --> 00:49:56.500
by the federal government that states can receive

00:49:56.500 --> 00:50:00.110
grants for. They've taken off. Speed cameras

00:50:00.110 --> 00:50:04.150
and and bike lanes, bike lanes under the assertion

00:50:04.150 --> 00:50:06.949
that somehow there are they are DEI, which is

00:50:06.949 --> 00:50:09.909
utter nonsense. But they've replaced this with

00:50:09.909 --> 00:50:12.630
men who are who have high T levels just yelling

00:50:12.630 --> 00:50:15.010
at you. Hey, slow down. Don't drive so fast.

00:50:15.389 --> 00:50:19.610
Get off that bike, wimp. No. Quoting from this

00:50:19.610 --> 00:50:22.329
article, speed safety cameras can substantially

00:50:22.329 --> 00:50:25.050
reduce speeding, according to the Insurance Institute

00:50:25.050 --> 00:50:28.670
for Highway Safety. In 2025, 338 communities

00:50:28.670 --> 00:50:32.050
operated speed safety cameras per IIHS data.

00:50:32.250 --> 00:50:34.369
They are important tools for a comprehensive

00:50:34.369 --> 00:50:38.070
approach to safety, an IIHS spokesperson said

00:50:38.070 --> 00:50:40.010
in an email. Getting drivers to pay attention

00:50:40.010 --> 00:50:42.670
and slow down protects everyone using our roads.

00:50:44.469 --> 00:50:49.130
Duh. Like how I don't understand why what is

00:50:49.130 --> 00:50:54.050
woke like a sure bike lane. The only look, the

00:50:54.050 --> 00:50:58.250
only the only justification here in the mind

00:50:58.250 --> 00:51:02.130
of, I guess, Transportation Secretary Duffy is

00:51:02.130 --> 00:51:05.750
he doesn't know how to ride a bike. He just wants

00:51:05.750 --> 00:51:09.210
political points from the group of voters who.

00:51:09.929 --> 00:51:12.989
Our anti speed cameras are essentially and the

00:51:12.989 --> 00:51:15.610
secretary of the Department of Transportation

00:51:15.610 --> 00:51:19.670
is pandering to anti safety folks for political

00:51:19.670 --> 00:51:21.869
points. I mean, it's pretty pathetic. I want

00:51:21.869 --> 00:51:24.630
more speed cameras because then we can eliminate,

00:51:24.690 --> 00:51:27.070
you know, cops who are hiding on the street doing

00:51:27.070 --> 00:51:30.469
really the hall monitors job as an adult, like

00:51:30.469 --> 00:51:32.349
pulling people over for speed and just speed

00:51:32.349 --> 00:51:34.969
cameras do that. They'll be better. And I think

00:51:34.969 --> 00:51:37.250
he's against bike lanes because he never learned

00:51:37.250 --> 00:51:40.460
how to ride a bike. This is so sad. sean duffy

00:51:40.460 --> 00:51:42.400
come on over i'll teach you how to ride a bike

00:51:42.400 --> 00:51:45.239
big boy it'll be fine no no worry this should

00:51:45.239 --> 00:51:48.380
have been a gas light this week it's just non

00:51:48.380 --> 00:51:53.179
-stop gas light all week but but who who is there

00:51:53.179 --> 00:51:57.460
a lobby saying get rid of speed cameras is it

00:51:57.460 --> 00:52:01.340
is it david letterman No, there's it's just it's

00:52:01.340 --> 00:52:04.760
it's literally just people who are entitled speeders

00:52:04.760 --> 00:52:06.900
who are complaining to their representatives,

00:52:07.380 --> 00:52:10.340
their council people, their federal reps. They

00:52:10.340 --> 00:52:13.000
think, you know, that this is a they come on

00:52:13.000 --> 00:52:15.139
all sorts of nonsense, a violation of their privacy.

00:52:15.179 --> 00:52:18.260
Like, look, I get if it's, you know. a flock

00:52:18.260 --> 00:52:21.079
camera or something that is tracking you and

00:52:21.079 --> 00:52:24.340
isn't secure. And people can hack into flock

00:52:24.340 --> 00:52:27.760
cameras and figure out everybody's gone by stoplight.

00:52:27.780 --> 00:52:30.420
Like clearly there's a privacy issue there, but

00:52:30.420 --> 00:52:32.800
a speed enforcement camera that's not tracking

00:52:32.800 --> 00:52:34.900
that type of thing. And it's just singling out

00:52:34.900 --> 00:52:37.539
people who are not just over the speed limit,

00:52:37.559 --> 00:52:39.840
but well over the speed limit enough to get a

00:52:39.840 --> 00:52:44.059
ticket. I don't, I, you know, and we know, we

00:52:44.059 --> 00:52:47.010
know. that they reduce crashes. We know they

00:52:47.010 --> 00:52:49.550
reduce fatalities, and we know they reduce injuries.

00:52:49.690 --> 00:52:52.289
Like, it's not even up for debate. And for the

00:52:52.289 --> 00:52:54.889
Secretary of Transportation to say, essentially,

00:52:55.130 --> 00:52:57.210
the federal government no longer supports this

00:52:57.210 --> 00:53:01.349
when they know it is working to save lives is

00:53:01.349 --> 00:53:03.949
just terrible. There's no other way to put it.

00:53:04.909 --> 00:53:08.570
I got busted in D .C., and then as a result of

00:53:08.570 --> 00:53:10.289
that, I started to observe the speed limits.

00:53:10.449 --> 00:53:13.440
So I was on the Southeast Expressway. Driving

00:53:13.440 --> 00:53:16.159
at 45 miles an hour, which is the surprisingly

00:53:16.159 --> 00:53:19.840
low speed limit for that. And I got to tell you,

00:53:19.900 --> 00:53:22.980
people were not happy with my performance that

00:53:22.980 --> 00:53:27.860
day or any other day. That's going to be in your

00:53:27.860 --> 00:53:30.480
tombstone. People were not happy with my performance.

00:53:34.300 --> 00:53:36.139
Speaking of not happy with performance, let's

00:53:36.139 --> 00:53:41.309
go into recalls. First up, BMW. That's right,

00:53:41.409 --> 00:53:45.409
Bavarian Motorworks, 29 ,119 vehicles, the 2018

00:53:45.409 --> 00:53:50.650
to 2020, some dumb name. It's a plug -in hybrid,

00:53:50.909 --> 00:53:53.010
530e xDrive. Yeah, it's their little bitty plug

00:53:53.010 --> 00:53:57.250
-in. Yeah, 2017 to 19, another xDrive thing.

00:53:57.690 --> 00:54:04.579
2018 to 2020, 530i performance. If you have a

00:54:04.579 --> 00:54:05.760
plug -in hybrid. The important thing is that

00:54:05.760 --> 00:54:08.659
if you have a tiny plug -in hybrid, BMW need

00:54:08.659 --> 00:54:11.460
to park outside. Yes. Yeah, because in certain

00:54:11.460 --> 00:54:15.280
cases, corrosion inside the starter relay can

00:54:15.280 --> 00:54:17.880
also lead to a short circuit and possible overheating

00:54:17.880 --> 00:54:20.579
of the starter. In an extreme case, oh, yeah,

00:54:20.699 --> 00:54:23.099
this could cause a thermal event during vehicle

00:54:23.099 --> 00:54:24.679
operation when the vehicle is parked and the

00:54:24.679 --> 00:54:29.440
ignition is off. Thermal event means fire. Yes.

00:54:32.599 --> 00:54:34.980
And this is, as Michael said, listed as a park

00:54:34.980 --> 00:54:38.300
outside event. But beyond that, it looks like

00:54:38.300 --> 00:54:39.900
they're just going to replace your starter and

00:54:39.900 --> 00:54:42.139
it'll be at the end of August. So park outside

00:54:42.139 --> 00:54:44.199
for a couple of months and get your get your.

00:54:45.679 --> 00:54:47.699
appointment scheduled as soon as possible, and

00:54:47.699 --> 00:54:49.699
then you can park in your garage again. A friend

00:54:49.699 --> 00:54:51.380
of mine has one of these little tiny cars like

00:54:51.380 --> 00:54:53.699
that. I don't know if it's a hybrid. A friend

00:54:53.699 --> 00:54:55.840
of mine had one, too. He said he hardly ever

00:54:55.840 --> 00:54:58.539
had to buy gas. Yeah, she's got one, and it's

00:54:58.539 --> 00:55:00.039
like, I was in it once, and I was like, it's

00:55:00.039 --> 00:55:02.940
like being in an iMac from 1998. I'm like, this

00:55:02.940 --> 00:55:05.179
is so, like, the color tones. And the gas tank

00:55:05.179 --> 00:55:07.380
is tiny, too. I guess. It's just a couple gallons

00:55:07.380 --> 00:55:11.179
or something. But the names are dumb. Next up.

00:55:12.070 --> 00:55:16.210
Ford, that's right, world champion, heavyweight,

00:55:16.269 --> 00:55:22.550
high -T company. Ford, 387 ,911 vehicles, the

00:55:22.550 --> 00:55:28.829
2020 -2027 Aviator, 2020 -26 Explorer. On affected

00:55:28.829 --> 00:55:31.409
vehicles, it is possible that the easy entry

00:55:31.409 --> 00:55:35.309
seat tip and slide switch on the second row outboard

00:55:35.309 --> 00:55:38.449
seats may intermittently bind or become stuck

00:55:38.449 --> 00:55:41.019
in the down position. This binding or sticking

00:55:41.019 --> 00:55:43.300
may result in the seat tipping and or sliding

00:55:43.300 --> 00:55:46.639
unexpectedly, potentially while driving. Oh,

00:55:46.880 --> 00:55:49.539
thank God it's not like the driver's seat because

00:55:49.539 --> 00:55:52.119
that'd be really bad. But so I'm stuck in the

00:55:52.119 --> 00:55:54.719
back of this vehicle and my aviator because I'm

00:55:54.719 --> 00:55:57.719
cool. I'm flexing. And then it just sliding around

00:55:57.719 --> 00:56:01.380
on me. Is that what the deal is? Yeah, I mean,

00:56:01.400 --> 00:56:03.320
I think the biggest concern here is that in a

00:56:03.320 --> 00:56:06.159
crash, what, you know, what could happen here?

00:56:06.599 --> 00:56:10.199
You know, we really like our seats to be static

00:56:10.199 --> 00:56:14.599
and non -moving on their rails in a crash. So

00:56:14.599 --> 00:56:18.920
that's the problem. Ford, you know, it looks

00:56:18.920 --> 00:56:20.900
like owners are going to be waiting a long time

00:56:20.900 --> 00:56:24.239
on this one. So either avoid use of that seat

00:56:24.239 --> 00:56:28.880
or. Be really careful about ensuring that it

00:56:28.880 --> 00:56:32.739
is fixed before you leave for wherever you're

00:56:32.739 --> 00:56:34.659
going, because you're going to have a while here.

00:56:34.760 --> 00:56:37.139
It doesn't look like owners are going to get

00:56:37.139 --> 00:56:40.920
an actual remedy for this until the start of

00:56:40.920 --> 00:56:46.500
next year. Next up, Hyundai 47 ,749 vehicles,

00:56:46.599 --> 00:56:51.039
the 2025 Kona Electric, the 2026 regular Kona.

00:56:51.559 --> 00:56:55.260
The rear center seatbelt buckle assemblies in

00:56:55.260 --> 00:56:56.880
the subject vehicles may have been manufactured

00:56:56.880 --> 00:56:59.599
outside of specification, which could affect

00:56:59.599 --> 00:57:02.980
their performance during a crash. Oh, the supplier's

00:57:02.980 --> 00:57:05.980
stamping die was not maintained properly, resulting

00:57:05.980 --> 00:57:13.960
in wear from overuse. Oh, come on, boo. Bad buckles.

00:57:14.340 --> 00:57:18.300
September 11th is when owners will probably see

00:57:18.300 --> 00:57:23.159
their notification. And hopefully they can get

00:57:23.159 --> 00:57:25.880
that fixed. They're still using kids to assemble

00:57:25.880 --> 00:57:29.980
cars or is that? No. I think they got busted.

00:57:29.980 --> 00:57:31.679
And remember, it wasn't them directly. It was

00:57:31.679 --> 00:57:33.519
just the third party suppliers. One of their

00:57:33.519 --> 00:57:36.000
suppliers. That happened to be directly related

00:57:36.000 --> 00:57:37.880
to the company or something. Down in Montgomery,

00:57:38.059 --> 00:57:43.289
Alabama area. Next up. Oh, this is the first

00:57:43.289 --> 00:57:45.269
time I believe we've seen this company in the

00:57:45.269 --> 00:57:47.510
recall roundup. Is this true? Is this the first

00:57:47.510 --> 00:57:50.369
time? Second time? It's all like not more than

00:57:50.369 --> 00:57:51.630
it's happened a couple of times. A couple of

00:57:51.630 --> 00:57:56.510
times. All right. Zooks. That's right. Zooks.

00:57:57.150 --> 00:57:59.670
The car isn't going forward. Is it going backwards?

00:57:59.809 --> 00:58:03.070
Who knows? One hundred and five vehicles. This

00:58:03.070 --> 00:58:05.690
is just the Zooks, the automated driving system.

00:58:05.769 --> 00:58:08.150
That's the name of the vehicle. Reduced visibility

00:58:08.150 --> 00:58:10.590
while driving near an active emergency scene

00:58:10.590 --> 00:58:13.090
can increase the risk of a crash or impede first

00:58:13.090 --> 00:58:18.550
responders. This is fascinating. So let me repeat

00:58:18.550 --> 00:58:22.329
this. Reduced visibility while driving. So as

00:58:22.329 --> 00:58:24.989
a human driver, you understand what that means.

00:58:25.230 --> 00:58:28.210
Oh, maybe my A pillar is way too big and it's

00:58:28.210 --> 00:58:30.409
blocking my view. Maybe the defroster didn't

00:58:30.409 --> 00:58:33.280
work. I can't see anything. Now the Zooks are

00:58:33.280 --> 00:58:36.480
supposed to have cameras, radar, LIDAR, ultrasonic.

00:58:38.119 --> 00:58:42.320
What does reduced visibility mean for Zooks?

00:58:44.619 --> 00:58:49.820
I don't know. We're all at a loss. I mean, it

00:58:49.820 --> 00:58:53.539
sounds like, you know, this is an active fire

00:58:53.539 --> 00:58:56.119
with heavy smoke. This isn't like what we all

00:58:56.119 --> 00:58:59.590
had in the Northeast. Last week and maybe maybe

00:58:59.590 --> 00:59:01.969
we'll have again soon with the wildfire. This

00:59:01.969 --> 00:59:05.630
is an this is a Zooks that was near of, you know,

00:59:05.630 --> 00:59:08.750
a very active fire and apparently wasn't able

00:59:08.750 --> 00:59:11.809
to respond appropriately. Yeah. From how did

00:59:11.809 --> 00:59:16.429
they fix that? I just I wonder. How did they

00:59:16.429 --> 00:59:18.809
fix this so quick? Yeah, it says the vehicle

00:59:18.809 --> 00:59:22.329
encountered heavy smoke that obscured an active

00:59:22.329 --> 00:59:28.409
emergency fire scene. OK, what? you so the there's

00:59:28.409 --> 00:59:30.150
too much smoke i couldn't see there was fire

00:59:30.150 --> 00:59:35.590
like what like that just seems wrong ah like

00:59:35.590 --> 00:59:39.570
what is going on here but okay shouldn't lidar

00:59:39.570 --> 00:59:43.210
work through smoke i mean unless the smoke particles

00:59:43.210 --> 00:59:47.059
are massive No, LIDAR should not necessarily

00:59:47.059 --> 00:59:49.139
work through smoke. Oh, LIDAR won't work through

00:59:49.139 --> 00:59:51.760
smoke. Okay, what about... Not necessarily. It

00:59:51.760 --> 00:59:54.860
depends on a lot of variables, but not necessarily.

00:59:55.340 --> 00:59:58.480
I'm assuming this is very heavy smoke. Radar

00:59:58.480 --> 01:00:00.280
should work through smoke. Radar should be, okay.

01:00:00.920 --> 01:00:04.139
But the thing is, it's saying it encountered

01:00:04.139 --> 01:00:07.019
heavy smoke and it obscured the scene, and so

01:00:07.019 --> 01:00:11.800
it entered the scene. Heavy smoke, no, don't

01:00:11.800 --> 01:00:15.679
go. No human would make that mistake. Okay, there's

01:00:15.679 --> 01:00:17.820
a couple people I know, but they're not that

01:00:17.820 --> 01:00:21.460
bright. Maybe it has a faith -based control system.

01:00:23.079 --> 01:00:26.639
Again, designers of these systems, I'm going

01:00:26.639 --> 01:00:28.360
to make it sound very easy, and I know it is

01:00:28.360 --> 01:00:31.440
not. If you encounter a situation you don't understand,

01:00:31.780 --> 01:00:37.260
leave that situation as safely as possible. That's

01:00:37.260 --> 01:00:45.599
just my end of rant. So what's the deal? I love

01:00:45.599 --> 01:00:47.380
it says planned dealer notification date. No

01:00:47.380 --> 01:00:50.739
dealers. Planned owner notification date. No

01:00:50.739 --> 01:00:55.940
owners. So has this been fixed? Isn't 105 every

01:00:55.940 --> 01:00:59.280
vehicle they've got? Yeah. It must be every vehicle

01:00:59.280 --> 01:01:03.219
they have. Anyway. Yeah, it looks like, you know,

01:01:03.300 --> 01:01:08.059
it looks like they said they're doing an updated

01:01:08.059 --> 01:01:11.289
software. will be released to all impacted Zoox

01:01:11.289 --> 01:01:15.510
vehicles. And they don't specify much, you know,

01:01:15.530 --> 01:01:19.070
about the fix, but. Trust us, bro, it works.

01:01:20.090 --> 01:01:24.429
Lastly, Ford Motor Company coming back, you know,

01:01:24.429 --> 01:01:29.530
with an uppercut, 288 ,314 vehicles, 2016 to

01:01:29.530 --> 01:01:34.510
2019 Explorer. The roof rail covers may become

01:01:34.510 --> 01:01:36.550
loose and detached from the vehicle while it

01:01:36.550 --> 01:01:40.880
is in motion. A detached roof rail trim cover

01:01:40.880 --> 01:01:43.000
can separate from the vehicle, creating a rose

01:01:43.000 --> 01:01:45.159
hazard for other drivers and increasing the risk

01:01:45.159 --> 01:01:48.980
of crash. Is this just to tighten the screw?

01:01:49.639 --> 01:01:53.820
Is this what happened? You know, I don't know.

01:01:53.880 --> 01:01:57.159
I think that there have been a lot of recalls

01:01:57.159 --> 01:02:05.139
by Ford for, you know, all of this stuff on the

01:02:05.139 --> 01:02:10.739
outside of the vehicle. It keeps falling off

01:02:10.739 --> 01:02:13.179
and endangering other road use. I was reading

01:02:13.179 --> 01:02:15.239
an article about this the other day. They just

01:02:15.239 --> 01:02:18.099
had a lot of issues with this kind of problem.

01:02:19.300 --> 01:02:22.159
Not just roof rail covers, but other pieces of

01:02:22.159 --> 01:02:26.420
the vehicle exterior. So, you know, I think this

01:02:26.420 --> 01:02:32.239
is Ford's 58th, 59th recall of the year. And

01:02:32.239 --> 01:02:34.679
these are older vehicles. So this, this has been

01:02:34.679 --> 01:02:36.860
a problem that's been going on for quite some

01:02:36.860 --> 01:02:39.260
time. I mean, this is a decade, 10 year old vehicles

01:02:39.260 --> 01:02:42.579
in some cases. Yeah. I think Ford needs to set

01:02:42.579 --> 01:02:44.619
a focus on quality for their customers, maybe

01:02:44.619 --> 01:02:48.239
focus on quality for their employees. So they

01:02:48.239 --> 01:02:51.420
don't keep making these types of mistakes. Happy

01:02:51.420 --> 01:02:55.500
worker, happy products. I don't know. I'm going

01:02:55.500 --> 01:02:57.860
to speculate that it's adhesive wearing out.

01:02:58.260 --> 01:03:01.400
They probably. They probably glue these parts

01:03:01.400 --> 01:03:06.079
in, and after a few years, the glue just becomes

01:03:06.079 --> 01:03:09.960
brittle and breaks or something. That's pure

01:03:09.960 --> 01:03:12.559
speculation on my part, but that would... And

01:03:12.559 --> 01:03:15.699
you're incorrect, because according to the remedy,

01:03:15.880 --> 01:03:20.260
it says they need additional pushpin fasteners.

01:03:20.320 --> 01:03:24.239
The pushpin fasteners are present, then no...

01:03:24.239 --> 01:03:27.019
It seems like a little pushpin, little piece

01:03:27.019 --> 01:03:29.179
of plastic is broken off. It doesn't even seem

01:03:29.179 --> 01:03:33.539
to be. adhesive so anyway with that listeners

01:03:33.539 --> 01:03:37.119
that's our show uh before we go i do have one

01:03:37.119 --> 01:03:39.539
plea to listeners if anyone would like to send

01:03:39.539 --> 01:03:41.800
michael brooks a new set of headphones that would

01:03:41.800 --> 01:03:44.300
be great he's been suffering with these in ear

01:03:44.300 --> 01:03:48.119
earbuds the entire time he can't find his headphones

01:03:48.119 --> 01:03:52.179
a nice set of uh you know sony studio monitors

01:03:52.179 --> 01:03:55.460
he'll be happy that'll be great or you know anything

01:03:55.460 --> 01:03:58.659
like that because You know, we have to see his

01:03:58.659 --> 01:04:00.699
face and we can see the discomfort. And it's

01:04:00.699 --> 01:04:03.980
not just from everything that I say. Yeah, I'm

01:04:03.980 --> 01:04:05.780
having to tilt my head to keep the earphone.

01:04:06.380 --> 01:04:08.840
Michael would be happy to model them for the

01:04:08.840 --> 01:04:13.119
publicity, you know, for the PR group that's

01:04:13.119 --> 01:04:16.340
providing the headphones. His picture's available.

01:04:16.739 --> 01:04:19.199
Oh, I wasn't even suggesting like Sony or Shure

01:04:19.199 --> 01:04:20.639
or somebody send him a headphone. I was figuring

01:04:20.639 --> 01:04:24.159
out a happy listener. And with that, till next

01:04:24.159 --> 01:04:27.630
week, we'll be back. Bye -bye. bye bye thank

01:04:27.630 --> 01:04:33.670
you bye bye for more information visit www .autosafety

01:04:33.670 --> 01:04:34.210
.org