EP15 - Has technology ruined cruise control? Drew rides in a Waymo!

Tesla Autopilot. Hyundai Highway Driving Assist 2. GM SuperCruise. Oh, and a Waymo robotaxi. This episode has it all.

Joe and Drew talk about their experiences of each and ask the crucial question: what’s it all for?

📨 Subscribe to Looking Out - The Newsletter, the partner newsletter to this podcast. https://automobility.substack.com

⬇️ Download Drew’s free trend report from CES 2024: https://www.studiophro.com/ces-2024-report-download-form

🤝 Connect with us on LinkedIn Drew - https://www.linkedin.com/in/drewpasmith/ Joe - https://www.linkedin.com/in/joseph-simpson-92863a2/

🔗 Links Dude, where’s my self-driving car - the latest from The Verge: https://www.theverge.com/24065447/self-driving-car-autonomous-tesla-gm-baidu A crowd destroyed a driverless Waymo car in San Francisco: https://www.theverge.com/2024/2/11/24069251/waymo-driverless-taxi-fire-vandalized-video-san-francisco-china-town

📸 Follow us on Instagram https://www.instagram.com/lookingoutpodcast/

That’s it for this episode! Thanks for listening.

If you like what you hear, please leave a review for us on your favourite podcasting platform. It helps other folk like you find us!

And you can sign up for Looking Out - The Newsletter, the sidekick to our podcast, here: automobility.substack.com

---

Click ‘Read more’ for the transcript

Transcript:

Note: This transcript is machine-generated and may contain some errors.

Pre-roll

[00:00:00] Joe Simpson: There’s an argument that these systems act as if you like, um, you know, they’re, they’re guardrails to stop you going wrong, to catch you when you screw up, to mitigate when you basically lose attention. but if they don’t provide the safeguards that they’re supposed to in a consistent way, if they’re not foolproof, then there’s obviously big holes that start to appear in that safety net.

[00:00:28]

Intro

[00:00:29] Joe Simpson: Hello, I’m Joe Simpson.

[00:00:30] And

[00:00:33] welcome to Looking Out The Podcast, in which we connect the dots across mobility, design, and culture.

[00:00:42] Drew Smith: And coming up in this, episode 15 of our beloved podcast, we talk about that most vexed of issues for the automotive industry: automated driving and advanced driver assist systems. Now I should say upfront that neither of us are experts in safety systems. Neither of us are experts in the legislatory environment that surrounds the use and design of these systems.

[00:01:12] But we are people that think very deeply about the user’s experience of these systems and how expectations and misset expectations around them have led to some pretty challenging situations for a number of car makers over the past. I don’t know, decade or so, if we think about when they started being introduced.

[00:01:35] So today, what we would love to do is talk about our recent experiences with, well, Joe, you had GM’s Super Cruise

[00:01:46] In your hot little hands, and I recently was able to try Tesla’s Autopilot with Autosteer on the hellish freeways of Los Angeles, Hyundai’s Highway Driving Assist 2, in the Bay Area, and most excitingly of all, Waymo’s RoboTaxi, which Is a fully autonomous, robotaxi system that’s operating in the Bay Area.

[00:02:13] So, Joe, with that out of the way, there’s one more bit of context that I would like to set. And that is, and it may be surprising to a lot of people that, listen to this show, I’m far from expert with using these systems. I have a Volkswagen Up! in which the cruise control, does not actually work.

[00:02:32] and I drive. A

[00:02:34] But design or failure.

[00:02:36] no just, just failure, it’s just a bit shit, typical Volkswagen group stuff. I have a 1990s Mercedes which still has what I regard to be one of the finest cruise control interfaces ever put into a car. And apart from a period of time when I was working in your current hometown of Gothenburg, and it was my professional responsibility to be au fait with these systems, I don’t actually use them a lot.

[00:03:04] So it was really interesting for me to come at them with an interested novice’s mind, and try and wrap my head around them, in a live environment.

[00:03:18] Joe Simpson: And I think it’s a really good point, Drew, because although I probably have a bit more experience there, before I moved to Sweden, I was,playing part time car journalist and getting different cars every week and making it my business to try out these various systems. I think you raise an important question, which is how do people safely, and securely understand how these systems work and their operating domains and what they can and can’t do And I think as you said at the top of this show we come at that very much from a design point of view which means we are trying to look at it from a problem solving point of view and from an experience point of view.

[00:04:02] So in other words when we look at these systems, what problem are they actually trying to solve?

[00:04:08] And what kind of experience are they creating for both the driver, but other people in the car? I think that’s a really good starting point. and a kind of really good set of questions to kick us off. Drew, from your experience, any answers to those things?

Drew’s experience of Tesla Autopilot and Hyundai Highway Driving Assist

[00:04:26] Drew Smith: Well, I guess, you know, my, my experience of the Hyundai and the Tesla systems were based as a rental driver, effectively.

[00:04:33] I rented, a 2020, Tesla Model 3, from Turo, and I did the same with the Hyundai Ioniq. so really coming at these cars completely cold. And, it was the first really extended period of time that I’d spent, driving a Model 3.

[00:04:50] And the first question that comes to mind is,how do I actually activate these systems, right? and if you’ve never driven, either of these cars before, it’s actually the Hyundai that kind of wins. It’s a partial victory in this space because, there’s at least a recognizable set of icons on a set of steering wheel buttons that look vaguely cruise control ish that you can start messing around with.

[00:05:18] with the Tesla you actually don’t know that the, that the stalk to the right of the steering wheel is actually the primary interface for how you activate and deactivate the system.

[00:05:31] Joe Simpson: Is there any labeling or iconography on that stock? I can’t remember. It’s a while since I’ve driven a Model 3. Right.

[00:05:46] Drew Smith: Tesla’s creator, Elon Musk, I decided to fuck around and find out. And I think this is, this is The nature of these systems when you’re coming at them cold is you are doing a lot of fucking around and finding out to determine what the operational boundaries are, what the capabilities of these systems are, and if you’re moving from vehicle to vehicle brand to brand, system to system, You kind of have to keep some sort of mental log of how each of these systems work, and that to me, at the level of the industry, seems like a bit of a failure,

[00:06:33] Joe Simpson: It

[00:06:34] Drew Smith: the cognitive load that you have to,undergo each time you get into a new vehicle is pretty significant.

[00:06:42] Joe Simpson: and I guess we can say that maybe for a lot of people, that isn’t the day to day reality. They would have one car, they’d get used to it.

[00:06:50] Nonetheless, it reminds me of something that I a friend of the show Lorenzo Wood talked about in the past, and that’s the concept of design patterns. And how, if you think about it, the way we operate things and the way that certain products and technologies tend to become adopted at mass scale is when we have patterns that are understood and consistent. So

[00:07:15] Drew Smith: Yep.

[00:07:16] Joe Simpson: Cars became much more commonplace when you drove them via a steering wheel with a gear shifter that you use by hand and three pedals. A clutch, a brake, an accelerator, rather than, as in early cars, a lever by the side that operated the brakes, and a layout of pedals which differed from car to car.

[00:07:44] Drew Smith: And tiller..

[00:07:44] Joe Simpson: say, right, right. So, I think the point you’re making is, if, ADAS, as these systems really are, so Advanced Driver Assistance Systems, systems which are, they’re pertaining to make our driving safer and to help us out and make it more relaxing. Then if the aim is to do that, and if the aim is to get greater adoption of these systems, then surely the first question we should be asking is, shouldn’t there be some standardized mode of operation, or means by which you activate them, such that it is consistent from car to car?

[00:08:24] Drew Smith: and what I find fascinating is that, a company like Mercedes Benz is rushing to standardize turquoise teal as the external lighting color to, indicate when level 3, level 4, level 5 driver assistance systems are activated before anybody’s worked out any kind of intelligible standard for, How the systems actually, how you interface with the systems in the vehicle.

[00:08:50] and. if we come back to the Hyundai and the Tesla, Hyundai does this thing where, in terms of its interface, they try and surface all of the various parameters right there on the steering wheel. So you can choose between regular cruise control. you can then have,speed dependent kind of start, stop, active, adaptive cruise control.

[00:09:15] you can then have lane keep assist, with cruise control. You can also have lane keep assist without cruise control on its own. then you have Okay, like what distance do I want to keep from the car in front and all of that is going on the steering wheel. So you spend an awful lot of time when you first get into the vehicle trying to work out what all of these various things are.

[00:09:41] Tesla, interestingly, sticks all of that into a set of submenus within the center screen. So you kind of set and forget how you want the car to operate and depending on whether you do a double tap or a single tap on the Column stalk.

[00:10:00] Joe Simpson: Mm.

[00:10:00] Drew Smith: it will activate one of two, two different modes. so once you’re up and running and you understand how the system works, the Tesla’s interface is great.

[00:10:12] But if you are in that mode of fucking around and finding out, having all of those options right there on the steering wheel, so you’re not having to dive through sub menus while you’re driving is really interesting, from the Hyundai point of view.

Joe’s experience of GM SuperCruise

[00:10:22] Joe Simpson: I think this is where I, it’s useful for me to step in and give some context with Cadillac Supercruise because I think, what’s most interesting about, Supercruise is it kind of the way it’s set up circumvents a bit both of those issues.

[00:10:37] So for those people who don’t know, Supercruise, does a bit what it says on the tin and I think that’s one of its big advantages.

[00:10:44] It is cruise, but cruise control, but super. and it only works on highways. It only works on about 250, 000 miles worth of LIDAR mapped, usually central reservation divided highway in the U S. So GM have been out, mapped this with LIDAR, certainty about it. So it’s very domain contained. So I think that’s the first point.

[00:11:13] Drew Smith: It only works on highways. Whereas. As far as I understand, Tesla’s system, or certainly FSD, which I think is a level beyond the one you were trying, can kind of work anywhere in a slightly shonky way. Joe, it doesn’t work. It doesn’t okay. Okay.

[00:11:28] shodky way. let’s be very clear.

[00:11:31] Joe Simpson: the Cadillac is domain constrained. So I think that’s actually useful in the first instance. It only works on highways. Then the way that the system operates is that you have to have cruise control engaged, again, via the steering wheel. And then, when it determines that it can work, which is a slightly unknown set of parameters, but it flashes up in the binnacle, Super Cruise Available.

[00:11:57] And then you press the button again to activate it, and you get, and I think this is a really important part of the system, a display in the top part of the steering wheel. So you’ve probably seen it, we can put a kind of an image in the, cut it into the video here. But there’s a kind of display in the top part of the steering wheel, which is either red, blue, or green.

[00:12:20] As it activates, it pulses blue, and then when it’s active, it’s green. And then it will say, Super Cruise active. And that’s green. And once that’s green, what I’ve found over several hundred miles is you learn to trust it’s got it, then you actually don’t need to worry about it. But then there’s an important extra factor, which I know that Tesla don’t have, which is it is monitoring you.

[00:12:43] And you can see just peeking over the top of the steering wheel hub, there’s a driver monitoring system. It does check that you are watching constantly. You have to be alert. And then if it bails out, so at times, and it bails out generally when it comes to a stretch of highway that hasn’t been mapped, there are roadworks, or something unusual happens, for instance, there’s an emergency vehicle, or a vehicle suddenly appears from the hard shoulder unexpectedly.

[00:13:15] It will bail out and it does so in quite a dramatic way. The display on top of the steering wheel will flash red. It pulses the seat and it tells you that it’s disengaging. And in the systems, one of the most interesting points is to understand what happens at the moment of disengagement, and how collaborative the system is with you.

[00:13:38] And I found Super Cruise, of all the systems I’ve experienced, actually, the most convincing in these, in that because I was genuinely paying attention, because it was telling, it was making me pay attention, but also because of what it did, I was back with the wheel doing the steering doing the driving within a kind of within a second or so That’s notwithstanding that the system is also using radar stereoscopic cameras a suite of sensors rear radar so a level of sensor suite and redundancy that again the Tesla doesn’t have.

[00:14:16] so I found it a fairly convincing system and I would stress it’s almost because of the limitation of the operational domain. And it’s quite pessimistic approach to things. Um, you know, roadworks. No, no, you take it back. There might be something we haven’t mapped here. This, I can’t do this. And that actually makes it, when you’re using it, a less stressful system.

[00:14:40] Because when that green bar’s on, you’re like, I’ve got to the point where I think this is pretty good. Yeah.

How do advanced driver assistance systems fail?

[00:14:45] Drew Smith: And I think, and I think, one of the things that we talk a lot about in, in service design, and I think it’s very much at play here is how do these systems fail? what kind of experience are you subjected to as a user when the limits of the operational domain are Now, GM, actually makes that experience far more, trust building by being very clear with you about when those operational limits are reached.

[00:15:15] obviously because neither, Autopilot or Hyundai driving highway driving assist, whatever it is to, Neither of those systems are operating within constrained domains. So that operational domain is basically always shifting depending on the conditions in which the car finds itself. And what that means in the Hyundai is It felt very much like being in dialogue with the system.

[00:15:50] It was, relatively speaking, pessimistic.

[00:15:54] It was,

[00:15:55] cutting in and cutting out, fairly frequently. And I likened it to, kind of, a toddler,on a bike with training wheels. it’s wobbling along and it’s, and,

[00:16:06] Joe Simpson: it go for a few seconds and

[00:16:08] then

[00:16:09] Drew Smith: you know, you kind of let it go, but you never really, really let it go.

[00:16:13] Joe Simpson: Hmm,

[00:16:14] Drew Smith: Now, that to me was actually a less stressful set of interactions, although it required a far greater degree of ongoing vigilance than my experience with autopilot, which was, it was like you’ve taken the training wheels off, and that toddler was going for its first ride without the training wheels, and it’s like, I’ve got this, I’ve got this, I’ve got this, I’ve got this, Fuck!

[00:16:41] I haven’t got this! Right, and, and it just, bang, you’re, you’re gone.

[00:16:45] Joe Simpson: one of those YouTube meme videos.

[00:16:48] Drew Smith: and all of a sudden you’re back in control of the vehicle.

[00:16:52] That actually made for quite a stressful experience, particularly, when you’re doing like 75, 80 miles an hour, keeping up with traffic on freeways around, around Los Angeles. Like, when all of a sudden it let go, the stakes felt that much higher, whereas because with the Hyundai, as I said, You’re in this dialogue with the vehicle. It, it, it feels like a, dare I say it, a co pilot

[00:17:17] Joe Simpson: yeah.

[00:17:18] Drew Smith: rather than an, an, an autopilot.

What’s the point of these systems?

[00:17:21] Joe Simpson: And Drew, I have a question for you relative to these three systems that we’re discussing. Which is a bit of a blunt one. What’s the point?

[00:17:30] Drew Smith: Good.

[00:17:31] Joe Simpson: As you see it, what, beyond perhaps the one for me, which came to mind with the Cadillac, which was the, There is a magical quality to it. Watching a car steer and do things itself is something we’ve grown up with in films and seen and perhaps all hope for one day.

[00:17:50] And there is this, idea of brilliant technology being indistinguishable from magic. And there was that slightly magic thing, but I felt it was a novelty. And after a while I found myself asking, but what’s really the value of this? And it sounds like it’s magnified to a much greater degree with the Tesla and the Hyundai.

[00:18:08] Drew Smith: Yes. and it’s a dangerous novelty. Um, because if I come back to, if I come back to, say the cruise control system in my Mercedes, it only operates above, I think it’s 40 kilometers an hour.

[00:18:26] Joe Simpson: Mmm,

[00:18:28] Drew Smith: I have to maintain total vigilance over that vehicle all the time whilst I’m using that system. if a car pulls out in front of me and I’ve got cruise control set at 120km an hour, I’m just going to run into the back of that car.

[00:18:41] The challenge I feel with these systems is, particularly with the systems where the operating domain is not clearly defined, is that they do encourage people to fuck around and find out.

[00:18:58] Joe Simpson: Right, take risks.

[00:19:00] Drew Smith: just how far can I push this thing? And, you know, in the interest of research, I pushed them pretty far.

[00:19:08] And, you know, I, I, I tried autopilot on the, the winding roads up through Topanga Canyon and down into Calabasas. And, boy,

[00:19:17] that

[00:19:17] Joe Simpson: did it do?

[00:19:18] Drew Smith: That was amusing at one o’clock in the morning. Again, because looking at a hairpin bend, the car was like, I’m doing it, I’m doing it, I’m doing it! Oh my

[00:19:27] god, I’m not doing it!

[00:19:29] let’s assume that one of the use cases or one of the consumer needs that these systems are meant to answer is for more relaxed driving.

[00:19:43] Joe Simpson: Mmm,

[00:19:44] Drew Smith: I would argue that the reality of the extra vigilance required. To manage these systems, to determine when and when they will and when they won’t work actually leads to a more stressful driving experience.

[00:20:04] Joe Simpson: I’m glad you said that because in the interest of research for knowing we were speaking about this tonight, I went out and read A few commentators and a few sort of, informed viewpoints on this, and several people said that. Several people said, this actually makes it harder work. I have to pay more attention, it’s more mentally taxing.

[00:20:27] if it’s not making it more relaxing, what is it doing?

[00:20:34] Drew Smith: Well, is it, is it making it safer?

[00:20:36] Joe Simpson: Is it

[00:20:37] Drew Smith: be curious, I’d be curious to know if there’s any, if there’s any data on whether, Advanced driver assist systems, not the city based stuff, like, you know, automatic emergency braking in city environments and, reverse emergency braking and stuff like that.

Do these systems make our roads safer?

[00:20:52] Drew Smith: but I’m saying, in, in actual highway driving, are these systems actually making us safer or not?

[00:20:58] Joe Simpson: And I think this is where we get into the regulatory question, which we promised we wouldn’t touch because we don’t have expertise on it. But I think we should be clear. What we’re talking about here is a combination of essentially adaptive cruise control, so distance based cruise control, in combination with a, lane keeping aid or a lane centering aid.

[00:21:21] And, at a very simple level, the way that the bodies such as IIHS and Euro NCAP feel about this is that they’re saying, Okay, a system that is able to maintain a distance from the car in front and should that car stop quickly. Also, stop quickly and respond potentially quicker than a driver can and bring you to a stop. That’s safe. You won’t run into the back of that car.

[00:21:44] And a system which is able to keep you in a lane and even should you get distracted, stay in that lane and stay in the centre of the lane and not ingress into another lane or collide with a vehicle alongside it. That’s also safe. Now, in theory, that’s great.

[00:22:00] In theory, that should be reducing accidents. As you say, it’s quite hard to find data on this that verifies that it is reducing accidents. And when you combine those two things with variable interfaces, and marketing such as Autopilot, it seems, and I’m stating the obvious here, that you then have a recipe for people fucking around and finding out, and unfortunately finding out to their cost that there are deep limitations to these systems as the many Tesla accidents that are reported in the media, and sadly deaths, while these systems have been engaged highlight. So I think, I come back to my question of So what are they for? The question of safety is a sort of unanswered one. but I think there’s a really interesting question which sits around the idea of attention. And we’re in a world now where our attention spans are really fought for and the average person’s sort of ability to focus on one task for a long time is diminishing because of external factors.

[00:23:15] Drew Smith: Right.

[00:23:16] Joe Simpson: Driving and driving for a significant period of time is actually now one of the most continuous focused activities that any of us will do. If you choose to drive to, I don’t know, the sort of, you choose to drive to the other side of Sydney and you don’t stop, that’s probably the greatest and focused piece of concentration on the single task that you will do in your day.

[00:23:44] Drew Smith: Are people actually focused solely on the task of driving that

[00:23:47] Joe Simpson: Of driving. So, so this is, this is where it gets very interesting with these assistance systems. on one hand, you say, people are so distractible these days. Our attention spans are reduced. you might have people in the car, you might be doing phone calls, and you,you might be tempted to pick up your phone.

[00:24:04] Drew Smith: That There’s an argument that these systems act as if you like, um, you know, they’re, they’re guardrails to stop you going wrong, to catch you when you screw up, to mitigate when you basically lose attention. but if they don’t provide the safeguards that they’re supposed to in a consistent way. If they don’t, if they don’t, if they’re not foolproof, then there’s obviously big holes that start to appear in that safety net. And it is, it is worth repeating outside of keeping a gun in your home, um, or living in a state with open carry laws, getting in and driving a car is probably the most lethal thing anybody does during their day to day life.

[00:24:58] You know, you’re talking about getting into and piloting something that weighs probably at a minimum, 1, 200 kilos and anywhere up to 2, 800, 3, 000 kilos these daysand we’re introducing systems that confuse who is in control of that weapon.

[00:25:18] Joe Simpson: right. And this

[00:25:20] Drew Smith: like a pretty shitty idea to me.

[00:25:24] Joe Simpson: the context of, come back to America, where we both use these systems, fatalities and injuries are going up.

[00:25:36] Drew Smith: Right,

[00:25:36] Joe Simpson: Despite the fact that, manifestly, cars are safer. Cars are safer than they’ve ever been. So why are those fatalities going up? they’re going up because more people, pedestrians and cyclists, are getting injured and killed.

[00:25:57] Drew Smith: yep,

[00:25:57] Joe Simpson: And you could attribute some of that to the fact that vehicles are getting bigger and heavier, as you’ve just said. But some of it could be that the people in them are more distracted.

[00:26:05] But the other thing is that given that our vehicles are safer than they’ve ever been from a sort of passive crash perspective, what’s happening in those vehicles in terms of, distraction, confusion, diverted attention that is maybe causing people to make mistakes or lose control of the vehicle.

[00:26:28] And I’m not here to be on a high horse about saying we need to ban all these systems, far from it. I’m not here to say we need to take screens out of the vehicles, but I’m saying that It feels to me like an extension of our conversation with CES, where at the moment, technology is being applied. It’s not being applied consistently, and it’s not being applied with the benefits that we might have been led to believe it could provide.

[00:26:55] Drew Smith: Yeah,and I think this lack of, it comes back to this lack of standardization, and this lack of, I would say, a considered and coherent response across the industry on how these systems are designed and deployed and communicated about.

[00:27:17] And, and we are, we are running so far and so fast and so loose. I would argue with these things. before we’ve really got an understanding of how to deploy them safely, locked down,the, perhaps the, I’m not so sure it’s a counterpoint to all of this. because there are plenty of people who will take issue with what we’re about to talk about is something like Waymo, right?

Drew’s experience of a Waymo Robotaxi

[00:27:48] Drew Smith: So for people who aren’t familiar, Waymo, is a robo taxi company, currently operating in a couple of cities in, in the United States. And thanks to Ed Niedermyer, who’s a friend of Looking Out and the person you absolutely want to listen to, to understand

[00:28:07] like the regular, likethe regulatory environment, for example, around these things, or if you really want to go soup to nuts on the engineering of them, his podcast, The Autonocast is absolutely superb on this topic.

[00:28:19] But thanks to Ed, I got access to Waymo, in San Francisco. so before we get into having a conversation about my experience there, why don’t we take a break and watch a little bit of video that I took from that experience.

[00:28:36] Uh, well, hello. Uh, welcome to San Francisco. I am currently standing on the corner of 26th and Sanchez. I’ve just had a delightful breakfast roll at, uh, the Noe Cafe over there. And I’m about to order my first ever Waymo.

[00:28:58] Drew Smith: Thought I might take it down to the X headquarters. Um, I can’t take it to Cruise, which has fallen off a cliff. But I can take it to X, uh, which is another cautionary tale of what happens with inflated expectations and technology. So, uh, let’s see how we go. I will see you again when the car turns up.

[00:29:19] All right, so the car should be turning up any minute now and I have to admit there is a little bit of social pressure associated with like how this thing exactly is gonna work. Um, not really sure what it’s gonna be like when it gets here. Dare I say what people are gonna think of me? I just took a photo for a woman in the middle of the koi fish there and uh, she said be safe in that thing.

[00:29:45] Yeah, I wonder if my travel insurance covers this, so let’s see how we go.

[00:29:49] Oh, there we go. And here it is. It’s about to turn right.

[00:30:00] Alright, what’s gonna happen?

[00:30:08] Okay. And now it’s pulling over.

[00:30:15] Oh, and it’s gonna keep going. Where’s my car going? Alright, here we go. Pulling over to the curb. And, hazard lights on.

[00:30:34] Riders

[00:30:39] are only on when you’re connected to rider support. So sing your heart out. We can’t hear you. You can also use the app for passenger screening to speak to a rider support agent at any time. In the rare case of an emergency, please keep your seat belt fastened and remain in the car unless there’s an urgent need to exit.

[00:31:04] Rider support will connect with you and provide assistance. Finally, relax and for riding with us.

[00:31:13] Okay, so this is a lot, like a lot has just happened and I realized I lost a little bit of video there when I changed the camera angle. So I get into the car, there’s this very pleasant voice. And, oh, oh, oh, oh, oh, okay.

[00:31:29] Um, kind of welcoming to the car by name, and uh, encouraging me to buckle up my seatbelt. I didn’t quite get that done in time, so it started beeping at me. Um, and, I have to say, it’s, it feels, and I never thought I would say this, so far, and we’re what? Like, a couple of minutes in, at most. This feels like the most natural thing in the world.

[00:32:02] And that’s amazing. Like both amazing and terrifying.

[00:32:08] I have to say, this thing is far more confident at driving on the wrong side of the road than I am, that’s for sure.

[00:32:16] Also, these must be the most reliable Jaguars in the world, right?

[00:32:24] What an absolutely mind blowingly incredible experience this is.

[00:32:30] So, one of the things that’s really interesting is that at the back of the center console, and it’s going to be a little hard to capture in this light, um, we have this screen which is not giving full situational awareness, um, and it’s certainly not like Tesla’s for example where they are showing, you know, the difference between trucks and cars.

[00:32:58] Uh, but you can see the other vehicles around you. It’s really nice because it lets you know that it, that it is seeing them. Uh, but just a little while back there, we passed a traffic cone and the traffic cone was rendered in full color. So you know, things that represent a hazard, uh, being represented, um, sort of with, with greater salience on the screen, I guess we could say,

[00:33:26] I don’t know what the play music button does. Let’s, Oh, that’s neat. Have we got, uh, this feels appropriately San Francisco. Oh, that’s so cool. That is so cool. Well, it hasn’t started yet. Where’s the volume? Oh! Oh, what a turn! Well picked Waymo. Very, very well picked.

[00:33:56] That was cute. The guy on the crosswalk just waved us across. I’m not sure Waymo, I’m not sure Waymo understands those kinds of pleasantries. Interesting point that, you know, we so often interact with the drivers of vehicles when we’re pedestrians if, you know, we’re happy to cede the road to them or we want to stick our hand up to suggest that they should stop in preference to our needs.

[00:34:24] I mean of course Waymo is going to stop in preference to our needs. But yeah, that guy waving us through. Waymo don’t care. He’s just gonna wait. Or she. Or they. Who knows what Waymo is.

[00:34:37] Alright. Oh, there’s the building. There’s the headquarters. Down ahead. Almost at your drop off. Check the app for walking directions.

[00:34:50] For your safety, the doors will remain locked when we arrive. Pull the handle twice to exit. The first pull unlocks. The second opens the door.

[00:34:59] I wonder if Elon’s paid the rent today. Is there anybody there? Is there anybody home? Is it just like five malnourished engineers sat behind a glowing bank of screens drinking soylent green. Who knows? Let’s find out. See if I can find a basin or a toilet to carry into the foyer. Let that sink in. Alright.

[00:35:34] So, drop off in one minute. Where is it gonna leave me?

[00:35:42] Uh Turn Oh, okay. We’re going to go straight ahead.

[00:35:51] And, Oh, there’s a car right there. It’s gonna, where is it going to lead me? Oh, we’re going around the block! Oh, we’re going for a three minute drop off now. Right, so, because it couldn’t find a place, I’m guessing, that it was happy to pull over, we’re gonna go around the block again.

[00:36:13] That’s interesting

[00:36:14] check the app for walking directions.

[00:36:21] For your safety, the doors will remain locked when we arrive. Pull the handle twice to exit. The first pull unlocks, the second opens the door.

[00:36:30] Oh, I see. Finding a spot to pull over. Yes. Good Waymo. Pulling into a back alley. You happy with this? You happy? You happy Waymo? You’re here. Please make sure it’s clear before exiting. All right, let’s get out of this thing. All right, pull twice, open the door, and I’m gonna get out. All right, bye bye Waymo.

[00:37:10] And,

[00:37:32] there it goes. Off into the fog. Off into the fog of San Francisco. Wow. Just incredible.

[00:37:41]

[00:37:42] Drew Smith: Alright, I want to capture some really quick impressions before the feeling leaves me. That is by far and away the most batshit, crazy demonstration of technology I think I’ve, I’ve ever experienced. I have to admit that I got into that thing based off my experience with, I mean it’s such an unfair comparison isn’t it? Autopilot and what Hyundai’s doing thinking I feel a little sketchy at times the reality is there was absolutely nothing sketchy about that experience at all and I think it shows just the massive gulf that exists between what we’re able to buy, uh, as consumers, and the stories that Tesla in particular are telling about the capabilities of its products, and what the reality of real autonomy looks like.

[00:38:45] And, just gonna put it out there, it’s pretty fucking cool.

Interlude

[00:38:51] Drew Smith: So before we jump back into the main section of the podcast, I figured while I was in San Francisco that maybe my reaction to this whole Waymo thing was just a little bit overblown.

[00:39:03] So, I decided to take a second trip with my friend Greg. Now, I won’t show you the whole trip, we did record the whole thing, but I did want to capture Greg’s reaction after the fact.

[00:39:16] Greg is not a car guy like me, he’s not necessarily a technology guy either. He’s just a regular guy who lives in the Bay Area, deals with that traffic all the time. And so I thought it’d be really interesting to get his point of view on how he felt about the ride. Here’s what Greg had to say.

Greg’s reflections

[00:39:35] Greg: I think I thought it would be a little bit more, like, giving up control, but I just felt at ease just sitting back there. And, and in terms of the way it drove, like, what, what kind of human driver would you equate it to? A good, safe Kind of moderate driver. Right. I would do it again. Hahahaha Was there anything that was disconcerting about it?

[00:40:15] No, I think because our trip was short it wasn’t long enough to be. I don’t know how I’d feel on the freeway, like on and off ramps. I think there’s some freeway interchanges that are, you know, six lanes and then there are two, so those things might take a little while to get used to. It’s just a short city hop.

[00:40:31] They’re perfect. And what effect do you think, like, riding with somebody else had on it? I think that was huge for the first time. Right. I think that getting in by myself for the first time would have been a whole different experience. Like, at least I’ll die with Drew. .

[00:40:50]

Joe’s reflections

[00:40:51] Drew Smith: So, uh, what do you reckon, Joe?

[00:40:55] Joe Simpson: well, it looked amazing. I am, I’m a bit in shock. We’ve just been talking about ADAS systems and essentially our sense that Super Cruise excepted they’re a bit sketchy. And then here we have A fully autonomous car without a driver in, navigating the streets of San Francisco, and you, who I think people will, maybe label as a skeptic at times on certain subjects, essentially cock a hoop at the experience.

[00:41:26] It looked like, at least from afar, my sense is that It was impressively competent, there were no mistakes, you rode for several minutes, it did a great job, it did it quite assertively, and that somehow, very quickly, it made you feel comfortable with it, and as if, yeah, I’ve, this thing’s got my back, it’s, you know, I’m, I’m trusting it. Talk us through your Sort of, what was going through your mind as you were getting in and then you started riding?

[00:42:03] Drew Smith: for a start, this was,this ride was set against the backdrop of the Cruise cluster curse. and of course all of the challenges that, Cruise faced, the cruise brought upon itself, which ultimately led to its license to operate being canceled as far as I can remember.

[00:42:19] and so I had an expectation that, Okay, yes, the system would nominally work,

[00:42:29] Joe Simpson: mm.

[00:42:30] Drew Smith: um, but there were, there would be moments, during which I would feel, I guess, sort of particularly unsafe. And,

[00:42:41] Joe Simpson: did you feel unsafe at any point?

[00:42:43] Drew Smith: Never. Never. And,from the way the app signals the arrival of the car to How the, the voice within the vehicle welcomes you and sets expectations and helps you understand what’s going on to the way that the displays within the vehicle Which you could see in that video, are showing a level of situational awareness, but not everything.

[00:43:18] Joe Simpson: Right.

[00:43:19] Drew Smith: those displays are showing you enough to know that the car is seeing the critical things. the displays are showing you, for example, traffic cones. In bright orange, right? We can see the traffic cones, but you know, buses, trucks, cars, um, and humans are all sort of rendered in sort of soft greys.

[00:43:42] Yeah, so you know it’s focused on the right stuff. If you like. what is interesting obviously is that the car or the system doesn’t take any liberties like a regular cab driver would. So in the opening of that video you saw that, the car I thought was going to pull up, but it wasn’t satisfied with the space available to it in front of where I was standing.

[00:44:11] So, it drove on to find a safe spot, um, and I had to, I had to, I had to kind of sprint after it.

[00:44:20] Joe Simpson: and did, did it, did it sort of tell you it was doing that in the app? I mean, this feels like potentially the slightly grey area of the system right now, the whole pick up and drop off, and where it goes on the street when there’s not space, and how it makes that feel safe, and how it communicates to you.

[00:44:40] Drew Smith: For sure, so with familiarity, I would have known that, when the car was finally parked, and available for me to get in, I would have received a notification around that. But, because this is my first time using it, I’m like, oh shit, the car’s running away!

[00:44:57] Joe Simpson: Yeah, you’re used to an Uber leaving within

[00:45:00] Drew Smith: must catch my car,

[00:45:01] Joe Simpson: there. Yeah.

[00:45:04] Drew Smith: is interesting about that, and I guess, look, the, what would happen if it was a cab driver is they’d just double park, they’d just pull up wherever.

[00:45:11] and, Waymo’s clearly made a decision that’s not in its best interest to do. Where that gets interesting, though, was at the end of that journey. as we saw in the video, I ended up having to go quite a long way around the block because the first parking space that the vehicle thought it had available to it, again, it was not satisfied.

[00:45:31] So it added, I don’t know, probably another two or three minutes onto the journey until it could pull up at a space, which it considered. Um, again, like a human cab driver just would have pulled up and gone, yeah, whatever. We’re good. Whack on the hazard lights and I’d get out. So, I think Greg’s reflection was really interesting.

[00:45:58] I said, when I, when I asked him, what kind of human driver is this? I think he said, a safe, moderate human driver. and I think your word around assertive is good. the programming of it is definitely not aggressive, but intentions are clearly signalled and acted upon, and you’re never really left in any doubt that this thing has control.

[00:46:31] Joe Simpson: Yeah, it didn’t strike me in watching those videos, it didn’t strike me that it was hesitant or pussyfooting around at any point. And I feel like that was probably a key part of building trust. like a confident driver that knows what he’s doing, or she’s doing, and knows the area.

[00:46:49] Drew Smith: exactly, and there was one section in that video where the car pulled what I thought was it was a punchy move, you know it was it was not that it was not the move of a demure computer It was somebody going: no, there’s a space there. It’s safe for me to go. I’m gonna go for it

[00:47:08] Joe Simpson: yeah.

[00:47:10] Drew Smith: And of course, these are all things that those systems that we were talking about before simply can’t do

The misdirection of the SAE autonomy levels

[00:47:17] Joe Simpson: I think a couple of things about this are one is that when we compare that system to the ones we’ve been talking about before, I think there’s been a sort of direction of travel in terms of mindset and message that started off with, we’re going to get to autonomous. This is a few years ago.

[00:47:36] We’re going to get to autonomous driving quite quickly. And on the way there, we’re going to go through these kind of SEA levels of assistance to arrive at that. And I perhaps, perhaps naively thought that. One of the things that would do would help condition people to being used to cars that progressively, did things themselves.

[00:48:02] But actually, I feel like what we see in the experience of the ADAS and the Hyundai and the Tesla compared to the fully autonomous Waymo is a world apart, which doesn’t really have any sort of relationship.

[00:48:20] Drew Smith: Now, now, Ed has a really interesting take on this, and when he expressed it to me over breakfast in Portland, I was like, oh my god, this is fucking genius. Part of the problem with talking about autonomous cars, and levels of autonomy in cars, is that We bring a whole bunch of assumptions about what a car is

[00:48:49] Joe Simpson: mmm,

[00:48:50] Drew Smith: and we mash that up with some fairly vague ideas about what autonomy means, and there’s not even, there’s not even firm agreement within the SAE levels about what level 1 is, what level 2 is, 3, 4 or 5, you know, people talk about level 2 plus or level 2 plus plus.

[00:49:09] Like, what is that? You know, how, how as consumers are you ever supposed to get a clear idea About the capabilities of the vehicle and Ed said, just talk about them as robo taxis. You know, or talk about it as a taxi bot. As soon as you call something a robot.

[00:49:28] Joe Simpson: it changes the

[00:49:29] Drew Smith: don’t talk about it as a car it fundamentally changes your mental model. And this is something that the Ed’s been talking about within sector for quite a while. And

[00:49:39] Joe Simpson: Cars are things we own and love and fetishize and use to project our own image and drive in ways which perhaps reflect our personality or kind of create body language. A robot, RoboTaxi, is something that assists us and takes away chores and does things that, you know, are just menial and that maybe we don’t want to do, or that just, in this case, take us from one place to another.

[00:50:08] Drew Smith: and there’s an expectation with a robot that we don’t take control of it, right? that it is programmed in such a way that it serves our needs and it serves our needs reliably and safely. With a car, hmm, like that’s something that for 120 years we’ve controlled.

[00:50:27] Joe Simpson: Yes.

[00:50:28] Drew Smith: So.

[00:50:29] Joe Simpson: the machine. Oh.

[00:50:36] Drew Smith: dab of oppo, like all of that stuff. And that’s, that’s resolutely not what these things are about.

But does Waymo make sense?

[00:50:45] Joe Simpson: And yet, I guess to just close on this, if I step back and I put a bit of a journalist hat on here, I would say, watching those videos, it’s deeply impressive what this vehicle can do, and as a RoboCab, yes, I can see how these things may have a role in the future, may become at some point in our lifetimes, something which we expect to see on more city streets.

[00:51:10] If I take away the, novel amazement at it doing the driving so competently, I feel there’s a load of edge stuff, which is quite important. Like, where it pulls up, how you get in and out, you being able to say, I actually just want to bail here that will need sorting and potentially in that trade off against a cab, it’s probably not doing a better job than at the moment.

[00:51:40] Is that your sense too? If you step back from it, or was it so enthralling and amazing? You’re like, no, no, no, a hundred percent. It was better than taking any cab I’ve taken in my life.

[00:51:49] Drew Smith: no, not necessarily. and I think there’s also a whole bunch of cultural, like social and cultural implications that come into play, which we could probably fill an entire show with a conversation about. I think it’s an extremely impressive demonstration of where we’re currently at with technology that’s considered safe to put on the roads.

[00:52:14] And it’s useful as a comparison to show just how far away from that the systems, that OEMs talk up, are. do I think individuals riding in robo taxis are the solution to the problems of traffic and moving people from A to B and stuff from A to B in our cities? No, I don’t. I don’t think it, I can’t imagine a city full of Waymo’s being any better than a city full of, individual drivers in their cars. Um, but at the level of a technology demonstration, it’s just hugely, hugely impressive. And it shows what a part of a future mobility system could look like. But it is not, it’s, it’s not the be all and end all.

[00:53:06] and, I don’t think, I don’t think Waymo positions it that way either.

[00:53:10] Joe Simpson: And I think all credit to Waymo and Alphabet for really what they’ve done. They’ve done this steadily, quietly, consistently, and with, a sort of safety culture around it. I think that maybe other, players in the space, mentioning their names, Cruise have,maybe taken liberties with.

[00:53:30] But through

[00:53:31] Drew Smith: to not drag a woman 20 feet or drive into wet concrete or shut down outside a concert because the cell network went down. Anyway, Joe, you were saying? Messily.

[00:53:47] Joe Simpson: prognosis? What’s your, paint me a picture 20 years from now. I think you’ve said what you needed to say about the Waymo. You’ve said, I can see a Robocop future. But with the cars, let’s go back to the cars and the cars we drove. How do you feel that’s going to play out?

The need for industry standards

[00:54:03] Drew Smith: I think until there’s a, as I said before, I think until there’s a set of kind of standard interfaces, clear definitions of operating domains, standards around security, data privacy, where these systems work and how they work we will continue to run into Situations where the technology lets us down.

[00:54:31] Joe Simpson: Yeah.

[00:54:31] Drew Smith: and fundamentally undermines our trust. Because, you have Everybody’s running at this issue in slightly different ways, promising slightly different things. Offering up slightly different constraints, you know, think about the system in the Mercedes S class, which can operate in the state of Nevada up to 30 miles an hour so long as it’s not raining, which I mean, in Nevada, not such a problem, but. Okay, like, how useful is that? Back to our question, way back in the podcast. So, yeah, my, my, my personal take, my non expert take is that until This stuff is standardized and everybody’s heading in the same direction.

[00:55:21] We will keep getting unpleasant surprises. We will keep undermining trust in the technology. In a way that, Waymo and Waymo’s approach to this stuff is trust building.

[00:55:34] Joe Simpson: And I think that’s a really neat sort of circle of the square for car brands for the OEMs. I feel like the question here is how to build this stuff slowly, sensibly, in a way that makes sense and doesn’t over promise that allows consumers to actually gain value and build trust and Us to all drive safer

[00:55:59] While there’s basically and frankly led by tesla a marketing war going on With a race to bring additional features New functions and names which over promise and I think for that for all our criticism I think what gm have done with super cruising cadillac is the right approach and I think for anyone who wants to understand where The best in the business is in these systems right now, that’s the one I would advocate you try.

[00:56:29] Sadly, you can only try it on a Cadillac product in the U. S. on certain roads. So if you’re listening to this in Europe, probably not gonna happen.

[00:56:38] Drew Smith: Sorry, babes.

Outro

[00:56:42] Drew Smith: Uh, that brings us to a close of this, episode 15 of Looking Out The Podcast. if you’ve liked this show, uh, go on, hit that thumbs up button, uh, if you’re

[00:56:56] Joe Simpson: smash that bell!

[00:56:58] Drew Smith: Yes, smash that bell. it helps enormously with the algorithm as if we really care about the algorithm. if you’ve been listening to this in your favorite podcast app, why don’t you share it with somebody?

[00:57:09] Spread the love. We love spreading the love. before we go, I’d love to do a massive shout out to Chris Bonelli uh, from Waymo and to Ed Niedermyer for making my ride in the Jaguar iPace RoboTaxi, possible. It

[00:57:27] Joe Simpson: thank you to

[00:57:27] Drew Smith: absolutely epic, so thank you so much for organizing that. Coming up in our next show, we’re going to be addressing another vexed topic. We love vexed topics on looking at the podcast. We’re going to be looking at the new Macan EV! And he

[00:57:44] Joe Simpson: photoshop pens out

[00:57:47] Drew Smith: he he

[00:57:51] Looking Out the podcast was written and presented by Drew Smith

[00:57:58] Joe Simpson: and Joe Simpson.

[00:58:00] Drew Smith: This is Drew Smith, thank you so much for listening, and for watching

End Note

[00:58:03]

[00:58:03] ​

---

Looking Out:

Subscribe to: Looking Out - The Newsletter

Subscribe to: Looking Out - The Podcast

---

---

---