Intro

SPEAKER_00 0:00

Hi, and welcome to the second episode of A Case for Debating podcast. I hope you enjoy this episode. Joining us today is my friend Henry. Hello. I'm so excited to be on this. Okay. Would you like to introduce what we are debating today, Henry?

SPEAKER_01 0:20

So today we are debating if we should have all self-driving cars on the road or a mix of the two.

SPEAKER_00 0:29

Yeah, so if normal cars should eventually be banned. So we both wanted to do a certain side. So this is the first episode where we actually had to flip a coin. Henry won the coin flip. So Henry will be defending the side that self-driving cars should eventually completely take over the world. And I'm defending the side that we should have cars that humans are able to drive. We have the results for the

Last week's results

SPEAKER_00 1:05

last episode, which were me and Alton debating about whether AI will be able to be stopped. And if it can be stopped or if it's not gonna be stopped. So, Harry, do you have a guess on who won?

SPEAKER_01 1:22

Um That was kind of a hard one to decide, but I think I'm gonna guess that AI will not be stopped.

SPEAKER_00 1:37

Yes, you are correct. That is what the voting form has said. There were only two responses, and they both said team unstoppable. So hopefully next episode we'll get a few more responses, but that's fun. I won that last episode. Okay. So without further ado, let's get into it.

Hearing from both sides

SPEAKER_00 2:16

Okay. First, I feel like with having all self-driving cars, cars just get so much more expensive. Like you have to have these super expensive hardware and sensors, and they have continuous software updates. And for all cars to be self-driving, eventually they'd have to have some like database, I feel like, that all the cars would communicate on. But I feel like it's gonna be harder for people to buy self-driving cars because of these costly sensors. So people can't who can't afford these more expensive cars could be left not being able to transport themselves.

SPEAKER_01 3:14

Yeah, and uh I think you are right about self-driving cars being expensive to buy. But I think that since people probably aren't going to be using self-driving cars, if you bought them individually, you're probably not going to use them 24-7. So um if you look at Waymo in Los Angeles and San Francisco and many other cities throughout the US and other countries, they own their whole fleet of self-driving cars. And I think it's a bit more expensive than Ubers, but it's basically the same as calling a taxi and having it come to you, and they are everywhere, they're super easy to use, and I think that um they could be ideal for someone who maybe doesn't have their own car yet, who doesn't know how to drive, or who doesn't feel good about driving.

SPEAKER_00 4:11

Yeah, so you're saying that they're everywhere, but I like I've been in Waymo's in Scottsdale, Arizona. But one thing with Waymo's is they are only in certain parts of bigger cities that have well-paved roads. These aren't replaces that get snow, and they'll probably have more updates and have self-driving cars be able to handle it. But right now it's looking like self-driving cars, like fully self-driving cars like a WMO, can't really go beyond like they can't go to dirt roads, they can only stay on well-paved roads in major cities where they know the roads and everything.

SPEAKER_01 5:02

Yeah, so I do think you make a point, and I also think that technology will get better throughout years. But right now, what we have is artificial intelligence, it it's only really good at a small range of subjects. We we right now are working our way, which um sure many people listening to this have heard of artificial general intelligence, which means that it's pretty good at a lot of things, basically. So I think throughout the years, once we develop AGI and eventually artificial superintelligence, um then I think these cars will be able to adapt to a wide range of cities and towns and rural areas.

SPEAKER_00 5:52

Yeah, I think I've heard I don't know when the release date is, but they're projecting AGI by like 2030 or like 2050, I don't remember what it is. But also the other thing with having a machine essentially drive our cars on our road, like they would work if we have a lot of good updates, which I expect we will. They will be able to handle these bigger things. But the only problem is we still have things like extreme weather, like maybe they can pull over and do stuff like that. But there's also like if they it there's power good flip failures, but also there's like cyber attacks. So if someone were to hack this big system that's essentially giving all the data to all of these self-driving cars, well, that could create a major incident and hurt a lot of people.

SPEAKER_01 6:54

Yeah, so I think that definitely is a concern for self-driving cars. However, they have worked really hard on the safety and they always have lots of emergency backups. Some people talk about if we can make them communicate with each other, we can make them just skip intersections entirely, but I'm not sh I don't think we could do that though, because you are you you do have a point there if a power failure happens on one of them, the whole system breaks. So I think it would be safer to do regular car driving, but these would still uh all surf driving cars would still be faster than a mix of the two because they wouldn't cause as many traffic jams, they wouldn't break too hard and they don't go over the speed limit, which a lot of crashes are related to going over the speed limit.

SPEAKER_00 7:56

Yeah, so we're really the thing is if we had all self-driving cars, all self-driving cars could probably just go like a hundred miles over the speed limit at that point. Because if we're all in this data connection, which I said could get hacked, like things like that could be bad, but then I feel like you could just increase the speed limit and just like staying at it wouldn't really matter. Another uh point I wanted to get at was there's a lot of people that are driving for their jobs, like rideshare companies like Lyft and Uber, those are all just like retired people from their job who still need to make make their living. Like there's truck drivers who that's their job too, and there's delivery drivers like more locally in neighborhoods, and that's how a lot of people make their living.

SPEAKER_01 8:54

Yeah, so job losses uh are one of the concerns for AI, and I uh wish that they didn't lose their jobs, but I also think that having this many self-driving cars would create tons of jobs. Like maybe when you drive self-driving cars, sometimes knew like if the technology isn't fully ready yet, you have to gather lots of models of the cities, so I think and also maintenance technicians for these cars and to clean them. So I also think that this will create a lot of new jobs too.

SPEAKER_00 9:34

That's interesting. Um I definitely see argument on that, like having jobs to adapt to it. But the thing with driving for a job is it doesn't require much skill. You really need a license and they need you need to be trained according to your company standards. You don't need like tech expertise or know how to operate whatever machines they would use to do maintenance on them. So I feel it's easier for people with less experience to do a driving job.

SPEAKER_01 10:11

Yeah, so I uh I also think that a lot of jobs that they produce, like cleaning the cars and maybe plugging them in at the facilities, because these probably will be all electric cars. I think that those are pretty relatively easy jobs, and also for training the car you just have to gather a 3D map of the whole city, basically.

SPEAKER_00 10:37

Okay, but the whole debate here is really having all cars self-driving. So, like in rural areas, people are we're like there's only dirt roads, people aren't gonna spend time just mapping out everything. And I feel like people there would benefit from just having a car that they could drive to their work that's probably pretty far away into a little bit of a major city. So, like if you had to drive two hours from your rural area into a city, you're going through t like in some parts of America, you're going through a lot of dirt roads and places that aren't really worth mapping out.

SPEAKER_01 11:27

And yeah, but like my point previously, once we develop AGI, um I think that it will that cars can pretty much drive anywhere self-driving cars. Um, but I think the main problem here that I take from this is the lack of infrastructure. Because you are right about the fact that we aren't just going to build a giant place with filled of tons of Waymo vehicles because uh it's kind of unreliable, but I think we could schedule them, like have a Waymo schedule to pick you up every weekday at 7:30 a.m. and then drop you off later in the day. I think that we could maybe arrange something like that, so then we could at five o'clock in the morning, then they could have all the Waymo's drive out. And I think it would make a lot of people's lives easier. And I also think that these technologies will get cheaper over time.

SPEAKER_00 12:38

I'm interested to see what happens when AGI comes out and like how we think of it as just like a general intelligence and how that will be incorporated into things like self-driving cars and autonomous vehicles. But I don't really know at this point what's that gonna do. So if it's something like it can navigate really well and see objects like falling power lines, that would be great. And like know where construction zones are and things. But I don't know how that's gonna incorporate, but right now I don't think we could at all replace all the cars with self-driving cars, especially because of these roller areas. But I if AGI turns out how we think it's going to be, and we're able to apply it to this, I'd still be worried about things like glitches under extreme conditions in snow, and they could misread things like construction zones, and there's even stuff about a Waymo, and this was like an off chance, and it's pretty likely that this won't happen once we have more advanced vehicles. But there was like a glitch in the Waymo system, and it stopped right in the middle of a crime scene.

SPEAKER_01 14:22

Yeah, so um I think you are right that even AI sometimes also does make mistakes and hallucinate, but so do human drivers too. And they do a lot more than AI. Definitely, for sure. Because human drivers can get drunk, they can get tired, um, they can get distracted. And while AI cars still can make some of the same mistakes, they make far less mistakes.

SPEAKER_00 14:53

I agree with that, but on a major level, 99% of the time, the AI will be smarter than the human for a self-driving car, but there's still so many people who like couldn't have self-driving cars, and they do some things that a human would almost like never do. Yeah, uh I think you are right about that.

SPEAKER_01 15:25

Because AI the there are multiple types of AI because we don't have a GI yet. The AI that uh that powers Waymo's is different from the large language models that we use, like Claude, Gemini, Chat GPT. And it's also different from the algorithms on social media, for example, that uh know what that can predict what you want to watch. I think uh you mentioned something about integration, but Waymo just launched the Waymo Ohi, which is their customized self-driving car vehicle uh that also is integrated with Google Gemini. So these cars also get updated a lot. And if part of the car fails, then there's usually a lot of emergency backups because they do take safety really seriously.

SPEAKER_00 16:25

Okay. I have a counter that because I wouldn't really actually call what self-driving cars to use AI, because I would call that a mixture of a vision system, LIDAR, and radar using sensors, which is coded in the systems. That's really what it is, and I feel AGI right now is more promised for these large language models. So I think that it could be developed for things like self-driving cars and autonomous robots and vehicles, but I feel like right now the focus is more getting AGI with our large language models. So I think it's gonna be a while longer before we see that with any kind of robot or vehicle.

SPEAKER_01 17:22

And Yana, you are right about that. AGI is still a long way off, and the artificial intelligence that we have right now still hallucinates, and I think again, this technology will get a lot better. So let's just assume that it's like 2032 by the time that it's all self-driving cars, and we have AGI and they can resist all kinds of things, and we have the newest computers. But right now, I don't think having all self-driving carts is necessary uh is necessarily practical right now.

SPEAKER_00 18:02

Yeah, so what do you think what year do you think it would be practical to have all self-driving cards?

SPEAKER_01 18:12

Well, again, like you said, it really does depend on um when AGI comes out, but technology will still improve even if we haven't invented AGI yet. So I think that it in select cities like Los Angeles and New York soon, I th or maybe actually maybe not. New York because I know that there's a lot of taxis there. But I think that maybe like in 2035 or 2040, it'll be practical to have all self-driving cars in um in cities that are really fond of them.

SPEAKER_00 18:58

I feel like that would be weird. Imagine just like driving to a city and then being, oops, can't use my car anymore. Park it at some like lot and then take a self-driving car into the city. That'd be pretty weird to go from driving your car to not be able to, just in a specific area.

SPEAKER_01 19:17

I mean, yeah, and you're right about that. That would be weird to do, which is why there's also a benefit of having the whole country have full self-driving. But I think that AGI and full self-driving is still a while off. I think we should wait until we have better technology until we actually uh put this put all of this technology into full use.

SPEAKER_00 19:46

Okay. That I'd agree that we still need more time for things to develop before we think about this. Once AG AGI comes out, as I said, uh I'm interested to see how these things develop in the future. Yeah, I'm definitely excited for that. Okay, that is the end of our main episode. After this, if you keep listening, will be our post-show. So stick around if you are interested in Waymo's. Okay, Hadri, have you ever been in a Waymo before? Yes, I have been in uh many Waymo's. In different cities or just one?

SPEAKER_01 20:37

Uh just two cities, Los Angeles and San Francisco.

SPEAKER_00 20:41

Cool. I've heard about them there. I've only been I've been probably like three times in Scottsdale, Arizona. So I rode one home to a hotel one night. Uh it was like a hotel farther away from the city, so we didn't really want to walk. It was late at night, it was super cool, and it pulled right up to the delivery lane. It was amazing. And then we went one other time that took us to dinner. It was super cool.

SPEAKER_01 21:11

Yeah, they definitely are super cool to ride in. Um and I also heard that in some cities you can use um like Uber to hail them.

SPEAKER_00 21:22

Oh yeah, they like partnered with Uber in some cities. Have you uh have you have you heard of like the other ones in San Francisco? I think I don't know if it's San Francisco or LA, but it's like called Ride or something. I forget what it is. It's like Oh, um yeah, I saw those.

SPEAKER_01 21:45

Um I don't remember what they were in a cart called though. I think they were like owned by Amazon or something like that.

SPEAKER_00 21:54

Oh weird. And then there's oboe taxis, which are from Tesla and a lot of them these days still have a person sitting in them and I think it's really weird. I've never been in them but I've heard about them and I think I've seen one in Scottsdale before.

SPEAKER_01 22:14

Yeah actually I think you can um like buy them and then like they join the fleet. It's kind of like it's kind of like just like I think you can buy one of those actually and you just get a car. I I don't know whether it can join the fleet and just become one or not. I also think it's interesting how they don't have any LiDAR that the right ones might but they rely I think mostly on cameras which I find interesting and also I don't think it's really as effective as LiDAR.

SPEAKER_00 22:45

Yeah I find it weird all Tesla system and I feel like it's not effective as effective too as like LiDAR and radar as Waybo's and other ones use.

SPEAKER_01 22:58

Yeah that's definitely yeah I find that odd when I first heard it they must have something against LiDAR because LiDAR is more accurate than cameras because it can see shapes. It's kind of like making a 3D model of whatever in front of you.

SPEAKER_00 23:19

I just think they just wanted because like a laymo when you see all those spinning radar things and the big LIDAR block it looks it doesn't look like a very polished car.

SPEAKER_01 23:36

Like it looks like there's a ton of sensors and stuff and it looks pretty ugly with the thing spinning on top well yeah I think although I think the reason for that though is that they're just using Jaguars the the Waymo Ohi is actual owned car. So I think it'll look a bit better. I ha I have a picture of it somewhere.

SPEAKER_00 24:01

Yeah. Jaguars I did yeah I don't think um jaguar even exists anymore. I think Waymo just buys cars from them and that's their only customer. I I think Jaguar still exists does it uh yeah so Tesla I feel like is trying to see if they can just rely on like Jript cameras to just put in so it doesn't look as technical as things like a Waymo or other cars.

SPEAKER_01 24:36

Yeah I think the Waymo Ohi is because they're making their own car not just adding their stuff into like I know with the Jaguars they just like modify the cars to add their sensors on them but with the OHI one the they uh can integrate it directly into it without needing to like place it above the sunroof or without it looking like um they just took a radar sensor and screwed it onto the outside of the car.

SPEAKER_00 25:08

Yeah. Well um I hope the listeners that tuned in to this post show this week enjoyed it and I hope you guys listening enjoyed the full episode and and I hope you come back next time.

SPEAKER_01 25:29

Got anything to say before we go Henry forget to vote we have it on the podcast website and there's a link in the show notes.

SPEAKER_00 25:39

Okay bye see you next time bye