Waymo shows off its futuristic “transportation as a service” vehicle | Hacker News Reader

Waymo shows off its futuristic “transportation as a service” vehicle

arstechnica.com

35 pointsrbanffy71 comments

It's the "Johnny Cab" from Total Recall!

https://www.youtube.com/watch?v=eWgrvNHjKkY

Waymo has demonstrated capability (in a limited sense) with full self-driving so I'm excited for this to take off.

What concerns me are the screens. They are going to play ads on a loop. Phone ads to car ads to house ads. We'll get the no billboard demands because your screens are the boards!

It's very exciting, but my first thought was that I'd look like jelly after that thing gets t-boned by a F-150 or some other monster truck. I'm very curious if it even has airbags, the console looks so minimal.

I know that these taxis will be in slowing moving city traffic, but nonetheless it's an automobile and should have (IMO) all of the safety of a modern day sedan.

"These technologies come with no compromise on safety. All SEA-M-based vehicles will meet global five-star safety standards as well as top safety pick requirements from the Insurance Institute for Highway Safety (IIHS)."

ahhh, I read too fast. Thank you for this.

is there still any way you'd want to get in this? there's not anything manual and it's honestly a huge statistical gamble riding around in this thing. Even with airbags, why would you risk your life? At least a steering wheel and pedals for example so you could take over if you need to. There's no way you would be safe in the middle of an intersection if someone came flying through a red light, which as a conscious being you could see from a distance when you look both ways as far as you can see. You'd be like oh snap someone is flying the car's not gonna stop! Current state of autonomous driving is not sophisticated enough to predict erratic human behavior. An example is a car crossing over your lane that's turning left from oncoming traffic without any stop signs/lights. Human drivers will sometimes cross at the very last second turn on the brakes at the last second, cut you off at the last second, and the car will probably slam on the brakes so hard it will send you from the second row to the first row or break your neck from whiplash even if you're only going 20-30mph, it's a hard stop.

> Current state of autonomous driving is not sophisticated enough to predict cars crossing lanes that's turning and crossing your lane from oncoming traffic.

This is just completely false. AV software stacks absolutely can and do predict off-nominal behavior like you describe. In general they predict better than humans, too.

you're literally talking about using a lidar and some object detection (for car and other large objects) to predict human behavior. you obviously don't realize that the human behavior is in the other car with eyeballs, mouth, head movement, body movement. that part is not being tracked. that part is human behavior. and that's where the car movements start. so everything that the "AV" is predicting is super old information because the human is full seconds ahead of doing whatever the computer is reacting to later by seeing a large object move seconds after the human brain has caluclated it's behavior. the computer is not solving the human part which you claim it does.

this is something that humans do seamlessly and you don't realize because you're probably robot. but we look eachother in the eyes, hand wave, point, etc. not always (these are the asshole drives) but you get the point. humans are superior and always will be until it's completely computer cars on the road that are networked so they can be networked the same way humans are networked to gether.

The computer sees cars that are going to run red lights, at great distances, with complete certainty, and without overlooking any of them. The example you have chosen is one in which the self-driving system is indisputably superior.

the transitions are so off an irregular that you're gonna get hit or rear ended

are you a robot or real person? i'm jc . when you're merging into a right lane from a left lane for example on a two lane road the av cannot tell if the person is gonna let you in. humans are supreior because we can communicate with eachother and you can predict what the other car will do because you can see a persons eyeballs, body lang, etc and communicate with them without words which we all do as we drive. hand waves, pointing, head nods, etc. a computer has no fucking idea. and unitl they're all networked or something it's gonna be hard.

dude literally watch this video https://www.youtube.com/watch?v=kJD5R_yQ9aw where a car is stopping in busy road trying to merge into the next lane. are you serious?

the tech is so bad still

Alas, essentially 100% of accidents happen with a human behind the wheel. Humans vastly overestimate their ability to avoid accidents when in control

If I recall correctly, collision testing is only done between vehicles of equivalent size. I do not for the life of me know why that is, but it's a pretty big problem for the safety standards.

This thinking makes US keep moving towards driving literal tanks on the road.

It would be great if we could end this arms race of bigger and bigger cars - they're less efficient, worse for pedestrians (bad sight lines and likeliness of injury if hit), take up more space and most of all - completely unnecessary in most cases!

It looks like an extremely typical minivan to me. The pillars are really thick, so they're almost certainly loaded with airbags.

And the true ’murican solution is making this car a 8-ton tank, not removing the monster trucks from our roads :(

Moving from screen to screen will become the sole purpose of our lives.

A step closer to this: https://www.youtube.com/watch?v=_xToQ4cIHkk

What is the purpose of our lives currently?

I think it's to perform some economic activity so you can 1 pay taxes on it, and 2 buy shit. Did I miss anything?

Could you explain what you're getting at? Taxis have been a part of modern cities for a long time, and they serve a purpose beyond moving humans "from screen to screen." This looks to me like an attempt to make a more modern, hopefully cheaper taxi.

I think gp is referring to the number of screens on the inside of the vehicle, and how basically everything we do now revolves around a touch screen.

Besides, I don't see how this will ever be cheaper than a 2010 prius being driven by some guy. The equipment in this machine will easily cost six figures, and the software will be a constant engineering expense, whereas the traditional taxi is just $50k for the vehicle, another $10-20k for the license (more in large cities), and then you just get a guy to drive it. Maybe economies of scale will drive the price down, but the guy driving the cab has never been the cost center.

If it works(i highly doubt we will see fsd in the next 100 years) it will run 24/7 with no cost associated with taxi drivers(i.e unreliability, health insurace etc). The tech will only get cheaper so it will innevitably be more cost efficient than a human.

Are you referring to fsd as frame shift drive from Elite Dangerous?

Probably it's "Full Self-Driving." The other kind of FSD is a bit further away, I'm guessing. :P

Geely owns, among other things, the company that manufactures the London black cabs. You know, the kind of car that drives you anywhere you want without you having to concentrate on the road.

https://www.wired.co.uk/article/levc-geely-london-electric-b...

For a very high price.

If Google is actually focused on this project, they could be terrible for Uber and Lyft in the future.

Especially if they integrate the service it with Google Maps. Use Map Ad revenue to keep prices lower than competitors.

But, obviously too soon to know if Google will keep working on Waymo.

> If Google is actually focused

That’s unfortunately a big IF in modern Google era

I think getting a license for full commercial services in California can be an inflection point. Waymo is currently serving outside SF downtown (edited: outside, not in) without too much hiccups, so I don't expect significant technical blockers for future expansion into the entire bay area and then the LA and its metro. Looks like this is their plan as well.

https://blog.waymo.com/2022/10/next-stop-for-waymo-one-los-a...

There will be several automakers who do not have world-class driving automation. Google already provides the infotainment and dashboard interfaces for Volvo with Android Automotive. VW made a mess of their in-house software development. Car makers will look to buy-in autonomy systems. Google has it end-to-end, including all the user-facing bits and the cloud backend. I expect they are committed for the long haul.

Transportation as a service? I think there is a word for that TAaS... no TAX... TAXA??

Sounds way less efficient than just building a train.

I welcome any and all efforts by our local governments to build better bus, subway, and train infra. Any day now…

THIS

Trains do not solve last mile.

Last mile I can walk/cycle.

Just get rid of traffic jams and 9 lane highways, and diesel emissions.

Trains would actually sort a great deal if the network is wide enough and frequent enough.

Not everyone can walk and cycle. I am glad you can.

> Last mile I can walk/cycle

You can, but can everyone else? I see quite a few disabled, seniors and pregnant people on trains. Not to mention people who are exhausted from their day.

Their point might be, don't let the perfect be the enemy of the good, perhaps. I was in Seoul, South Korea recently, and their public transit was simply insane. could get anywhere in the city with merely at most a 2 minute walk away from both my current location to the bus/train stop, and the end bus/train stop to my intended destination. For transit to longer areas, there were trains every 30 minutes at least to for example the coast from the center of the country.

I saw plenty of disabled and elderly and they all were able to take the buses and trains, and they used a cab for the last mile.

In contrast in the US, we don't have the transit at all, unless you live in a city like NYC, we just use cars everywhere.

> Their point might be, don't let the perfect be the enemy of the good, perhaps.

Agreed, we should be making it easier and more convenient to take mass transit instead of driving. If improving last mile transit options beyond walking/biking helps to get more people out of cars and onto trains, that's a good thing. We just need to make sure we build enough mass transit to support that use case.

I know the US has sucky trains, but commuter rail in Massachusetts is not electrified, so it is pretty dirty. I also recall a study of diesel busses running regularly scheduled routes. Because the load factor over the whole day is low, buses do not save much if any fuel or emissions.

Solving the above problems plus being on-demand and door to door is a huge benefit.

Yeah, if you don't factor in needing the government to build trains.

Americans will do anything but invest in transit. It's been shown over and over again that simple electrified rail is by far the most cost effective, green, and efficient transit solution.

The great thing about robo-taxis is that they don't actually solve a problem. If I showed somebody two images of inner city traffic jams, they wouldn't be able to tell me which is humans driving gas engine'd cars, and which is self driving electric vehicles.

> If I showed somebody two images of inner city traffic jams, they wouldn't be able to tell me which is humans driving gas engine'd cars, and which is self driving electric vehicles.

The city with self-driving cars will economically outperform the city with human-driven gas cars.

Just like America economically outperforms Europe. The best and brightest from Europe/cities with manual human drivers will leave their dying cities behind and move to the futuristic, inspiring, and highly-compensating self-driving car city/America.

Europeans are upset that America is getting away with treating the underprivileged so poorly, yet America constantly outperforms Europe $-wise. And yes, everything in life is about $.

America with good public transit would definitely outperform dollar-wise America without it.

China is investing heavily in robotaxis too, and also has active deployments in large cities (which also have transit). I'm not sure why you're trying to make this about Americans.

> The great thing about robo-taxis is that they don't actually solve a problem. If I showed somebody two images of inner city traffic jams, they wouldn't be able to tell me which is humans driving gas engine'd cars, and which is self driving electric vehicles.

They absolutely solve problems. They just don't solve the one particular problem you mentioned. For example, they solve the problem that cities don't have enough parking spaces.

This is correct, but public transit is a sort of political matter so will take a lengthy time to take off. Which is a sad portrait of modern American politics.

Over a long enough run, maybe, but the roads already exist. You don't need eminent domain, debt-financed billion-dollar public works projects, and years of disruption to existing urban transportation corridors to install rail.

On the other hand, at least making some of these driverless buses or common-destination carpool shuttles would be more efficient over any span than single-passenger coupes.

These aren't meant to compete with trains. They're meant to get you from the train station to your actual destination.

Brief: you can design ANYTHING it doesnt even need any legacy human controls. Complete freedom.

Google: how about a minivan?

> Current state of autonomous driving is not sophisticated enough to predict cars crossing lanes that's turning and crossing your lane from oncoming traffic. This is just completely false. AV software stacks absolutely can and do predict off-nominal behavior like you describe. In general they predict better than humans, too.

you're talking about using a lidar and some object detection (for car and other large objects) to predict human behavior. you obviously don't realize that the human behavior is in the other car with eyeballs, mouth, head movement, body movement. that part is 100% not being tracked. that part is 100% human behavior. and that's where 100% of car movements start. - so everything that the "AV" is predicting as what you call "human behavior" is super old information because the human in another car is full seconds ahead of doing whatever the computer is reacting to later. the computer sees objects move around and these movements are seconds after the human brain has caluclated it's behavior and already executing. the computer is not solving the human part which you claim it does.

watch this video https://www.youtube.com/watch?v=kJD5R_yQ9aw where a car is stopping in busy road trying to merge into the next lane. are you serious?

> Alas, essentially 100% of accidents happen with a human behind the wheel. Humans vastly overestimate their ability to avoid accidents when in control reply

> The computer sees cars that are going to run red lights, at great distances, with complete certainty, and without overlooking any of them. The example you have chosen is one in which the self-driving system is indisputably superior. reply

when you're merging into a right lane from a left lane for example on a two lane road the av cannot tell if the person is gonna let you in. humans are supreior because we can communicate with eachother and you can predict what the other car will do because you can see a persons eyeballs, body lang, etc and communicate with them without words which we all do as we drive. hand waves, pointing, head nods, etc. a computer has no idea. and unitl they're all networked or something it's gonna be hard.

humans communicate seamlessly and if you don't realize it, it's probably because you're a robot. but normally we look each other in the eyes, hand wave, point, etc. not always (these are the asshole drivers) but you get the point. humans are superior and always will be until computer cars are networked so they can commuicate behavior to eachother the same way humans do.

so vis-à-vis the transitions are so off an irregular that you're gonna get hit or rear ended

The socially awkward and the easily car sick will avoid these like the plague, considering the seating arrangement.

The socially awkward seem to ride buses with little fuss, I’m sure they’ll figure this one out too

Given a choice, they'd pick something else, and since there will eventually be many options it may even eventually be something you could select in the ride hailing all.

The variety of psychological issues that people present with is very broad. We are not going to design the world around them and have any sort of success

oops, I was looking at a chinese version in pictures lower down the page.

Why? It looks like bog-standard 2 seats in front, 3 in back layout with more legroom.

Are we looking at the same article? One of the pictures shows two seats facing backwards.

The only photo with backwards facing seats in the article is labeled

“Check out this wild train interior on the Chinese version. It has a centrally located table and rear-facing front seats. Truly a "living room" layout.”

Which is describing a completely different vehicle.

I missed that, thanks.

That's only for a particular variant.

It would be more awkward if the seats were facing each other.

They are facing each other.

Edit: oops, I was looking at a chinese version in pictures lower down the page.

I'm not sure I understand what you're getting at. This looks like a standard minivan layout, but without a driver. I'm pretty socially awkward myself, and I've always used taxis and such without any issue.

It seems I'm seeing a different article from everyone else. I see seats facing each other when I scroll through the pics.

Edit: oops, I was looking at a chinese version in pictures lower down the page.

Reply on news.ycombinator.com