Hacker News Reader

Comment by amckenna | Hacker News Reader

A lot of people are bashing the car's appearance, but I think people forget that putting the first driverless cars on the road is as much a PR challenge as it is a technological challenge. Truly autonomous driverless cars is a huge shift in the way we have operated for almost 100 years. There will be a lot of caution and resistance from political groups, concerned citizens, entrenched interests, etc. The car that they put forward first needs to be non-threatening, safe, and easy to adopt.

Given Google's stake in Uber the car will be part of a fleet that can be summoned by a mobile app, not some product you go out and buy. Because there will be no dealerships and individual owners, they don't care about attracting buyers for the vehicle - it doesn't need a cool factor. What it needs is to be non-threatening and safe so you will feel comfortable getting in one and going for a ride.

Additionally, the first car on the roads will just be making in town trips and will be limited to 25mph - no highways or major arterials. This means it makes more sense for the car to be compact, light, and similar to a Smart Car, than a Camry or SUB.

View on HN

> The car that they put forward first needs to be non-threatening, safe, and easy to adopt.

Moreover, I think Google wants their car to be recognizable as a driverless car. They don't want it to blend in, because people would (at least for now) think that was "spooky"—that any car anywhere around them could be being secretly driven by a robot.

Instead, the visual distinctness means that it's instantly clear that these are their own thing, in much the same way that golf carts, scooters, and backhoes are—despite all being wheeled vehicles that can be driven in car lanes—their own thing. Because they're recognizable, it makes it possible for you to build a separate autonomous response into your subliminal "driving event loop" for the class of "driverless car", the same way that you will react separately to seeing any of the above in your lane. Which is exactly what people want, in order to feel in control: the ability to notice and compensate for the driverless cars around them if they all encounter some pathological edge-case.

(Just as an example of what I'm talking about—and this is a completely untested hypothesis on my part—I bet that driverless cars come to a complete stop a lot more frequently, in surprising places, because their sensors tell them something is standing in front of the car—e.g. a pigeon. A regular driver knows a pigeon will just sit there minding its own business as the car approaches it, but fly away right before the impact would happen; a driverless car might not know that. Because of this, your driving algorithm for being behind a driverless car will likely have a component similar to driving behind a school bus: "this vehicle makes frequent sudden stops, and I need to be ready to stop too. I should be more car-lengths away from it than are strictly necessary.")

oh I sure hope it takes more than a pigeon, but I see your point

I hope it doesn't; consider the PR disaster that'll happen the first time a driverless car hits poor little Fluffy the cat. Nevermind that cars with drivers do that all the time.

Point of fact: Fluffy the Cat was only hit nine times. All the time indeed...

Not pigeons right in front. But the driverless car often sees dangerous spots (like debris on the road, or an accident site) before any human can see. So they might stop or slow down earlier.

(I think I learned that in one of the tech talk videos about it.)

Plastic bags actually. A plastic bag looks pretty solid to most sensors and cars have been known stop suddenly when one flies in front of the car.

And they should stop, and err on the side of caution. An anecdotal digression, if you'll allow me: I'll never forget the time when a plastic bag was gently flitting across the highway a friend and I was traveling on, in that nonchalant manner plastic bags usually do, and how we, not giving it even a second thought, drove right through it... Except we couldn't, since it turned out there was actually a small picnic-type propane tank inside the bag, rolling on the asphalt, and our car's motor was totaled by the impact, though thankfully we didn't suffer any serious injuries.

This is why you never run over anything in a car. Ever.

No, they shouldn't. I'm glad you weren't hurt, but your freak accident aside, most plastic bags are just plastic bags and human drivers are not expecting vehicles to yield to them. (There is also a difference between a plastic bag moving on the ground and one flying around.) If we are erring on the side of caution we shouldn't be creating a more probable hazard to avoid a less probable one.

And if we remove humans entirely from the equation what we get is traffic that comes to a stop because none of the vehicles can avoid or get past the plastic bag which floats around more-or-less randomly.

That sounds plausible to me. Apparently, even things like exhaust from another car in cold weather can interact with the laser in a way such that the car can mistake it for a large object directly in front of the vehicle.

Source: tech talk at my university by a self-driving car engineer.

Well I guess we wont get these in Europe (due to rain) but there are lots of cars that look like that here already.

The appearance of the car is really important. If it looked too fun, like a Ferrari, some may think it's not safe. It reminds me of "helper" robots in movies about the future. They always have such unintimidating appearances. This car appears to be attempting the same thing and I'm ok with it.

sure appearance can convince a person of a well designed product. But I prefer, a well designed product, especially when my life is on the line over appearence. Tesla may come out with a super shiny beautiful self-driving car, but I will probably go with a dorky Google car if it has X miles (where X is a large impressive number) under its belt.

In fact, now that you mention it, this is pretty much Baymax-the-Car. It just needs to puff up when it accidentally drives into a lake.

I have to be honest - the last thing I think of when I look at it is "safe." Maybe that's just because I'm an American and feel the need to put as much iron between me and everyone else as I can.

It's cute, and non-threatening looking, and probably the right design for what it is, but people are still going to look at it and think "this is going to get destroyed if a real car hits it."

There are (at least) two kinds of people Google needs to convince. First, people who will sit in a self-driving car; and that's already done. Second, people who share the road with a self-driving car. You are right `unthreatening' is the better word than the ambiguous `safe' here.

PS I think you are too American there. They don't drive tanks on the road in every country. But I see your concern.

Fair enough.

This is also a prime opportunity to question the idea of traveling at 85 mph in 4,000 lb. metal boxes that get 13 mpg.

In my town, many parents put their newly-licensed teenage children behind the wheels of enormous SUVs in the name of safety, simultaneously reducing the safety of everyone around them. If tanks were street legal, we'd probably have 16 year old kids barreling down the road in them.

Actually theres been a lot of research that increasing speed limits have more benefits than not. Most accidents that happen on high ways resulting in fatalities are due to driving under the influence. A higher speed limit has been shown to decrease accidents, decrease ticket fines, and decrease fuel consumption.

Making it illegal to buy SUVs is moronic even if buying them is moronic in itself. Lots of things in life are dangerous, and we're going too damn far with needing to make anything with a potential of danger illegal. I'd much rather own a Range Rover which is far more comfortable, enjoyable to drive, accommodating, while reserving the capacity to take me to Yosemite a few times a year.

Aren't most tanks driven by teenagers already? Of course, they get very specialized training, but it still shows that teens can handle that responsibility under the right circumstances.

Actually, I don't really know: how old are the soldiers in a typical tank crew? I guess they might have a couple of years in the service before they get into a tank, so maybe they're early 20's rather than teens.

Real tanks mostly do not drive threw dense citys, so its mute.

> Real tanks mostly do not drive threw dense citys, so its mute

Tanks are not usually quiet, whether driving or throwing.

> the first car on the roads will just be making in town trips and will be limited to 25mph

I think the rest of us are going to start loathing these timidly-driving things: "Oh f%#k! I'm stuck behind a self-driving car again, now I'll be late for my meeting."

Or not: http://www.wsj.com/articles/traffic-engineers-say-slowing-do...

Uber have the right idea here; own the platform, not the infrastructure. Lets face it, a cars spends most of its life stationary, waiting for you to drive it. Not so the autonomous vehicle. Autonomous vehicles are freed from the shackles of parking so why not lease your car to Uber and have it earn money as a cab? The future will be a bad time to get into car parking.

> The future will be a bad time to get into car parking.

Perhaps, this will be the case in the long term. However, until then perhaps we could implement some sort of price surging where rental costs of parking space skyrockets when demands get high. Uber is doing the hard work of making price surging palatable. Everyone else just needs to ride on their coat tails.

SF and other cities already have variable, usage based parking fees[0]; I imagine these will only become more widespread.

[0] http://sfpark.org/

That is already common in most places. Most parking structures I know have different rates for different arrival and stay lengths. A lot of street parking has free and paid times. Demand based parking has been around a long time.

but the jutification for surge pricing doesn't make sense with parking. The entire premise (that I have not seen data to verify, but the premise nonetheless) is that surge pricing increases the number of drivers on the road. Are we going to build more parking spaces for half a day?

"Price surging" would become a fancy word for price gouging.

Surge pricing also limits demand. And that's a good thing; it's better if there's an expensive spot available, than no cheap spots.

only better for rich people. Rent seeking does not follow free market principles (due to fixed supply), and so loses any form of moral highground in my opinion.

Surge pricing is not rent-seeking. If the surge pricing was not warranted by the situation, then it creates a trivial way for competitors to eat the business's lunch by offering the same service at a slightly lower price.

"Uber is surge pricing right now but Lyft is not, so I'll get my ride with Lyft"

Surge pricing is a legitimate function of a properly functioning market. When demand has skyrocketed, or supply is constrained, prices _should_ rise. This is the mechanism by which markets enable resources to be allocated efficiently. Whether someone is poor or rich, if they desire a thing more than other people do (for the sake of argument, other people of similar wealth), then they should have a mechanism by which to express that preference. That preference is the price they're willing to pay.

If the area is congested, and you _really_ want to go somewhere right now, then surge pricing gives you two ways to accomplish this. (1) To the extent that capacity is fixed, the ability to out-bid someone else means that you can take the capacity from someone else who doesn't desire it as much as you do. (2) The higher price incentivizes more supply. Uber drivers who see an exceptional surge price will make themselves available in order to take advantage of higher wages. A higher surge price is pure profit for the driver, if he can make himself available to drive right now versus some other time in the week

Lastly, no supply is fully constrained. If the price of parking in a city skyrockets, then this creates a strong incentive for parking businesses to find a way to make more parking available. This might mean buying existing buildings, demolishing them, and converting them into parking structures; or it might mean that in the routine course of building new structures, the high price of parking will make it economical to dig a few stories deeper and provide a few more levels' worth of parking. Price and incentives always influence the operation of the market.

Thanks for the explanation!

Surge pricing in Uber's case is not rent-seeking if surge-pricing actually increases supply (i.e. more drivers on the road). My gut says it doesn't ( don't think that drivers get out on the road fast enough), but I don't have the data to prove that.

But in the case of parking spaces, I don't see how me being able to out-bid you for the last space makes me a more qualified candidate for that parking space. I'm paying a higher price for the same good, the number of spaces isn't going up, and the owner got richer without adding any value. Textbook definition of rent-seeking if I've ever seen one.

High parking tension means that people will try to make more parking, but is that the most effecient allocation of resources? What about alternative solutions, like mass transit? This is a digression, but you mentioned it with supply.

On a more pragmatic level I don't really think the ultra-rich have that much more of a right to a parking space downtown on a Saturday any more than I do, and a parking space is more or less a public good (and should be shared as such).

> Surge pricing in Uber's case is not rent-seeking if surge-pricing actually increases supply (i.e. more drivers on the road). My gut says it doesn't ( don't think that drivers get out on the road fast enough), but I don't have the data to prove that.

From what I understand, fear and hatred drive public policy more than objective data does. I doubt it is going to change any time soon.

You won't lease or own self driving cars. That makes no sense. You will just subscribe to a drive service if you drive a lot or just pay per ride and have one pick you up.

I easily see you scheduling regular events (leave for work at 8, for example) and your car will be out front waiting each day.

No way. In the city, yes. But in the somewhat wealthier suburbs many people will spring for the "self-driving" package. It might only be a ~$10k add on or less by my estimates.

Most people will want to be able get in the car without having to request and wait for a ride (most people already have this luxury).

Sure, at first. But then your daughter has the SUV and your husband has the sedan and you want to get groceries and your sons want to go to two different friends houses, so you end up using sepf-driving cars all the time anyway, and you realize it's actually quite nice not to have to worry about maintenance and cleaning and charging. And you realize you could save $300 a month. And oh, you can just order a 4-wheel-drive vehicle pre-loaded with kayaks for your camping trip? And... and... and.

We haven't imagine half of what cars-as-a-service will entail. Some people will always hold onto their vehicles, but in the end I don't think a hunk of plastic and metal will be able to compete with a fleet of special-purpose vehicles supported by a staff of thousands.

Many rich families own (at least) one car per family member. We are talking about luxurious add-ons which is why I bring this up.

Isn't owning typically cheaper than renting?

Anyways, it makes no sense that everyone on my block owns their own lawnmower, and yet here we are.

> Isn't owning typically cheaper than renting?

Depends how you account for all factors. Owning usually comes with some hassle and risk. Renting means you pay someone else to discharge those duties. Another factor is the cost of capital for you vs the company that would rent to you.

Having your own lawnmower probably has something to do with transaction costs. Ie it's a hassle to coordinate among people.

Renting a property or object that you are the sole user of is typically more expensive than owning.

Renting a property or object that you are one of many users of is typically less expensive than owning. For instance, you don't imagine that renting one of those steam-carpet cleaners from the grocery store once every year or two is more expensive than owning one, right?

(Renting is more expensive than purely your share of the use -- that is, if the steam cleaner gets rented by 100 people in a year, you expect to pay more than 1% of the yearly total cost of ownership of the cleaner).

So the concept is that say you effectively use maybe 1/5th to 1/3rd of a car's capacity in a ride-for-hire situation, and so you pay say 25%-40% of the total cost of ownership, allowing a reasonable profit margin for the owner. (Or you use much less, but in that case you probably were never going to own a car).

This doesn't seem quite right - if I own a car and use only 1/3rd of its capacity, I generally expect it to last 3x as long, so the unused capacity costs me nothing. Even your annual steam cleaner example will favor the purchaser over the renter, given enough time.

So the difference between renting and owning must come down to the cost of unused capacity, which means the fixed costs. Where those fixed costs are high, like a parking space in SF, you want to amortize them by maximizing utilization, which means sharing cars. But where those fixed costs are low, like parking in my driveway, unused capacity may be so cheap that owning costs less than renting, even if the car sits idle most of the time.

Cars degrade in value pretty strongly based on age, not just mileage. For example, a 2009 Toyota Camry with 5,000 miles on it, base 4-Door sedan, in excellent condition, shows a Kelley Blue Book price of about $13k (in my zip). It would've been about $18k new.

But your point is taken, a great deal of the equity burn on a car is mileage-related.

(Especially when you consider that if there are three users of a rented driverless car, it drives more than 3x the distance than would three separate owned vehicles for the users -- it must, after all, shuttle itself between users.)

In the short term, self-driving is going to be an optional activity, not an always-on activity. So for most drivers who have it, it will simply be a luxury. Certainly there will be services, but it may or may not be cost effective. We already have the option to schedule a car service every day at 8 am, yet most of us drive. Some of us have the option of busses and trains, which is more or less the same thing, but again, we opt to drive.

In the long term, your car isn't going to sit outside your office for 8 hours. It's going to pick up the kids, your take-out order from the Chinese place across town, and deliver your laundry to a laundry service. People aren't going to travel around in taxis that don't even have the benefit of a driver who tries to keep it clean. They're going to want their own space.

There are some caveats and exceptions, certainly. And self-driving vehicles and services may prompt more people to eschew car ownership, but the idea that people won't own or lease these things _in general_ is ridiculous.

But everyone wants to drive to work at the same time so there is very little saving from central provision (and thats the main use, at present in the US). It makes the rental model much less attractive.

For commuters, may be that's all there is to it. If I want to go out on a trip with my little one, I'd rather have a bunch of her favorite things in the car (without me having to lug it around every time), and have the convenience of not having to drive, which also makes it much easier to keep her entertained. I'm hoping she'd grow up by the time these are generally available, but I'm sure there are others reasons to own one.

Assuming Google wants to get into the car business, all it has to do is add a button to google Maps/now. I don't think ther's anything special about the Uber platform, and Google already integrates journey planning through the Calendar and Google Now interfaces (eg telling you that you need to leave in 5 min to catch that train).

You're right on the car parking. I think the City of Chicago actually made a very shrewd decision to sell their street parking revenue to a private operator for a 75 year period in return for a fixed amount of cash up front (obviously I'm oversimplifying). It was excoriated as a terrible deal by most people, but none of them seem to have considered the probability that there just won't be the same sort of demand for parking in even 10 or 20 years.

Do they really have driverless cars or just cars that run on essentially electronic tracks very carefully determined?

If someone moves the location of a traffic light, can a Google car still find the traffic light? Wbat if a bulb in the traffic light is burned out; is the car smart enough to negotiate the intersection with the failed light? What if a bucket of white paint fell on the road yesterday; is the Google car's image processing good enough to recognize this or to tell the difference between wet paint and dry?

Heck, when I drive, there is a lot for me to watch and analyze. A Google car can do that?

A Google car can do that?

Yes.

https://www.youtube.com/watch?v=TsaES--OTzM

They have videos of their cars handling common cases. But what about the corner cases?

I've had a sign disappear on me over night which changed the roads rules at a tricky intersection.

There has been a lot of construction in my area so 2 lane roads had a lane closed off and a traffic controller managed traffic across a single lane.

I've even seen a road that you could drive into and out of, become a road that you can only drive into. There had been traffice controllers during road works managing a single lane all through construction and then the day that it was finished, one direction became a dead end.

Did you even watch the video?

The google cars have no memory of what was "yesterday".

or if someone parks a flatbed with a ramp in the road with lines painted up. Will the Google car just drive onto the back of the truck? Would this be how you kidnap the cars and occupants?

I don't care how it looks, it if still cannot drive in bad weather, snow, and worse needs detailed up to date maps, its not truly autonomous driving, its like programmed driving with a few extras.

When the cars can come to you where ever you may be, broke down on the side of the road for example, and take you anywhere regardless of conditions then it can be heralded.

As it stands now its nothing more than the automated cars shown as high tech in Jurassic Park.

This is silly. I bet you can't drive through a swamp in a jeep, because you haven't acquired the skill. But that doesn't mean the rest of your driving doesn't qualify as real driving.

Perhaps because what you describe is not autonomous driving, but human driving. And it's been through a 100 year trial and found to kill over a million people worldwide each year, prevent the elderly and disabled from living full independent lives, and be a huge time and money sink that many humans have to accept.

Snow, heavy rain, being able to go everywhere with up-to-date maps -- these are problems which will find solutions in time. I have no doubt that up-to-date maps will be no problem, considering these vehicles constantly collect centimeter-accurate 3D maps of where they drive and upload them to Google.

And being able to pull over to the side of the road to pick someone up? A simple software feature. There's nothing preventing that NOW.

Don't get caught up in thinking that these things have to be perfect upon release to have a huge benefit.

Everyone seems to be overlooking the fact that using an autonomous car as a cab would be illegal. Current permits are rather specific about having an manufacturer's employee behind the wheel at all times. I'm not an expert, but I wouldn't expect any state to lighten up the requirement of at least having someone behind a wheel for the foreseeable future. I think most patrons of Uber don't really want to be sitting behind the wheel.

"I think most patrons of Uber don't really want to be sitting behind the wheel." - if the vehicle was self driving, then absolutely.

I can imagine a scenario of 24/7 tech support VC'ing at the push of a help button or if Google notices the car has malfunctioned.

Getting the laws changed will take time. When presented as a n option will it be in demand? You bet!

This generation of cars doesn't have a "wheel" at all -- there is literally no steering wheel and no way for a human to operate it.

So legislation will clearly need to change before these are street legal.

"It doesn't need a cool factor"

That was the death knell of Google Glass (at least in its current incarnation). Seems that Google hasn't learned much.

Google Glass is a camera on your face. The self driving car revolutionizes transportation internationally forever.

One makes you look like a hipster. The other is primed to save tens of thousands of lives a year if it gets rapidly adopted.

One weighs grams. The other can run you over and crush your skull.

Perception matters. This initial design is so insanely non-threatening it achieves its appearance goals. It does not need to sell units, it needs to sell the idea.

AFAIK, Google Glass was explicitly a prototype, meant to visually advertise itself as a prototype to anyone who looked at it—in the same way that, say, a breadboard or a caseless Arduino visually advertises itself as a prototype.

Seeing someone wearing Google Glasses was meant to put you in mind that they were testing something and you could help them test it, not that they were using a piece of everyday consumer electronics.

And this car is pretty much the same: it's meant to stand out (see my other comment above: https://news.ycombinator.com/item?id=8785463). So I wouldn't say they needed to learn anything.

As the technologies are accepted and commercialized by other parties, then they can make one that looks normal-ish.

Google has tried hard to make Google Glass fashionable in its current incarnation. That's why they are inserting it into New York Fashion Week, getting designer frames, etc.

That's back when they tried to be like Apple.

/pedant mode

Death knell, not death nail. A 'knell' is the slow solemn tolling of a bell at a church funeral. This has been a public service announcement.

Reply on news.ycombinator.com