Hippie Hedgehog posted:

How, exactly, does this differ from regular car prices, today?

Millions of people in the US are poor, and will never be able to afford a new car, even the cheapo ones like the Chevrolet Spark ($13,875) or Nissan Versa S ($12,855). So they ride a bus to work, or carpool, or buy used cars.

I guess I'm not sure where you're going with your argument that autonomous cars will cost a little more than regular cars.

Peanut President posted:

The problem with autonomous cars is all the major problems they fix can be fixed with tech that already exists called a choo choo train.

Peanut President posted:

The problem with autonomous cars is all the major problems they fix can be fixed with tech that already exists called a choo choo train.

wolrah posted:

Autonomous cars work on the infrastructure we've been building our society around for 60 years.

wolrah posted:

There are exactly zero passenger rail lines within 20 miles of my house. The closest station is at the airport and pretty much only goes downtown.

As many problems as there are still unsolved in autonomous cars, I think solving those will still be easier than convincing Americans that investing in rail infrastructure to a level that would actually make it useful for most of us is worth doing.

Even in the denser areas that can more easily support mass transit systems that pre-existing density poses the problem of where to put the new rails, since Amtrak shows us clearly how well any freight priority system works out.

fishmech posted:

Autonomous cars will cost a shitload more than regular cars because you need all the extra sensors, processing hardware, possibly also network subscriptions. Additionally in order to actually function those sensors must continue to be cleaned/maintained and it'll definitely not work correctly if components outright break/get stolen.

I'm not sure what's so hard to understand about this for some people. It's not just a little extra expense.

Javid posted:

A car that absolutely cannot be driven by a human is also worthless if one tiny electrical part takes a poo poo, as you can't even get it to a shop to be repaired. I doubt any of them will come without even some rudimentary manual controls, even if it's some lovely tiny steering wheel and pedals that fold out of the dash when needed.

Hopefully somebody comes up with an effective retrofit process for older cars.

Javid posted:

A car that absolutely cannot be driven by a human is also worthless if one tiny electrical part takes a poo poo, as you can't even get it to a shop to be repaired.

fishmech posted:

Autonomous cars will cost a shitload more than regular cars because you need all the extra sensors, processing hardware, possibly also network subscriptions. Additionally in order to actually function those sensors must continue to be cleaned/maintained and it'll definitely not work correctly if components outright break/get stolen.

I'm not sure what's so hard to understand about this for some people. It's not just a little extra expense.

Hippie Hedgehog posted:

I'm convinced that, in a few decades, the car with manual controls will be more expensive than the autonomous one, simply because the autonomous one will be produced in higher volumes.

EoRaptor posted:

I'm going to be pedantic, and say it won't be the cost to produce the car, it will be the cost to insure it. Self driving cars will consolidate the insurance industry, and change the standard for accident investigation, because a self driving car can be interrogated with 100% accuracy. A human will have to pay to offset that kind of advantage. Few will have the means to do so and fewer will probably want to.

Peanut President posted:

The problem with autonomous cars is all the major problems they fix can be fixed with tech that already exists called a choo choo train.

babyeatingpsychopath posted:

Javid posted:

A failed drive belt can be diagnosed with the unaided human eyeball, costs $10 at any auto parts store in the universe, and can be replaced by one person with - at most - a wrench, anywhere. Embedded Chip #8043x0f8c deep in the dash getting six molecules of water on it and frying so the autonomous driving system just fails to boot with an "Oopsie! :-(" error is just game over, enjoy getting assfucked by the dealer for thousands to order and replace a piece of silicon the size of a dime.

wolrah posted:

One major problem they solve, at least in the US, which stops trains in their tracks...

Lack of tracks.


Autonomous cars work on the infrastructure we've been building our society around for 60 years. There are exactly zero passenger rail lines within 20 miles of my house. The closest station is at the airport and pretty much only goes downtown.

As many problems as there are still unsolved in autonomous cars, I think solving those will still be easier than convincing Americans that investing in rail infrastructure to a level that would actually make it useful for most of us is worth doing.

Even in the denser areas that can more easily support mass transit systems that pre-existing density poses the problem of where to put the new rails, since Amtrak shows us clearly how well any freight priority system works out.

Javid posted:

A failed drive belt can be diagnosed with the unaided human eyeball, costs $10 at any auto parts store in the universe, and can be replaced by one person with - at most - a wrench, anywhere. Embedded Chip #8043x0f8c deep in the dash getting six molecules of water on it and frying so the autonomous driving system just fails to boot with an "Oopsie! :-(" error is just game over, enjoy getting assfucked by the dealer for thousands to order and replace a piece of silicon the size of a dime.

Kaal posted:

If your privately-owned Uber auto-taxi joins a Google car-convoy on the publicly closed federal highway, and the fifth car in line blows a tire leading to a fifteen car pileup, who's liable for your health coverage?

Javid posted:

A failed drive belt can be diagnosed with the unaided human eyeball, costs $10 at any auto parts store in the universe, and can be replaced by one person with - at most - a wrench, anywhere.

Hippie Hedgehog posted:

I'm sorry but I'm not sure what you're getting at. Maybe you could expand on that a bit?
By consolidate, I take it you mean companies will merge and jack up prices due to lack of competition, but how do you predict self driving cars will contribute to that?

Surely the cost of investigating an accident will go down when each involved car can be interrogated with 100% accuracy?
Surely the insurance premiums will go down, not up, when the number of accidents go down due to self driving cars having fewer accidents?

Yes, a free market always has a tendency towards monopolies, but we have at least some semblance societal safeguards to prevent that.

EoRaptor posted:

I think you’d be surprised how much cost can be cut from a car that doesn’t need to support, enable, and obey a human pilot.

No need for 360 degree glass
No steering wheel or instrument cluster
No forward facing front seats, so fewer airbags

Electric helps further, as maintenance needs on an electric car are significantly lower than on an ICE car. The car can also drive itself to a depot for periodic maintenance, cutting down on those costs further ( breakdown towing will still be a thing, of course )

For the very low end service, an electric car with center facing hard plastic seats, a scattering of small porthole style windows, vinyl floors, and no console at all ( everything is controlled via phone) is probably going to appear at some point, and target the working poor who would normally by used cars and be cost effective for them vs gas, maintenance, and insurance costs they have each month on a beater used car.
....

Hippie Hedgehog posted:

Kind of like when the drive belt breaks on a regular car, and you have to call a tow truck?

The problem, that cars have multiple single points of failure, already exists. This is nothing unique to autonomous cars. Of course, the fewer components a machine has, the less the risk that one fails, but in fact electronic part are inherently less prone to failure than mechanical ones. Also, as pointed out, electric cars have significantly fewer parts and the future is electric.

If we can make combustion engine cars that run without failure for years at a time, there is no reason why we couldn't do the same with an electric one that's autonomous. It's a matter of applying two things:
* good engineering principles
* experience

The latter will take some time, they won't work perfectly.

Hippie Hedgehog posted:

Do you know what a cell phone cost in 1998?
What did a similar one cost in 2008?
What does one cost now?

I'm sure you see where I'm going with this. Systems that today seem very complex and are expensive to produce, become less expensive over time as volume increases. A very basic principle of manufacturing and engineering. Making a prediction that whatever seems very difficult and expensive today, will stay expensive forever, seems disingenuous.

I'm convinced that, in a few decades, the car with manual controls will be more expensive than the autonomous one, simply because the autonomous one will be produced in higher volumes.

Javid posted:

Is there a safer direction for a vehicle's occupants to be facing than directly forward? That actually makes SOME sense as a possible avenue of safety improvement that can occur if the driver doesn't even have to be pointed in the direction the car is going.

Offhand (and I don't claim to actually KNOW, just spitballing as a layperson), if everyone's facing backwards, it seems like head-ons could be less horrific, since you just do high Gs into the (presumably redesigned for this purpose) padded seat rather than smash forward into the steering wheel/airbag shrapnel. Getting rear-ended might be worse, but that's more likely to involve lower relative speeds? And the rear vehicle still benefits.

That said, none of that equals CHEAPER, just BETTER.

fishmech posted:

You can't use "electric cars are easier to maintain" as your basis for self driving cars being cheaper to do the same.. Those are entirely unrelated, you can make a self driving car on gas and you can make a non self driving electric. Dump that argument from your defenses of self-driving cars entirely.

A self driving car needs a massive amount of special sensors all around, to match a human's two eyes and to a limited sense ability to intuit speed and traction issues etc from touch. If those parts start breaking, the vehicle becomes "blind" in an important direction or subtype of detection immediately when it needs all of those things functioning to work. It's not like a modern car with that legally required backup camera, where 99% of the time it's unneeded and it's low res anyway. The car needs the sensors kept in good working order and cleaned so that the car can self drive at all.

And that will all cost more than not needing all those bits.

fishmech posted:

What are you actually trying to ask here? Because my mom bought herself a new phone in 1998 and it was $500 were it to be bought outright, instead she paid $50 upfront and the rest went on monthly billing. In 2008 a similar class of phone depending on feature set is either "completely free on contract with no extra payments, but a similar class of phone depending on who's buying it? That would be reasonably high end featurephone with slide out QWERTY and a tiny little touchscreen for lining up camera shots for... $500. And what does one cost now, well, I spent $950 on my Pixel 2 XL to just buy it outright, but other people can get basic smartphones for $20 and something that meets the general requirement of fitting into the cell phone market bracket that my mom's previous phones did is something like the :G she has now, which is a full smartphone for .... $600. So really prices didn't change much?

I don't know where you thought you were going with phones. We simply mashed together the concepts of the existing decade old PDA and the existing decade old cell phone and made the first smartphones, it didn't require levels of AI that doesn't exist yet to do the way car driving does.

And where you're going is wrong. The car will always need a ton of sensors which must be working as wella s a ton of extra computer equipment and probably a whole separate network communication network which will probably need subscriptions to pay. This will always cost significantly more than not having the sensors et al - that is to say a normal car.

Hippie Hedgehog posted:

I never said that "self driving cars will be cheaper to maintain because electric cars are easier to maintain". Please stop making strawman arguments.

The post I replied to said "A car that absolutely cannot be driven by a human is also worthless if one tiny electrical part takes a poo poo, as you can't even get it to a shop to be repaired." Which I thought was a poor argument against making self driving cars, because even though there is an inverse relationship between them, complexity does not preclude reliability.

So I posted to say that manually controlled, combustion engine cars, are already hugely complex machines, which the world's collective of automotive engineers have made go from things that break down all the time, to things that can run for years and years if you give the mandated maintenance.
Adding the switch to electric cars, as is already starting to happen, will contribute to an increase in reliability, which is why I mixed that into my post.

Hippie Hedgehog posted:

Dude, that's a really long spergy post that I guess you only had to post because I didn't clarify that I meant a cell phone with equivalent features. Sorry for that omission. I had a cell phone in 1999 that could make phone calls and send SMS, and store phone numbers in an address book, and play Snake. in 2010 I bought a similarly featured "dumb phone" for a 20th of the price, and it was a third the size, to boot. Of course cell phones have not only gotten cheaper, but you somehow missed my point that 30$ will buy you a pretty good phone these days, if you're just looking to call someone, even though you're obviously aware such phones exist.

You keep obsessing over details and (intentionally?) missing the bigger picture: I'm saying that a certain technology (say, lidar detectors, or collision avoidance software) will only get cheaper over time. Sooner or later autonomous cars will not be as expensive as they seem today.

I'm talking about long-term effects of the tech, over say 10-20 years or longer, but you keep on arguing like I said they will be cheap next week. As usual, you're hopeless to argue with because you seem to never take any argument at good faith. Instead you seem to assume everyone needs to be proven wrong. I wish you'd find a better use of your time (and ours).

(BTW, AI is in no way needed to make self driving cars. I assure you, no machine learning or AI will be involved in the software of the first generation of self driving cars, just as it is not today a part of Tesla "auto-pilot" or similar existing driver assist tech.)

wolrah posted:

There are exactly zero passenger rail lines within 20 miles of my house.

Peanut President posted:

And whose fault is that?

devicenull posted:

So back to roads...

How cold is too cold to pave? I'm wondering if we'll see completion of our road, or if we're stuck with raised manholes until spring. I'm sure the plow drivers will love that.

devicenull posted:

So back to roads...

How cold is too cold to pave? I'm wondering if we'll see completion of our road, or if we're stuck with raised manholes until spring. I'm sure the plow drivers will love that.

wolrah posted:

Note the next sentence. The rail that does exist basically connects the airport with downtown and then has 1.5 east-west routes (two lines that share track for about half the run and then one splits to run a few blocks south of the other one). It's the Browns of rail systems, technically there but pretty much useless.

I'd have to uproot my entire life and move out of state, get a new job, meet all new friends, etc. just to live near a useful rail system.

I like the idea of being able to take the train to get places, but not that much.

fishmech posted:

These proposed cost savings make no sense.

"360 degree glass" is cheap to have, doesn't weigh much, and people like to be able to look around themselves. Plus there's already plenty of vehicles out there with glass-delete on some models such as cargo variants of sedans and vans.
Removing a steering wheel entirely means your car now needs to meet a much higher standard of reliability if it's impossible to ever take emergency control, and instrument clusters are literally just cheap LCDs on many cars. "Removing the instrument cluster" on a lot of vehicles would mean a savings to the manufacturer of like a few dozen bucks tops.
There is absolutely no reason that "no forward facing front seats" means fewer airbags. That doesn't even make sense as a feature for such a vehicle to have either.

Electricity doesn't "help further" as electric cars already exist for non-self-driving cars. That's 0 cost saving for the self-driving car.

Using "small porthole style windows" is likely to cost more than simply dumping cheap glass into the large existing frames. There are already cars with lovely uncomfortable seats around, and this won't really save money. Also the working poor won't be able to afford such cars are you literally insane?

Or are you just dropping a huge assumption in the background that "oh by the way someone's spending billions of dollars a year to subsidize this service" because that's what it would take to get a taxi service down to "the working poor can afford this on a daily basis" levels.

EoRaptor posted:

I'm not sure if I want to reply here, because you don't seem to be interested in a debate. Your reply was duplicitous, deceptive and chose to pick specific meanings for other peoples words that support your use case, instead of more generally understood or common use meanings. I can see why you get such a bad rap in debate threads, but here we go anyway.

Glass will always cost more in a structure than the equivalent strength of steel. Glass provides no structure of its own, and must be supported by the surrounding structure. Removing the glass from a car and replacing it with equivalent strength steel makes the car cheaper to produce. You deny this, then later talk about vehicles that have had glass replaced with steel to reduce their costs. Small porthole style windows (see the style used by the virgin galactic unity vehicle) are cheap, universal and lightweight, and would meet the needs of providing a visual sense of a motion for a vehicle while keeping manufacturing costs to a minimum.

There is no extra standard of reliability needed when removing the steering wheel. The steering wheel already needs to work reliably in the event of several failures, and that same system can be used to turn the wheels, even if the steering wheel and steering column are removed.

My reference to instrument clusters includes not only the speedomoter, etc, but also the entire center console, radio, everything. I'm not sure what you imagine the price of these are, but please don't think doing away with them won't save a bunch of money in the cost of a vehicle.

With no forward facing seats, the driver and passenger airbag are no longer needed because a collision during motion won't have the same risks to the people inside the vehicle. Their removal is a significant cost savings.

The discussion up to this point had been about how people wouldn't switch to a for hire car service, with either per use fees, monthly fees, or some combination of the two. My point was that a for hire car service would appear that is cheap enough to make even people who buy very cheap used cars and have the minimum insurance on them would eventually switch because they would realize a cost savings, and that to get pricing down that low, the for hire service would populate it's fleet of cars with some very minimum, no frills vehicles, which would be designed and built for this purpose, not retrofits of any existing vehicle. An electric vehicle is considered part of this cost savings, because the people I was talking about are currently only buy ICE engined vehicles, so the savings in gas cost are included in them being incentivized to switch.

You also bring up taxi services. While Uber is out to break the back of taxi unions by exploiting legal loopholes (or just acting illegally) and pushing systemic risk onto the shoulders of those least able to bear it (their drivers) to reduce costs, that wasn't ever brought up by me, so I don't know why you suddenly referenced it. Self driving cars, either owned or more likely as subscription based fleet vehicle, will be the end of human driven taxi services, put this has nothing to do with what Uber is doing (though Uber understands this, and is also pushing hard into self driving cars).

devicenull posted:

So back to roads...

How cold is too cold to pave? I'm wondering if we'll see completion of our road, or if we're stuck with raised manholes until spring. I'm sure the plow drivers will love that.

fishmech posted:

But you don't need to have "the strength of steel" where the windows are - that's what the "pillars" and other aspects of the car's frame are for. And replacing all that stuff with steel doesn't even make sense, it's not saving money. And no, cargo vehicles don't use blanking material because it's cheaper, they use it because you don't need the windows and the cargos might break or otherwise damage actual windows in many situations for what you carry.

There is an extra standard of reliability needed if there's no steering wheel. Because a car you can still drive ... can still be driven if the self-driving part craps out. It's kinda weird that you don't get that?

But the center console, the radio, those are all cheap components? And you still would want to have the ability to play music etc inside a self-driving car, even if it's just to pair your phone with. You're trying to economize by dropping out a few dozen bucks, it's insane.

This isn't true, you still need airbags in the car for an accident. There's all sorts of collisions that can happen that need an airbag. So no, you can't get a cost savings there unless you're willing to massively endanger the passengers.

These bare minimum no frills vehicles are still going to be very expensive, and thus costly to rent access to, since they need to be self-driving, which requires very expensive additional equipment. Your apparent response to this is to try to save maybe $100 a car by making it impossible to listen to music and being hard to look outside while making the car a death trap in case of accidents because you think you can just remove air bags and other safety features.

No, you're not getting this, I wasn't talking about Uber. I was talking about the fact that most people can't afford your ludicrous self-driving car rental system unless someone's going to be pouring billions of dollars into it, whether it be a government or private money. Because cars are actually quite expensive things, despite your asinine plans of making them into soviet drudge-mobiles to scrimp a few dollars off the price after dropping tens of thousands of dollars of expensive sensors and processing hardware in. Hence why costs to ride and use current similar services are high, too high for most people in most situations to use them on a daily basis.

EoRaptor posted:

Why would I put pillars in to support giant holes that I'm filling with glass if I don't need to? They just aren't needed any more, so I'll take that same weight of steel and spread it around the 'upper half' of a car. I get the same strength of 'car' to meet safety standards, and it's cheaper in materials, cheaper in manufacture, cheaper in design, and cheaper to maintain. Again, I'm referencing how the long tail of self driving for hire cars will play out, not the projected first offering (that doesn't yet exist and may not for years), which will be premium products.

Steering components need to be reliable and redundant, and they already are. Taking away the steering wheel itself doesn't change that. You are saying that the computer, sensors, and actuators also need to be reliable, which is 100% true, and will cost money. Those will start out pretty expensive, but eventually they will come down in price to not only be less than the added costs of a vehicle that needs to support a human driver, but a lot less. There will be no need to 'take over' driving a self driving car, any situation where the car is unable to steer itself would also be a situation in which the mechanism for steering had failed so badly it wouldn't work for a human either. You think there must be a method for a human to assume control, as some sort of ultimate fail safe, but it has been repeatedly demonstrated that the opposite is true.

The only needed airbags with rearward facing seats might be side impact ones. We only have driver and passenger airbags now because people love to drive into stationary objects while sitting in forward facing seats. Flip the seats around, and the seats will absorb the energy of a sudden stop collision. Of course, now stationary impacts (getting rear-ended) become a problem, but we don't see those at the same frequency or at the same speeds as forward motion accidents, and airbags may never be put in for that.

Self driving equipment is expensive now, but it's already below ten thousand for prototype SAE Level 5 equipment. The 'self driving' add on for Tesla Model 3 cars is only $5000, though I'm not sure if that reflects Tesla's actual costs. If we begin to produce the equipment in enough volume to equip onto most new cars, the cost will go down even further. I'm going to say, outright, that once we introduce the first SAE Level 5 self driving car that the public can purchase or rent/hire, that within a decade the costs for all the electronics and sensors to equip a car for self driving will drop below 500 current year dollars (whatever their inflation adjusted value). Most of the sensors are already in mass production for use on cars already, they are your parking, adaptive cruise control, lane departure warning, etc sensors, and they cost very little already.

A no frills car produced in a reasonable volume will be really cheap compared to todays cars. It just won't need to meet the same ascetic design standards, which will make both design and assembly cheaper. We are also reducing the number of parts going into the vehicle and making some of those parts much less expensive to make. This will be the k-car of electric vehicles, and it will be very inexpensive to make. If you think removing much of the interior of a car will save 'maybe $100 a car' just the radio alone costs twice that amount.

And yes, these will be soviet drudge mobiles. Not because the state mandated them to be, but because someone will figure out that if a working poor person is spending $200 a month on gas and insurance, that offering them a hire car service for $175 a month will be a viable proposition, and they will work towards making (or having made) a fleet car that only costs $100 a month to run, all in (amortised initial price, maintenance, everything). I've repeatedly said this is the long tail of self driving, and this is what that long tail looks like.