Tesla and Uber need constant hype to keep investors believing. Even Google needs to diversify away from its desktop ads business, or show it has significant room to grow. For a while last year I was taken in by the hype and thought self driving cars were really a lot closer than I previously admitted, but then I realised there is an agenda pushing these success stories.

Is it that you don't want self driving cars to happen or you don't THINK it will happen? I, for one, am super excited for self driving personal vehicles and will be glad to see the eradication of mass public transportation like trains and buses.

There's a strong case to be made that this isn't a desirable outcome in a lot of areas. Trains and buses are much more efficient at moving people than individual cars, even if they're self-driving. I'm also extremely excited about autonomous cars, but neither desire nor think they will replace mass transit.

edit: As a few people have mentioned, my suggestion that "trains are buses are much more efficient" is dubious. And I certainly don't mean to say they're better in all circumstances. It's more accurate to say that in certain scenarios, it still makes sense to use mass transit instead of self driving cars. Moving inside Manhattan, or commuting from a suburb to Manhattan, are examples where mass transit likely makes sense. With the obligatory disclaimer that I could be incorrect, this is simply my understanding.

I think you overstate how much better mass transit is than self driving cars. Transportation is fundamentally about traversing a graph, and transportation infrastructure is about maximizing the value of the flows on that graph. Mass transit only makes sense when the edges become a certain weight. The vast majority of point to point connections are far too light to support mass transit. Car transit is a bit better for these P2P and near P2P connections, but still pretty mediocre, because of coordination costs.

To a significant extent the issues with mass transit can be remedied by breaking the global graph up into connected subgraphs, and each subgraph can be similarly subdivided, recursively down to the point where the form factors and utilization rates cause the economics to break down. The edges between those subgraphs is mass transit.

Autonomous technologies make both point to point and mass transit transportation much cheaper. But labor (market or not) comprises a much more of car transportation costs than mass transit costs. Eliminating that cost inherently drives a shift from mass transit to car transit. Combine that with an efficient coordination algorithm to increase car utilization rates, and it gets a significant upper hand.

Each approach will still have its place, of course.

Just a nitpick here, but when you write

> breaking the global graph up into connected subgraphs

I feel like it's too conservative -- and takes "the transportation graph" as too fixed an entity. Consider how much the shape of the U.S. changed between 1917 and 2017 -- suburbs, the Interstate Highways, malls, fast food, huge parking lots, and on and on. The car literally gave us new ways of living with each other, for better and for worse.

So I think we should expect changes at least that radical in the next hundred years -- which suggests that we'll not just divide some abstract, eternal graph differently, but really truly reshape it.

I can't wait for super-fast trains + self-driving cars, AND I can't wait to see what kind of world they create.

That's an interesting approach to breaking the problem down, and a way in which I hadn't thought of it. I can see that I likely overstated my initial comment. I'll edit it to clarify what I meant.

Trains are great at moving people to train stations. Buses are great at moving people to bus stops. Cars are great at moving people to their homes and businesses.

When millions of Americans live in rural or suburban areas, what do you think is easier? Extending rail and bus routes to within walking distance of every house, or convincing everyone to move to a city? The answer is neither, and that's why cars exist.

Here's one that isn't a trick question: is it more efficient to build a rail line to my office and have a train run there or to have individual cars drive there? I don't know the answer for sure, but considering you may possibly need 15 trains filled mainly with people who don't work there converging on one location or 15 cars with one person each who does work there converging on one location, I'm almost willing to bet 15 electric cars would be more efficient than 15 trains.

There's a certain density requirement that mass transit needs that doesn't always exist. That's why we have cars.

And streetcars are great at moving people that last mile.

Take that road in front of your office, and just put rails in it.

If your local government can pay for a road in front of your house, it can pay for a streetcar. This applies to dense areas as well as rural areas. Rail lines actually cost less to maintain than roads, and they cost YOU less because you don't have to buy and pay for a car.

And you don't have to fear dying in a crash every time you set foot inside a streetcar, like you do in a car.

But would a streetcar go past my house? How far do I have to walk to get from my house to a bus stop or train station? Because I own 19 acres and my nearest neighbor is three miles from my house. The nearest town is 20 minutes by car. How would public transport deal with that kind of remoteness? My office is one end of the equation, my home (and the homes of my coworkers) are the other end.

It's surprisingly common when you leave the bigger cities. 90% of the country might live in cities, but the other 10% is spread out across the other 99% of the nation.

The urbanization statistics are also misleading. According to the US census, 80% of Americans are urban. The definition of urban includes the town where I live which has about 7,000 residents and I live on 5 acres and my neighbors more.

The move to cities stats are also way overapplied. College educated young people are indeed moving to a few mostly urban dense cores like Brooklyn. More broadly there's somewhat of a shift from true rural to the suburbs but most of those people will still need cars.

You are obviously not the median person, and the median solution will therefore not work for you.

Wait now we're talking median solutions? I thought we were talking about replacing all roads with rails/streetcars? That's what the grandparent comment was saying

So if I don't fit into this world, how do I get around? Obviously we already ruled out self-driving cars, since they'll never work. So I assume I drive my own car? Since I'm already driving my own car, why can't I just drive it to the grocery store? Do I have to drive to the train station and get on a train to get groceries, or are there still parking lots for me? How far am I expected to walk from the train to the grocery store even though I have my own car and could drive right to their door? Or are we assuming self-driving cars work for me but don't work for anyone else? In elementary school I rode the bus for an hour and a half (not even exaggerating) to get from my house to school, am I making that same trip every day to get to and from work? Even though it's 20 minutes by car?

We're talking millions of people across the country here, not an inconsequential number. We're talking about thousands of small towns where trains will NEVER work. Are we just relegating them to second class citizens who have to ride their horses to town? Damn it's so easy for people in big cities to come up with these clever solutions, of course everyone can just take a train or a bus. Except millions of people can't.

But they're just minor outliers, right?

If government built a road in front of your house, they can build a streetcar line in front of your house as well.

This is the cost of living in luxury with your own 19 acres.

And if you have your own farmland, you are richer than half the country.

It's really not. Calm down and don't pick up on one minor point and reply with a giant response. Obviously no one else thinks that cars are unnecessary.

No, it really is. So far I've seen people saying buildings older than "a few decades" should be torn down to move them closer to street car lines, people calling that "urban renewal", people saying that the problem solves itself because all young people are moving to cities as a rule, people saying self-driving cars will never work and every road should be replaced with street cars that lead to busses that lead to trains, but for the rural population the median solution wouldn't work so they'd have cars of course, but you wouldn't need them because you would just tear down your home and move to the city. Oh and also farmers are rich, but at the same time farmers have to move to the city.

Every single person in this thread lives in a big city and never stepped foot on land that didn't have a sidewalk next to the road. We're not talking "minor points" here, we're talking about 46 million Americans who don't have cable internet or cell phone reception, but somehow public transport will work there or we'll just force them to move.

I knew HN was a bubble but god damn, this is some next-level shit if you can just pretend 15% of the country doesn't exist.

No, pretty much all the things you're claiming people have said are egregiously uncharitable readings of their comments.

Someone suggested abolishing public transport might not be the best idea, and you seem to have interpreted that (and all subsequent defences of it) to mean we have to abolish cars instead. No one's suggesting that, they're just saying that we should use more public transport where it makes sense.

You're in the discussion about Apple's self-driving cars.

Do you think Apple is going to create a car for the 15% of the rural people?

This is why your limited market is irrelevant and worth exactly 0.

It's your responsibility to catch up with the world, not the world's responsibility to cater to you.

> If you think that's true for every Millennial across the country, you're crazy.

Of course it's not everyone, but it's the overall trend.

> Just don't have a job, just don't go to work, I'm sure I can afford to move to Manhattan after I quit my factory job in Springfield IL because it was too far off the rail line.

Manhattan is expensive because there's only one of it. If the jobs build around the rail lines, you can save a load of money by not needing a car, and split the difference with your employer. And sure, not every job will be able to do this - factories need space. But that's a small portion of our economy at this point, and getting smaller all the time.

"Trains and buses are much more efficient at moving people than individual cars, even if they're self-driving."

I've said this on here before, but I'll say it again: efficiency is not everything.

Individual houses (or even apartments) are less efficient than having everybody living in barracks, hot-bunking the beds in three shifts, and eating some sort of nutrient-balanced kibble in cafeterias.

Human beings have goals other than efficiency. One ignores this fact at one's own peril.

> Individual houses (or even apartments) are less efficient than having everybody living in barracks, hot-bunking the beds in three shifts, and eating some sort of nutrient-balanced kibble in cafeterias.

Efficient at what? The comment you replied to explicitly talked about what's "efficient at moving people".

Of course some people like cars for reasons other than how good they are at getting from A to B, but is that a matter of fundamental human needs, or something more like fashion? Certainly people who commute by car don't seem to actually enjoy doing it.

>being able to go where you want, any time you want

And self-driving taxis mean you can use them for those journeys, and rely on more efficient public transport for everything else.

> in a conveyance that doesn't smell like urine

This is not an inherent issue with public transport. I take the bus to work every day and it's never smelled of urine.

"And self-driving taxis mean you can use them for those journeys"

No.

You will have to call the taxi and wait for it to arrive, and (inevitably) you're going to have to wait a lot longer at peak times.

That is not the same as going any time you want.

Well okay. If you really feel that being able to take a trip to a less accessible location at a few seconds notice at peak times is vitally important to you, then you're still free to own your own self-driving car.

But I think most people would much rather wait a few minutes on those occasions and save their money.

(And FWIW, I think it could be fairly quick, since you don't have to pay a self-driving car to sit around in a small town in case someone needs it.)

Evidence is against it. We have cabs (also Uber, Lyft, etc.) now, but most people outside dense urban cores still choose to own their own car.

That's less true of the younger generation though.

If we stop pulling oil out of the ground efficiency will become everything. The whole western economy is based on cheap energy.

Public transit in the US generally fails at the last mile for me especially in the suburbs. Self driving cars would be awesome for getting me to the train station. Currently I'd have to take a bus for that. However the bus stops at literally every block which is insane. Extremely frequent stops are a general problem I've had in both the bay area and the Portland area as well as the few times I've used public transit in LA. I'm more than happy to walk 5-10 minutes but I'm not happy to sit in a bus or train that is stopping all the time and never gets up to speed.

I would love to have reliable and efficient mass transit. However, for most places in the USA, that seems to be politically and socially impossible. For whatever reason, Americans have almost completely rejected public transit.

For that reason, I'm more excited about self driving cars than about public transit.

How much energy does it take to move 500 cars?

How much energy does it take to move one train?

That's your answer.

How much energy does it take to operate a train station or get people to the train station? What is the time and monetary cost spent getting to and from a train station and waiting on trains? How many NYC or Paris or Seoul subways run at 3am?

Can't speak for Paris or Seoul, but in NYC probably about helf the lines run 24/7. Of course, that's one of the reasons the system is so hard to maintain.

Well there is, because you get autonomous trains today due to the massively reduced problem space.

My belief is that it just won't happen without infrastructure changes, such as new road designs and car-to-car communications regulations.

And I don't see those changes happening without an infrastructure overhaul, at which point you might as well create a streetcar infrastructure.

That would be terrible if it were to happen; it would mean either sprawling, inhumane cities or small settlements that are awful to live in if you don't conform to the norm. Possibly both. Personal vehicles just take up far too much land.

I suspect you would not be so happy when all the people that travel on the more space efficient public transportation systems buy cars and clog up the roads in front of you.

> It's absolutely never going to happen, ever.

Famous last words?

because local maxima vs global maxima?

For the "happy path" case, in good weather, etc., sure. But that's the first 90%. The second 90%, dealing with all weather conditions, all traffic conditions, arbitrary road hazards, crazy road conditions, etc., might take a lot longer. To the point that it's not crazy to say "it'll never happen".

That said, I do think we'll get there eventually, as I think we'll get to AGI (or even ASI) eventually. But I can't prove that either. And I can understand a certain measure of skepticism about the whole thing.

There was a day when I booted my computer, I'd have to run a few commands and enter a few keys and insert a diskette. If, at that point, you told me I'd one day have a computer that boots itself and starts programs by itself upon booting to the point where a reboot would be indistinguishable from merely turning the monitor on and off, I may have laughed at you. But here we are, in an era where when my Mac reboots I can open the lid and be looking at the same CNN article as before it rebooted.

I'm not trying to say self driving cars are as easy as building a modern OS, however difficult that may be. I'm trying to say, in the area of how advanced computers can become in a very short period of time, humans have historically been incredibly, actually laughable inaccurate.

Anyone who says "computers will never X" deserves to be shunned into nonexistence. Literally all of the proof is on the other side of their argument. Historically, computers have always been able to do exactly what humans have claimed they could not do. It's only a matter of time.

Throughout all of that development, computers remained remarkably dumb. We still program them manually and they do what we tell them to even if it's disastrous.

Driving on open roads with other humans is going to require elements of real intelligence that are qualitatively different than your OS example where they just did more programming, better, faster.

It's apples and dump trucks. What if computers are just the wrong approach for this problem?

Computers are always improving. Granted. Doesn't mean it's going to be in a direction that's useful for general AI and driving full-autonomous.

On the flip side of constant improvement is are the claims that we'd have full autonomous cars "real soon now" for each of the last 20 years.

In 2017, we're still far enough away no one will give a product availability date. This again points to a qualitatively different challenge than what came before, and the need to develop several new layers of tech (not just integrate what's there) to get the job done

One thing that isn't getting better: humans ability to drive cars. Actually with computers getting smarter, we might even be getting worse at driving cars because we're spending more time on our computers when we should be driving.

Take a modern computer and show it to a programmer from 1970 and try to convince him that computers are still remarkably dumb. I doubt he'd agree. AI always falls into this trap. Once it's been done, it's not AI. It's just an algorithm. AI is something else. And when we do something that only humans could do just a few years ago, suddenly it's not AI anymore either. It's just an algorithm. AI is something else. And when we do that, now it's just an algorithm too. AI is something else.

I have a robot vacuuming my floors right now. I just typed "vacuuming" wrong and my computer corrected me without asking, it just did it. A minute ago I asked my phone, using my voice, to turn off my bedroom lights and it did. Seems pretty mundane, I know. So too will self-driving cars in time.

Alan Turing established a gold standard for what intelligence is quite some time ago. Modern computers still fail the unrestricted Turing test.

No doubt he'd be impressed at the progress, but failing the test is failing the test.

There is a difference between behaviors which seem intelligent (your roomba) and actual intelligence. Seemingly intelligent behavior (like Eliza) only works when your counterparty isn't too probing or discriminating. It's like a Potemkin village of intelligence; looks great, as long as you don't look too closely.

Who cares about that other stuff.

Replacing half of all driving is still a world changing benefit to society.

The self driving technology that already exists could change the world with regards to the trucking industry alone, as trucking hours are mostly done in "easy" conditions on the highway.

Trillions of dollars will be saved even if self driving trucks only ever work in sunny weather on the highway.

These slow electric busses are all over the world on serious trials with real passengers.

For examples: https://www.theguardian.com/technology/2016/aug/18/self-driv...

But they have also been used in DC, Vegas, Perth, the Netherlands and other places.

It might not be a total replacement for cars but they are self-driving vehicles that are moving passengers now.