Driverless cars are stuck in a jam | Hacker News Reader

Driverless cars are stuck in a jam

economist.com

152 pointsscottlocklin534 comments

If we expect driverless cars to be further delayed and we continue building our cities to encourage micromobility (bikes, electric scooters, segregated lanes, car free streets, etc.) then at which point does the value of the driverless car become moot because we no longer drive?

Does this sound unlikely? Consider European cities like Copenhagen and Amsterdam which have encouraged cycling over the last two decades and all the other cities, like London and New York, that are embarking on ambitious transformations to create friendlier streets for people.

If you don't need to drive why would you buy a driverless car? Also, why would I prefer a driverless taxi if an electric car with a human driver costs no more?

Note that I'm only asking about city travel, not intercity transportation.

But why do we even need to cluster in massive cities, when today’s technologies allow for remote work for pretty much any non physical type of work? Accountants, developers, designers, architects, customer support, and so on can all be done remotely. That would reduce both the need to drive and to cram people in overcrowded public transport options.

Because cities make a lot of things much better and easier.

There are definately downsides, but they can be solved. We have systems to collect and clean waste so that is no longer a problem. When cars can be mostly eliminated and the ones that are left will be autonomus, that’s the traffic and air pollution problem mostly solved.

Then we can also have more of untouched land for wildlife and our general pleasure.

>Then we can also have more of untouched land for wildlife and our general pleasure.

I live in a rural area in Ohio. Any eligible land that isn't being tilled already gets converted to farm land. Plots are getting consolidated, treelines are getting leveled and burned to gain another couple acres of arable land.

I think if we all move to cities there will be some natural parks but most of it will be converted to ag and industry. If you get a couple million acres of privately held land, who knows what's happening in there.

True, that’s another problem to solve. How to reduce each person’s nutritional footprint while improving their nutritional intake.

Gravity is still an important aspect of an economy. People congregate near food sources(jobs). Companies congregate close to the other services they need because it costs less.

as long as our food is on one specific location so will we.

If the trend of subsistence farming because more than that we might see a shift in urban areas.

It's not just a matter of work. Good luck finding my local theater's shows in Greenfields, nowhere, USA.

Sure I can work from anywhere, but businesses (especially leisure) tend to concentrate where people concentrate.

Just like when I went trekking. I asked our guide once, _where do we have to go for the trail to be less crowded?_

He named a few places. Then mentioned there were no stairs on the slopes of the mountains he just named. No hot water for showers in the lodges on the way. Much fewer places to stay to eat or sleep. And not as much choice on the menu.

Wether it's around the Annapurna or in the US, businesses and services develop and concentrate where there they can find more customers. People go where they can find the products, services and experiences they long for, and the people they care about.

Many people like living in cities. There are many dozens of restaurants within walking distance of my home and quite literally thousands if I hop on the tube. How would my life be enriched if I lived in suburbia?

Many people like living in she suburbs. There they are exposed to less traffic and emissions from said traffic. We discussed this [1] last week and it seems less traffic has a major effect on air quality. The air you breath would literally be enriched if you lived in suburbia. But I do get your point, I live in a small town and when I travel to a place like Vancouver, it is fascinating to be able to choose from so many different venues and options throughout the entire giant city. [1]: https://news.ycombinator.com/item?id=21322291

I think both points of view are accurate. Cities offer a number of advantages but so do towns within commuting distance of cities.

Because you have some unique interest. In a city you can fill it, not in a small town. Do you want a tank of Argon - you won't get that in the smallest towns (argon is useful enough that many farmers have a tank for welding). What about a new magic trick - order from amazon and hope it isn't cheap junk, or buy from the local store and see it in person first - you need a larger city to support that store. Maybe you want see a Broadway type musical, the larger the city the better talent they can afford to act. (in small towns the main actor is still pretty good but the support actors are not very good)

The reason why is scaling of infastructure. It is cheaper to provide services for 1000 in one skyscrapper than 1000 in a small town. Even if we had FTL communication for full cyberpunk "jack in robot bodies remotely" those benefits would remain. It is true that there are also costs to the arrangement but cities are here to stay and growing even.

Add in doctors and other specialized labor which is important but thankfully rarer in need and have a larger "patient footprint".

That small towns are mostly dying is no accident or sinister conspiracy - they are victims of not being able to compete with scale barring some niches like say tourist towns.

The only implausible escape from cities as a situation involve not only teleportation but cheap teleportation which effectively scribbles over many spatial constraints.

The big reason is money. We now have large multinational corporations that service the entire world but have headquarters in one or two cities. So all the high paying jobs end up in a handful of cities.

I live in an completely unremarkably small city except for the fact that there is one big employer with a lot of money. The rents and house prices are only going up.

And many of the companies are not actually in the urban core. And people live over a fairly large radius from the central city.

The discussion always seems to assume that you live in a dense central city or in Nowhereville USA. In reality, a huge number of people live in suburbs, exurbs, and less dense parts of the city.

There's no shortage of local shopping. And "downtown" is easily accessible for those times when someone wants to take advantage of the resources (like better restaurants and theater) that are mostly less available outside of the city proper.

Physical interaction is still extremely important to societies and individuals; not everything can be virtualized or made remote without loss of effectiveness or efficiency, at least not yet.

For personal interactions, you have face-to-face talking and generally (in the context of clustering) access to many people with aligned interests to talk and interact with in person in real time.

Culturally, you have art installations, exhibitions, and all sorts of activities like musical concerts, theater, etc. that aren't exactly reproduced remotely.

In terms of work and hobbies, clusters can give access to tools, workshops, that might be too expensive to have at home. Say using a hackerspace/fab lab vs. having a complete electronics lab in your house or a laser cutter.

All of those should be more readily accessible (in terms of transportation time and cost) if you are well located within a large hub.

Of course, remote does have significant cost benefits in some respects (housing and supplies), and for some professions the benefits of cities aren't strictly necessary, but clustering and the physical world should not be underestimated or thought obsolete.

For two main reasons: efficiency and preventing habitat loss.

Higher densities of people make many things more efficient, e.g. shared heating in apartment building, making mass transit viable, sharing resources like food and water. Climate change is forcing our hand on transitioning to a higher-efficiency society, but we would have to do it soon anyway because we are running out of oil. When you can't enjoy the energy density of fossil fuels, efficiency will start to matter a whole lot more. Our current mode where a single person might travel up to 50 or 100 miles every day, just to live their life, is a ridiculously inefficient way of spending your time on Earth.

Secondly, habitat loss is the single greatest cause of the mass extinction of species we are currently experiencing. Most of that is agriculture, but urban sprawl certainly plays a big part. If we want to enjoy, you know, other species existing, we should be looking for ways to reduce the amount of land we need to occupy as humans.

Also, a bonus: kids that grow up in walkable neighborhoods have been shown to have the highest levels of upward economic mobility. So really if you want to do right by your kids, you should be moving to a walkable area and reducing your car use.

Lots of people don’t live in urban cores. And it’s outside those cores where people can drive significant distances where self-driving has the most value. I actually expect we will eventually see self-driving on limited access highways and other “easy” situations. Self-driving in cities and as taxis may be indefinitely postponed.

> Lots of people don’t live in urban cores.

That's what has to change. Both the location of people, and the definition of "urban core".

> That's what has to change.

Good luck with that. People like living in the suburbs.

It's where I live and when self-driving cars are available I'll move further out of the city.

Where people live is always a compromise. I want my mansion on 100 acres within a few meters of New York's subway (any station), and no more than 20 minutes from my job (in the midwest - I'd have to average 3x the speed of sound to make that trip time). The above of course is impossible, even if you are the richest person in the world.

Yep. Where I am now is definitely a compromise between convenience and space. A self-driving car will make it possible to increase both my convenience and space.

As a practical matter, meaningful change to the population density distribution will take decades, if not centuries. And it will not happen at all until we can find a way to make urban living less brutally expensive.

And yeah, urban core will need to be redefined. Not everyone wants to live in a skyscraper. We need to strike a balance.

Not everyone lives in cycling distance to work, not all are fit enough to cycle regularly (old age, disability, obesity).

In many countries such people use public transportation, but that's crowded and inconvenient for the richer folks who can afford cars.

Why do you think people still have cars in Amsterdam?

I think many will still have cars, although they'll have to be electric [0]. Will the driverless car be as cheap as the second hand Nissan Leaf? Also, perhaps the car isn't used to commute every day when the weather is good and you aren't battling through traffic.

In Europe cycling distance to work within a city is as quick as driving. If you aren't quite fit enough you can get an electric bike (very popular amongst the elderly).

I didn't consider the suburbs in my question but in a small to medium sized city (<0.5m) I'd suggest the average commute would be ~8km. That takes ~25mins on a bicycle. Now project a better micromobility infrastructure and cycling looks much more attractive. All those school runs just vanish.

So you'd have to rich and fat to still be driving. ;)

[0] https://www.theguardian.com/world/2019/may/03/amsterdam-ban-...

> Will the driverless car be as cheap as the second hand Nissan Leaf?

Eventually, yes. That is the march of technology. Especially if Elon Musk's bet that it can be done using only cameras is right - then the hardware cost is the same as an iPhone. If LIDAR is also required, it will take longer for the prices to go down.

This year I bought a 2-year old used econobox car (63 horsepower of pure fury). It already came with 360 degree "top-down-view" parking cameras, stereo camera auto emergency braking, lane departure warning, green light departure warning. And it's a mild hybrid. Once the software is developed, it won't take more than a decade to start coming down in price.

I am not at all bitter that my brand new 60K electric car does not have top-down camera ability. Or CarPlay. No, I am not bitter at all.

I drove a rental Nissan Rogue a few weeks back while we were in DC and it had top-down camera, auto steer, adaptive cruise, heated everything, etc. And that is not exactly the most desirable car you can buy. The technology really has trickled all the way to the bottom, but some of it needs to trickle back up...

I guess it's like hotels. Any cheap roadside hotel will give you free Wi-Fi, but the expensive luxury hotel will charge you $10/night for the pleasure (not sure how true that is anymore, it was a couple years ago)

Also, sometimes it rains.

> Not everyone lives in cycling distance to work, not all are fit enough to cycle regularly

And don't forget lack of showers at workplaces. If there wasn't a shower available where I work I definitely wouldn't bike to work.

> not all are fit enough to cycle regularly (old age, disability, obesity).

Obesity is one thing promoting modes of transportation other than driving is a hope to solve; where the developed world is showing highest rates ever and scientific studies showing that the chair of a car is unhealthy.

That may be happening in Europe but for 95% of Americans it's sadly a pipedream.

What is stopping American cities from changing their infrastructure investment from automobiles to micromobility? Is there one, big, standout reason that could be addressed? It's a change that takes decades and it's fighting against incumbents with deep pockets but it could be done.

Democracy.

Other than Manhattan, everywhere else a majority of residents own cars (Even in Manhattan I think it’s like 40% or something own a car). Moreover, the last 70 years of migration have led to a common pattern of people living in the suburbs and commuting via car to their workplace - replacing that with micromobility feels like a threat to their way of life and they will fight it vigorously.

What we actually see happen is even small tweaks like the 14th street bus lane in NYC get tied up in court and fought via democratic processes, when they do succeed it’s after a lot of effort.

Car ownership is declining and will continue to decline because it's an expense many people would rather avoid.

Car ownership in the US has been steadily increasing over time. There's a small increase in "no-car households" since 2010, but it's around a .5% increase.

https://investorplace.com/wp-content/uploads/2019/04/Screens...

What do you mean by micromobility?

Walk? There is no place you can walk to from your house. Our strict zoning codes have carefully kept anyplace you might want to walk to farther from your house than walking distance. Even if we eliminated all codes today (ignoring all downsides) it will be many years before it makes much a difference.

Bike? Cities in the US are too large and spread out. You can get farther in a car than a bike. There is some opportunity, but not much because few people live within reasonable bike distance of work so they will have a car: once you have a car you may as well drive: you can fit more in the trunk than the bike racks, so it is easy to be lazy.

In addition, streets tend to wind around, you often spend far longer than desirable getting out of your neighborhood. This is only a few minutes by car, but on a bike it is a lot longer and you feel that when you spend a lot of time going the opposite direction of your destination because that is how the roads go. This is really hard to fix without knocking down houses. This same issue mean that transit cannot work, even if density is high enough (it typically is!) the winding roads mean the bus can't run fast enough to get people anywhere in a reasonable amount of time.

Micromobility is dangerous. Want go check injuries per mile stats?

Tell you what, you do it and report back, eh?

While you are at it, check the traffic deaths due to motor vehicles. Do it for urban environments, which is the original topic.

Oh, and the improvements in life expectancy for those who take regular exercise by, for instance, cycling to work. Better offset those. I suggest using micromorts.

>then at which point does the value of the driverless car become moot because we no longer drive?

When nearly everywhere people live is sufficiently urban to for it to not be economically insane to deploy good enough public transit to replace cars in all those areas.

So not in the lifetime of anyone discussing it here. Maybe never if predictions about population leveling off hold.

If you don't need to drive why would you buy a driverless car?

I don’t anticipate many people will ever buy a driverless car.

The whole concept of buying a car is kind of ridiculous. You pay for it then you continue paying for registration, insurance, maintenance, park, etc all while it sits idle 90% of the time.

Driverless cars allow you to be picked up at your front door and dropped off exactly at the spot wherever and whenever you want for a low fee per trip. And we’ll need about 1/10 the number of cars that we have today.

A lot of those benefits can be had from taxis - and yet people still buy their own cars.

Taxis are expensive because they require a human to drive them. Hence the reason the SDC could change that dynamic. A shared car that does not require an expensive human to drive it.

Another model are shared cars, like dockless scooters and bikes. They are reasonably popular in the Netherlands for example. But they still require quite a bit of planning to ensure the car is available, and they are not always "in front of your door". Also you need to rent it for the entire time that you're away.

Irrespective, SDC will take some time. Like NLP, they work will within a constrained problem domain, but don't really generalise.

But I'm sure I have done the arithmetic and worked out that taxis would be a lot cheaper than buying cars - but I still ended up buying a car.

Mind you I live in a rural area and spend a fair amount of time (1 day week) in areas with no or very poor mobile phone coverage so that might put me off!

> Taxis are expensive because they require a human to drive them. Hence the reason the SDC could change that dynamic. A shared car that does not require an expensive human to drive it.

Human driver's only account for 50% - 55% of a cab's cost.

https://urbankchoze.blogspot.com/2014/08/what-taxis-can-tell...

https://www.quora.com/What-is-the-breakdown-of-where-taxi-fa...

"Only" half of the cost? That's more than enough to make it economically viable for a large chunk of the population.

Many people's cars are idle at the same time and in use at the same time.

Just because everyone's car is idle 90% of the time doesn't mean that we can replace the fleet with a driverless fleet that is 1/10th the size.

No state has an average commute time of over 31 minutes for solo drivers, while peak commuting time is several hours long. We could easily get by with less than half of the number of vehicles.

You buy a car because it is there for you. When you want to go it is always where you left it and ready to go. People tend to forget about this: you don't have to reserve your ride ahead of time, you just go when you are ready. To be useful shared cars will be idle most of the time as well: if they are not you cannot be sure one is ready and waiting for you when you decide you want to go. Otherwise you may as well take the bus which is cheaper than a shared car can ever be because more people are on it at a time.

Since there isn't much savings possible in a shared car without losing the whole advantage of a car you may as well buy your own: the ability to leave your golf clubs in it is more than worth the extra costs.

I buy cars so that I can leave my stuff in them instead of having to load and unload every day. That convenience is worth a lot. Whether the car drives itself or not is an entirely separate issue.

Somehow I doubt work would be willing to relax their weapons policies so that I could bring in the training weapons that sit in my car.

In most states they don't have a legal choice in the matter.

>I don’t anticipate many people will ever buy a driverless car.

Unless buying a fleet of vehicles over time produces better gains/results than other retirement/investment options (fossil fuel divestment or 401k replacements, plus having the convenience of immediate SDC access).

I think cars embody the classic american spirit of freedom and "wind in your hair" fleeting moments.

After gen-x, the cost of vehicles skyrocketed to parity with what modest accomodations used to cost, so the younger generations are writing their story with a much different frame of reference.

I enjoy the saying "In America,100 years is a long time and in Europe, 100 miles is a long distance".

> all while it sits idle 90% of the time.

Thought experiment: Most of your clothes sit idle 90% of the time.

Most of your clothes sit idle 90% of the time.

I have a cup full of pennies in my closet that sit idle all the time too and I don't care... cause they're pennies.

Similarly, the average car cost over $35,000 in 2018. That's a sizable expense for most people. Clothes in comparison, not so much.

Whoosh :)

I think the whole point of driverless cars is at some point, you can deploy car sharing services that largely replace private car ownership/usage. At least, that's what I'd hope, because it's an insane waste of space and resources that the vast majority of cars you see on the road only transport a single person around. I mean, I get that people don't care about their environmental impact, but at some point we're going to have to care.

Why do you think people using a driverless car service would care less about having to share their ride? In theory if people would just carpool much more right now, the number of cars on the road would go down a lot, and they could divide the costs. But somehow they don't.

I don't see why people who don't consider carpooling right now would order a driverless car that needs to pick up some other people before going where they want. So how would driverless cars reduce the number of cars on the road? Or to state it differently: why do we need driverless cars to get the average occupancy of cars on the road above 1?

Convenience. Habit. Cultural pressure. Regulatory changes. Etc.

It isn't as simple as "People don't do it now, why would they do it in the future?" There are many ways to make automated car sharing far more attractive to people than car pooling ever could.

This isn't a technology problem, it is a geometry problem. Car pooling/sharing cannot work very well because eventually the car needs to go down some "dead end" that you didn't want to go down to pick up/drop off somebody, and that time is wasted. If you are the last one on and first one off it is great, for everybody else it is wasted time where they are not making any progress getting to their destination. Everybody would be better off (less congestion - less expensive roads need to be built) if everybody carpools, but you are always personally better off to be the one person who doesn't.

It may seem like you can solve this by just picking up in the neighborhood. However there are not enough people in any given neighborhood to make that work. This morning exactly one person (car) passed me on my way to the bus stop - I don't know if this person would go the same direction as me or not: there isn't much opportunity for ride sharing within one neighborhood.

Transit has the same problems. It solves them by forcing people to walk to the stop/station. At least everybody feels like they are making progress. They also come on a schedule which means you can't leave when you want to.

They may not have to share a ride.

We (as a society) are fully capable of building single-person cars. It's just the economics of car ownership incentivize having four to five seats and a lot of trunk space just in case you need it, even if you aren't using them 99% of the time.

That's likely to change if the cars are being run by a driverless car service. They'll already have the numbers to know how many single-person cars to maintain in the fleets as compared to five seaters. Fuel / electricity will be cheaper for smaller vehicles. If traffic does notably increase, tolls and congestion taxes will come more into play and be lower for smaller cars. Single person cars could split lanes (like motorcycles).

At that point, the economic incentives change, and we could see different outcomes.

Convenince has historically proven the industrial era and beyond killer app. Availability is part of the issue - it would increase car utilization but probably not people in a car at once. They would care less about "my car isn't available because it is out working" if they could get a replacement. They also don't need to give their time driving passengers.

Self driving also could allow better "microroutes" which don't take you out of the way like public transit. It wouldn't address the stranger adversion however.

Ironically in the worst case it may technically reduce the average occupancy even if the total utilization rises and thus number of cars on the road drops for the same demand.

Even in the inner city there are times when a car beats everything else. Or more likely a truck. The store you get groceries from gets deliveries every day, and they probably should have a special waiver to drive on that car free street to get the truck to their back door (I've seen areas in Germany where the old part of town was effectively car free except delivery trucks)

Most people will not need to drive very often, so they would not buy a car at all. However everybody has situations where driving of some sort is the best option.

There are places where it's cold, where it rains and snows. So we can't bike everywhere.

The good thing about the driverless cars is that they solve the parking problems in the cities. We can even add lanes or pavements by removing all parking spaces in the city centers.

You can absolutely bike in the cold, and in the rain, and even in light snow. Also, public transit & busses can get you further without even the dreaded prospect of burning a calorie.

Removing free rental of public space to park your private vehicle is something I _do_ endorse though!

After commuting by bike for a number of years the only weather than I found sensible to avoid was very strong winds - I got blown off my bike onto the side of a lorry once, fortunately I was OK but I did get a scare!

In the Seattle area, if it’s 38F or below, I use four wheels. In 35 years of riding a motorcycle, the singular time I’ve dumped one on the street was black ice when bike’s temp sensor was screaming, “it’s 38F, ice! Ice!” One time it was right. :-( That, and I’ve dumped enough bicycles in similar conditions that I’ve just quit doing it. So total, there’s about a week of days every winter that I drive. I’m fine with that, because as I’ve mentioned elsewhere in this thread, I ain’t young anymore, and that collarbone is gonna snap one of these days if I don’t start exercising some caution.

> The good thing about the driverless cars is that they solve the parking problems in the cities.

They only solve them by doubling the traffic problems. I'm not a civil engineer so I'm not sure which problem is harder to solve, but this hardly seems like a free lunch.

> There are places where it's cold, where it rains and snows.

Like Amsterdam and Copenhagen?

> The good thing about the driverless cars is that they solve the parking problems in the cities.

They solve the search for parking by humans. But if you want to get the cars out of the city while they aren't used, you practically double the traffic and emissions.

Amsterdam and Copenhagen don't get very cold.

The canals freeze. That's sea water, too.

I'm not saying it never gets cold, just that the climate is much warmer than, say, Wisconsin.

I grew up in Southwestern Ontario and we had days were school was cancelled due to cold, not snow. They were worried about frostbite on any exposed skin.

No, not like Amsterdam. I love Amsterdam, but I'm so tired of it being used as an example of how we should all be on bikes. It has a very temperate climate. The weather in most of the US isn't as mild and cool as Amsterdam.

There are cites in the US with great weather, for example San Diego, Santa Barbara, all of Marine and Sonoma Counties in CA have great weather (sadly it is having wild fire problems right now). I agree that we could be biking in those places more.

The US as a whole has climate all over the place, from polar (N. Alaska) to tropical (S. Florida) to a desert. The coldest it has ever been in the Netherlands is -27 °C, but the continental US has gotten down to -57 °C (Alaska has reached -62 °C). The hottest it has ever been in Amsterdam is +41 °C while the US has reached +57 °C.

It's true that these are crazy extremes and we can't decide on public policy based solely on rare events, but there are big cities all over the US where it is too cold to bike. Consider:

Average Weather: days at or below temperature:

    City (population)   -18 °C    0 °C  
    -----------------   ------   -----
    Amsterdam   (821 K)   none    none

    Minneapolis (422 K)    23      148
    Chicago    (2716 K)     7      122
    Denver      (620 K)     7      157
    Milwaukee   (595 K)     7      126
    Detroit     (673 K)     4      115
    Indianapolis(872 K)     4      104
    Rochester   (208 K)     4      129
    Buffalo     (259 K)     3      124
    Cleveland   (386 K)     3      104
    Hartford    (123 K)     3      125

It also gets too hot to bike in the US:

Average Weather: days at or above 38 °C:

    City   (population)      38 °C
    -------------------      -----
    Amsterdam   (821 K)       none

    Phoenix       (1626 K)     107
    Las Vegas      (642 K)      70
    Riverside      (328 K)      24
    Dallas        (1341 K)      17
    Austin         (951 K)      16
    Sacramento     (502 K)      11
    Oklahoma City  (644 K)      11
    San Antonio   (1493 K)       8
    Salt Lake City (201 K)       5
    Houston       (2313 K)       4
    Kansas City    (489 K)       3

Some very hardy people do bike in the cold. See [1], the worlds longest winter ultra-marathon (running, biking or skiing 1000km, the last 500 miles without support!!!) It might be the hardest race in the world. (There are not very many other races where contestants have had to be evacuated by helicopter because they were being stalked by a pack of wolves.)

[1] https://www.iditarodtrailinvitational.com

Minneapolis has some of the highest percentage of Bike Commuters in the US at around 3.5% only falling after Portland at ~6% and San Francisco which is a nudge closer to 4%.

The entire west coast has weather at least as favorable to bike commuting as Amsterdam, yet the percentages of people commuting that way are still far lower than Amsterdam simply because the infrastructure isn't there.

People who think nothing of skiing or snowboarding in winter conditions are convinced that biking is impossible in those same conditions. It makes no sense.

Another factor to remember is snowfall. Where I live, we push snow to the side on roads, but it tends to stick around and accumulate for months at a time. Adding a secondary* infrastructure around bicycles would be costly.

But the big one is that the Netherlands has about half the population of Canada in an area the size of New Brunswick. Obviously they can afford better transport infrastructure than we can.

* Where I live, bike lanes become snow banks circa mid-January.

Where they expect annual snow and have segregated bike lanes you simply plough those, too. I've seen this in Sweden in Uppsala. You bike across a hard pack of snow, just like the cars on the roads. The cars need studded snow tires but the bikes are okay without. In the melt when you can have ice patches then studded tires on your bike are possible.

You mean places like... Amsterdam? Plenty of rain here. That won't stop people from biking. If the majority of people own a bike, but not a car, biking becomes the default. Even with rain.

No places like Wisconsin and Minnesota where there's week spans of high temps that barley get above 0F and wind chills that will give you frost bite if you're exposed for too long.

I live in Seattle, rain sucks to ride in and most people don't have facilities to change or shower at work. That's a luxury not a standard.

It's all a moot point anyways because all this assumes everyone lives within biking distance of work and that work trucks/vans don't exist for a valid purpose.

I went to school in MN, and regularly did 1 mile walks in all weather including highs of -25F. You learn to dress for it.

I keep forgetting that since the dawn of time, humans have been so sensitive to environmental conditions that it is a wonder we survived as a species. I get what you're saying, but in the end it boils down to "it must be 72F and dry anywhere I go". A little rain isn't going to kill you, leave the car at home today. Don't want to ride in the rain? Then don't ride in the rain, I certainly won't judge. But it is not some failing of the design of human beings that makes us incapable of doing so.

they solve the parking problems in the cities

By turning the general traffic lanes into de facto moving parking spaces?

So you're admitting that your solution involves a reduced quality of life?

Trench foot builds character.

>There are places where it's cold, where it rains and snows.

So what? Can you only bike on sunny days? Get better gear.

Hell I motorcycle year round even when it drops below freezing. When it gets real crazy I just take the bus for a day or two.

This is kind of silly. It's nice to imagine a world where people don't need cars. But most/all european cities, many of which we laid out before cars, still have cars.

We all have cars because cars are incredibly useful; Got 4 young kids and you need to get groceries for the week? You're not going to bicycle them to the store and all that food home again in the mountains. Live in the city and want to go hiking 20 miles away? You're not going to settle for 30MPH when you could be doing 70MPH to get there. The reasons for cars are endless. That's why we have them.

(BTW, I pretty much only bike in my city. But it's pretty obvious why people have cars.)

"Live in the city and want to go hiking 20 miles away?"

Yes, you want to use a car in this case. But, in the life I imagine, I would rent a car. I dislike driving intensely, but I can see use-cases in which cars are necessary. However, Americans have been largely forced into car ownership, despite the expense and hassle of owning one. A reduction in car ownership coupled with more pedestrian-friendly community designs is more my goal than making cars extinct. I would like my household to go from two to one vehicle.

If you have to rent a car to go hiking, you aren't going to hike very often.

And why is that?

Dont pretend like its normal to go through the hassle and expense of car rental just to do a routine activity

Are you thinking of full on rental like Enterprise or car sharing like Zipcar? Zipcar is pretty dang easy and it's what car-free people think when you absolutely need a car for a trip.

Thats more or less prohibitively expensive, because your hiking location would be outside of parking zone most likely, thus you're on the clock the whole time. Not my idea of a relaxing experience. Looks like zipcar is even station bound which makes it worse. You have to get yourself and all the stuff to the station somehow and then take it back from there again when you are done.

> Thats more or less prohibitively expensive, because your hiking location would be outside of parking zone most likely, thus you're on the clock the whole time. Not my idea of a relaxing experience.

An extra hour or two doesn't make a lot of difference when you're already booking for a weekend. Of course if you get stuck on a mountain overnight or something then you'd have a penalty charge, but frankly that would be the least of your worries in that case (and it would still be a problem if you had your own car but were worried about e.g. missing a day of work).

Of course hiring a car for a weekend costs a couple of hundred quid, but realistically you're going to do that what, ten times a year at most? You'd end up paying more in fuel, insurance and deprecation on a private car.

> Looks like zipcar is even station bound which makes it worse. You have to get yourself and all the stuff to the station somehow and then take it back from there again when you are done.

The "station" is around the corner at most. Walk 5-10 minutes, collect the zipcar, drive it to your building and load up, and then the same in reverse when you get back.

> Of course hiring a car for a weekend costs a couple of hundred quid, but realistically you're going to do that what, ten times a year at most?

Nice. The “at most” is an interesting rhetorical (and grossly dishonest) trick you used there.

Go on then: how many times did you actually go hiking over the weekend in the last 365 days? In my experience people greatly overestimate how much time they're actually going to be able to spend on their hobby; they plan things out with the idea that they're going to spend every weekend doing xyz, but life gets in the way.

Hiking? Once or twice. Then again I’m more of a watercraft and skiing person as the season dictates.

Anecdotal experience from NYC here:

Owning a cheap car is way less expensive than renting a ZipCar. My car's only cost beyond gas was insurance. And my insurance was cheaper than renting a ZipCar for one weekend.

Obviously it depends on the car you own, but ZipCar's here can certainly be more expensive and more inconvenient than owning.

When I looked at it in London, the monthly charge for a parking space in my building alone would've cost about 6 weekends' ZipCar rental. I could've rented a cheap garage further out and bought a cheap beater, but at the point where I'm walking further and driving a less reliable car (and likely less safe in a collision) the ZipCar ends up more convenient.

Do you park for free on the street?

When we visit Brooklyn we usually get free street parking.

> Of course if you get stuck on a mountain overnight

This is often my end goal when I go hiking

Yeah I was thinking of enterprise. Its just weird some people think you can survive the suburbs without a car.

You can. Just not American suburbs, apparently.

Depends on the density and transit available. There are some American suburbs of big cities where you can live with no car, but not many.

Also depends on the availability of sidewalks. Some suburbs are designed without them.

The parent of this thread explicitly states:

>Note that I'm only asking about city travel, not intercity transportation.

I don't know why these conversations always seem to descend into talking about the suburbs. Urban environments can (and should) reduce car traffic inside them. Talk about leaving the city to go hiking and suburban living (designed specifically around personal car ownership) are not what we're talking about here.

Car rental doesn't mean going to your local car rental shop. It more often means using a car share service like ZipCar, Modo, or others, and it's so convenient that I often use it despite owning a vehicle.

This isn't true, talk to folks in downtown Seattle. Anecdotally, I have a group of 5 that rents a car to go hiking about every other weekend. Splitting $70 between 5 people twice a month beats buying a car by... a lot. Know tons of others that do this.

On the subject of Seattle, there are county transit busses which go to some trailheads: https://trailheaddirect.org/

Depends how painful you make it to rent a car. If you're talking about a traditional auto rental where there is a half hour of filling out paperwork and doing checks with obtuse and baffling fees and gas options and everything then yes, that's an impediment.

A future where you have an Uber-like app where you summon an autonomous car that takes you to your destination for a nominal fee then who needs to own a car?

> A future where you have an Uber-like app where you summon an autonomous car that takes you to your destination for a nominal fee then who needs to own a car?

People which have items in their car that cannot be taken into the places they have to go, such as workplaces. Many people might not have that problem, but it is a problem.

I live in one of those european cities. We have at least six companies offering small electric scooters, even more for bicycles, pedelecs, larger electric scooter, cars, trucks everything. They litter the streets can be rented on a whim with your smartphone, totally hassle free. I used at least four different of these company in the last four days. BUT I still own a car, scooter and my own bicycles because it is super expensive if you use it regularly. Usually 1 Euro up front and then at least 20 cent per minute depending on the ride maybe even more. We have public transport but its depressing and annoying and also expensive. I usually go everywhere with my own bike but its broken down atm. I absolutely love driving and its basically the only way to get away from the city on the cheap! All those fancy new mobility solutions are only good for moving within the city and if you have a well paying job. I don't ever want to miss the freedom of leaving all of this expensive busy crap behind with my own car. Its all an addition but not a replacement.

And its true that if you have to rent you don't go hiking often because the money ticks away even while you are parked outside the zone (which would be the case for hiking) so you are always stressed out about hurrying. Also you always have to decide, do I want to/can I spend 50 Euros today just to go hiking? Or you know you could use your own car. Yes it also costs money but its way less and more spread out.

>Also you always have to decide, do I want to/can I spend 50 Euros today just to go hiking? Or you know you could use your own car. Yes it also costs money but its way less and more spread out.

That sounds like someone with very stupid money management. Unfortunately, this is probably the norm.

You're surely spending many hundreds of dollars/Euros per month to own and operate a private vehicle: loan, fuel, insurance, maintenance, repairs. If you go hiking every single weekend, that's 200 Euros (according to you), which still should be cheaper than owning a car. Most likely, you're not going to rent a car every weekend. So it should be pretty obvious that you'll come out way ahead if you sell your car and just rent a car for those occasional weekend hiking trips.

>I live in one of those european cities. ... We have public transport but its depressing and annoying and also expensive.

What city is that? The German cities I've visited had very inexpensive public transit, especially if you got multi-day (or longer) passes, which anyone living there probably would.

Cars in Europe are as a rule not bought on credit.

Well in that case you're losing potential interest on that money, and also (much more significantly) depreciation.

The last time I got any interest worth mentioning on money was more than a decade ago. Yes, assets depreciate, cars more than most. So don't buy new, buy second hand.

>The last time I got any interest worth mentioning on money was more than a decade ago.

What's "worth mentioning"? My bank gives me about 2% on savings; it's not huge but it's certainly enough to switch banks over. If you're not making any interest in a bank, you're at the wrong bank.

By my calculation, my car (including running costs) comes down to about 300 eur/month assuming I keep it for 5 years and get at least 1000 eur back when I sell or scrap it. (Reality: I'll probably get more out of it, or keep it longer)

Rentals start at around 28eur/day, so if I rented one every weekend (saturday+sunday), that'd be 224 eur plus fuel. Hardly cheaper.

Of course I don't travel every weekend, but when I do, it's more likely to be friday+saturday+sunday, and I do make week-long trips (and longer) a few times a year so it balances out, more or less.

Owning a car means I don't worry about making reservations, fetching and returning the vehicle, etc. It's always near the door, and I can go on a whim if a sudden need comes up, like it did just today.

> Owning a car means I don't worry about making reservations, fetching and returning the vehicle, etc. It's always near the door, and I can go on a whim if a sudden need comes up, like it did just today.

But it does mean you worry about parking, scheduled and unplanned repairs, car wash, etc.

I think there's still an untapped marker for on-demand car hire like Zipcar and similar.

Don't forget registration and inspection every 2 years (could be more or less depending on the place you live).

Good point, I forgot that too. Here in my state, yearly inspection is a bit of a PITA but pretty cheap, but there's a "car tax" ("personal property tax") that's usually hundreds or even into the thousands per year for newer cars. Other states have something similar but they call it a "registration fee".

It's once a year here, for cars as old as mine. It's not a big deal though; it takes less than an hour and you can usually go without even having a reservation. Costs less than one tank of gas.

I mean personally I go on day-hikes relatively regularly and I just catch the train to my starting point. I appreciate that isn't always possible though.

> However, Americans have been largely forced into car ownership, despite the expense and hassle of owning one.

Indeed, car ownership is a big part of most household budgets in the US. There's the initial cost (in the $25K to $35K range now), insurance, repairs & maintenance - cars are huge money sinks. So much capital tied up in cars. A few years back I estimated that there were 3000 cars in our tech company parking lot which represents $75Million in capital (at $25K each) that's sitting idle about 95% of the time. The combination of public transit, on-demand self-driving cars, better city design, telecommuting could allow us to redirect that capital more productively.

> in the $25K to $35K range now

Why are you buying a new car? You can get a very decent car for $8K. And if you did buy a new one, you don't have repairs and maintenance expenses.

> The combination of public transit, on-demand self-driving cars, better city design, telecommuting could allow us to redirect that capital more productively.

Doubtful. Public transit usually takes 3 times as long, so you are wasting human time in exchange for capital (there's a reason people buy cars - it's not because they want to waste money). Self-driving cars will never happen, and city design can't help all that much.

From your list, only telecommuting can help, but only if you do it full time.

(Also your dollar estimate assumes all new cars, which is far from accurate.)

"Options for people who don't want/need to have a car" is very different than "maybe no one should have a car". Cars are pretty necessary if you have kids in most of the US, for example. While there's an argument that much of the US was designed in a way that forces this necessity, the reality is that the road system in US exists, the US is gigantic, and mobility within the US is generally considered desirable.

It's good to encourage cities to find ways for residents to get around without cars, if only because it can increase social mobility by increasing options for people who can't afford one. But pretty much all of the evidence I've seen suggests that Americans very much choose to have a car once they can afford one, and that is very much rooted in the fact that they're phenomenally useful.

> Cars are pretty necessary if you have kids in most of the US, for example.

Yet, there is a significant portion of the US population that can not afford a car, and the next level to afford to maintain a car, or afford to adequately insure a car.

And yes, it hurts economic mobility https://www.nytimes.com/2013/07/22/business/in-climbing-inco...

If you read to the second paragraph of my comment, I acknowledge exactly that and agree that increasing non-car transportation options can be a good thing at the same time as cars being a pretty necessary (or at least substantially life-improving) technology.

In earlier times, there were hiking clubs (which might provide transport or car-pool coordination), or transit to recreational areas.

Denizens of the Greater Smoke Area today might care to cozy up 'round the fire with a tale of the Crookedest Railroad in the World.

http://www.friendsofmttam.org/railroad/crookedest-railroad.h...

We all have cars because cars are incredibly useful; Got 4 young kids and you need to get groceries for the week?

This only work that way because that's how your life is structured, not that it is a law of nature.

If you live near a grocery store, you wouldn't have to buy enough grocery to last you a week. You can just buy what you need.

And then go back to the grocery store every day. What a waste of time. You can also wash your laundry but hand. The reality is we have these appliances (including cars) because they increase our personal efficiency quite a bit.

Going to the grocery store ever? What a waste of time when you can order groceries.

Those personal appliance have very real cost, especially in term of how much space it takes up on the roads and the amount of parking lots allocated. All of these increased the distance you have to walk to get somewhere and increased congestion.

I saw traffic in Sagion, Vietnam. Believe me, cars stick out like sore thumb compared to motorcycles and bikes, and if everyone uses car, traffic would basically grind to a halt.

With the way we use cars in America, it's very inefficient.

Nah, the suburban life was a strange fashion that took hold for a few decades but never really made sense, and thus it's now in decline as people return to the city cores. For the older generation owning a car was a status symbol, but the younger generation doesn't care.

Check back with us in 30 years. I bet suburbs will still be hugely popular in the US.

That trend lasted about 10 seconds: https://www.cnbc.com/2019/09/29/millennials-are-fleeing-big-...

I like living in the city. My wife and I did it with our daughter until she was 5. But the problem with American cities isn’t cars. It’s that they offer no solutions for parents who aren’t rich. (And almost everyone ends up becoming a parent at some point.) Schools, family centric activities, transportation, affordable housing. You need a million dollars to buy a house in a good school district in many major US cities (if one even exists). There’s not much in the way of other kids around. (Babies, maybe, but how often do you see 10 year olds in NYC that aren’t tourists?)

So, that CNBC article summarizes an article in WSJ that cites 27,000 people leaving some of the largest cities in the US. The problem is that the changes in each city are actually much smaller than the margin of error for the data that WSJ uses (I think it was ACS 5-year estimates).

This post goes into more detail on why that narrative is most likely noise: http://cityobservatory.org/no-youth_exodus_signal-noise/

Edit: 27k is approximately 0.25% of all 25-34 year olds living in cities with pops over 500k (which is what WSJ reported on). Not an exodus at all and it's really a bit of a nonsense narrative that WSJ has been pushing hard.

People are still moving to the suburbs. They're growing at rates comparable to the city core.

https://www.brookings.edu/blog/the-avenue/2018/05/25/early-d...

If you've got 4 kids all under 4yo, you're kind of screwed no matter what vehicle you drive. For biking, you'd probably be looking at a Bakfiets with electric assist plus a trailer.

Presuming we've sufficiently eliminated car traffic and the roads are safe for kids, it's also likely that at least one of those kids can bike themselves to the store with you.

4 kids will fit easily in a minivan or large SUV. Lots of families manage that.

4 kids will fit after everyone has been dressed, got their shoes on, got their shoes on again, been lifted/chased into the car/bike. That's what I meant by "screwed". The choice of vehicle doesn't really matter.

Once they're older, kids can bike themselves or walk. Families also manage that.

I've heard of one car-free mother that moved seven kids around by bike on a daily basis.

If memory serves, she got a shout-out in this book: https://seattle.bibliocommons.com/item/show/2862141030

That's pretty cool.

Funny enough, I saw a mother of 4 biking today in Boston. It was raining and 50F/10C outside. She had a bakfiets with 3 onboard and a kid pedaling behind her. I don't think they were going grocery shopping though.

How do you find a place within easy transit distance of your work, your spouse’s work, a grocery store, your kids’ school or daycare, your kids’ activities, and your kids’ friends’ houses? How do middle class people do all that in a country where cities have made it illegal to build new housing?

Used to be when I lived in a semi-civilised place that the grocers was a half block from my tram stop and three blocks from the house.

Trolley carts.

https://www.homedepot.com/p/Easy-Wheels-Mini-Shopping-Cart-i...

When I lived in DC I was across the street from a grocery store and kiddy corner from my wife’s work. I was on the other side of town from my daughter’s preschool, and my office was between my daughter’s preschool and my house. (Dropping her off via transit would’ve taken an hour and a half easily.)

> Live in the city and want to go hiking 20 miles away?

No reason trains can't solve this. When I lived in NYC there was plenty of hiking off the NJ transit. Popular hiking destinations here in the PNW have huge problems with traffic that would be helped greatly by building rail lines to them.

Why create new rail lines when driverless cars can take you anywhere?

Does nothing about the traffic problem, massively wasteful of electricity and other resources that will have to go into replacing every single car.

There is a throughput issue. Driverless cars with 4 people a car in a lane (or a bus lane) will do ~6000/people/hour if you are optimistic. A train 'lane' will on the other hand will do 32'000/people/hour. Non-bus lanes do 1500/hour, so a train is equivalent to a 21 lane highway, one way. [0]

I think driverless cars are a great adapter / collector last mile technology for suburbs to go to the train station, but if somewhere has traffic problems now, it will have traffic problems with driverless cars too. After a certain point you need to put in trains.

[0] https://www.theurbanist.org/2016/05/26/the-supply-and-demand...

> Got 4 young kids and you need to get groceries for the week?

Most people living today don't, and never will have 4 young kids. We don't need to optimize the world for the most extreme outliers, at the expense of everyone else.

I think we just need to have more flexible options that don't require owning a car. When I lived in Vancouver I didn't own a car. I got around using transit, biking, or walking most of the time. I was also a member of a local car coop. If I needed a vehicle for a short trip or to run an errand I could rent one in one hour increments. I appreciated the flexibility of being able to rent a truck for an hour or two to haul some furniture or rent a car for a day or two for a short trip.

“Got 4 young kids...”

Seems like lots of responses from able-bodied individuals without kids, elderly parents or other dependants who are able to get around on scooters or bikes in any weather and have little tolerance for those who might not be able or willing to do the same.

I live in Orange County, Ca. and bike most places most days. My wife and I share a car and we walk many places. Interestingly, even getting to the Home Depot which is only 3.7 miles away takes 20 minutes, is a bit hair raising, and I can only get small items. My last job was about 6 miles away, took 40 minutes to bike and I only did it once because it was unsafe.

But I know we’re living a charmed life. My next door neighbor is recovering from cancer treatment. If he makes it to his physical therapy appointment, it’s a good day. I’m not going to tell him to “get better equipment” or that people bike in Amsterdam so he should to. Believe me he’d love to be biking again.

It doesn’t mean our cities cant move in a more walking/biking friendly direction. I’d love to have more safe places. But have some compassion for people who aren't as lucky as you.

> Got 4 young kids and you need to get groceries for the week?

In the meantime, not having a vehicle for errands is real life for ~20% of the population, and if you add safely (maintained vehicle) or responsibly (insured, with a valid license, no warrants), this number quickly approaches 50% of the population.

https://www.nytimes.com/2013/07/22/business/in-climbing-inco...

And then you bring up hiking, which ~15% of the U.S. population participates in anyway (and it's at its height)

This ignores the massive value SDC's can provide to the blind, disabled, elderly, and children, all of whom may not be able to hop on a bike or scooter.

I did mention taxis. The question would then be, why an SDC as opposed to a taxi? In the example you give a driver would be of more use.

Because a car your own means your stuff is in it. If you use it somewhat regularly it is cheaper in the long run to own your own SDC over a taxi (even assuming automation), if only because you get to decide when to clean it.

What you're saying is taking into account only big cities and young population, but in small places and suburbs you depend on cars to do anything, get anywhere. Old people can't be expected to walk far or drive scooters and bikes everyday. And population is mostly getting older, plus with the rise of remote work I presume that more and more people will move outside the big urban centers looking for slower pace and cheaper life, closer to the nature. That means cars will be needed more, not less, and driverless cars can save you a lot of time when commuting everyday.

> Old people can't be expected to walk far or drive scooters and bikes everyday.

Yet that's what one sees in Copenhagen.

And Copenhagen is well known for its high hills, right? Not everyone lives in Holland and Copenhagen and other flat places. Also, while it's definitely great that Scandinavians live their healthy and slightly ascetic lifestyle, it's not everyones cup of tea outside of those specific nations/cultures. Maybe it should be, by all means if this is what young people want they should do it one day when they grow old. But it's silly to expect from old people who grew up in some other time to now suddenly jump into that lifestyle, especially in places that naturally don't fit it. Walking short distances, great. Public transportation, also great. Bikes for everyone... well, great for those who want it, but leave people a room to dislike riding a bike (or anything else) for whatever their reason is.

"And Copenhagen is well known for its high hills, right?" No, but Oslo is, and they are gradually phasing cars out of the city centre

That is why you have electric bikes to tackle those high hills.

Anyway, young people biking more means that more older people will actually be able to bike, because exercise helps with increasing mobility as we age.

Barcelona seemed very walkable to me after a month I saw people of all ages on foot and bikes even during the summer heat wave and it is not flat at all.

Old people can't be expected to walk far or drive scooters and bikes everyday.

You need to give some serious readjustment to your mental image of “old people”. Perhaps the word you’re looking for is “disabled” or “differently-abled”, depending upon your preference. Some of those people, maybe disproportionately so, also happen to be old.

I’d type more, but this AARP member crashed his scooter on the way to work on Friday, and my shoulder’s still a little sore. Wear your helmets, mine cracked so that my skull wouldn’t.

I myself am 40+ and seriously obese, so I'm definitely counting myself into that "too old for e-scooter" group, but I actually had in mind average around-70-and-plus people around me, like my parents and their friends, the "real old people". They can walk but can't walk across the city every day, they can go on a leisure bike ride, but not ride it uphill (and we don't all live in flatland like Dutch), and you definitely can't seriously expect them to move around that way every day all around the year. You can't expect me either, btw, really sorry if I don't match nowadays modern lifestyle standards. Which are btw ridiculous as combination of walking and eco-friendly public transportation are much more sustainable (and for many enjoyable) solution than everyone zipping around in their e-scooters and e-bikes and whatnot which are all being charged everyday from the grid.

but I actually had in mind average around-70-and-plus people around me, like my parents and their friends, the "real old people".

Guess what, me, too! My anecdata are from a more active group, though. So we have differing views of the world, and that’s fine. I’ve never been overweight a day in my life, so I try to avoid general, “all ya need to do is...” It demonstrates a lack of empathy and a small world view.

But I would ask the same of others. Yes, there are people with limited mobility. I get that. But they’re still a minority, and does it need to be brought up every g-ddamned time the subject comes up that not everyone can ride a scooter? Most people can, can we talk about them for a minute before dive off into the edge cases?

We can talk about limited mobility options, no we will talk about them, just maybe later. But the reason we have cities over run with cars now is because way too many people who are otherwise capable think they can’t hobble more than a city block, and only under perfect conditions.

And some of us who are perfectly happy to walk to get to places where practical (which is certainly not where I live) consider scooters dangerous. So, no, I’m not going to ride a scooter or a bike in a city.

For the vast majority of old people their doctor would disagree: they need to get out and walk more. Yes slowing down is important, but their health would overall be improved by walking more. For those not yet old walking more is more a mental barrier than a physical one.

> Also, why would I prefer a driverless taxi if an electric car with a human driver costs no more?

Is that something that seems likely to happen?

But even assuming it does, I suppose the answer would be because you don't have to worry about finding parking for a taxi, or walking to wherever you parked last night. With a little forethought, you can step right out your door and into your waiting vehicle.

Works great in those cities, doesn't work for the rest of the world. Everyone else is sprawled out.

> "If you don't need to drive why would you buy a driverless car?"

Not everybody lives in a city, and even people who do live in a city want to leave the city sometimes. I use my bike whenever I can, but we still own a car.

That said, the less you drive, the more interesting a driverless car might become: we won't be used to driving, so a self-driving car might be safer for us.

And the main reason I prefer to take the train over a car is that I can read in the train. With a driverless car, maybe I can read in the car (though car sickness is also an issue there).

> we won't be used to driving, so a self-driving car might be safer for us

Very much this. I've lived in London for nearly 20 years. I haven't driven in nearly 20 years. I've barely set foot in a car in nearly 20 years.

If I ever needed to drive I'd essentially have to learn all over again (ethically, anyway; not sure about the legal situation). A driverless option for those rare occasions would be great.

> Also, why would I prefer a driverless taxi if an electric car with a human driver costs no more?

I believe the number one cause of death for Americans under the age of 30 is traffic accidents.[0] It's around 40,000 deaths per year (in the US alone).

Cars are getting safer, but humans aren't getting better drivers. Individuals humans will get better with experience, but experienced drivers cycle out and new drivers cycle in.

A 16-year-old doesn't get to inherit the experience of the 70-year-old they're replacing. A new self-driving car does inherit the experience of older self-driving cars. They also learn from each other in a way humans can't.

Human drivers also improve at a rate of O(1). Self-driving cars can share their experience perfectly, so they improve at a rate of O(n). I get a day of driving experience per day. Each one of a million self-driving cars on the road will acquire a million days of driving experience. They won't drive traffic accidents down to zero in a week, but in a few years, if we were to deploy them en masse, they could.

[0] Can't actually find CDC data about this anymore. "Accidents" is definitely still #1, but they no longer break out traffic accidents. I know in the past, that was #1. https://www.cdc.gov/nchs/data/dvs/LCWK1_2015.pdf

Wouldn't the safety of self-driving cars relative to human drivers also become moot if cars were no longer needed?

We are a LOT further away from cars no longer being needed than we are full self-driving cars.

Who's "we"? It seems to me that not all parts of the world are equally dependent on cars.

This is an odd comment considering how humans have lived hundreds of thousands of years without cars, and only a few decades with them. You’re saying that a solution that’s as old as time - and used by billions of people today - is not as realistic as a technology that hasn’t been developed yet?

We've had wheeled chariots and carts powered by animals for almost as long as we've had civilization (about 5000 years). Cars and trucks are simply more efficient (a defensible statement... https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5102965/ particularly with electrification), self-propelled versions of these.

So absolutely we will have cars and trucks well after we're able to develop driverless versions of them.

> get to inherit the experience of the 70-year-old they're replacing

A lot of 70 year olds should not be driving.

You're misusing big O notation

It’s pretty clear that the author isn’t using the technical definition of big O but rather is using it in a descriptive way. It’s not ambiguous in any way so what’s the harm?

I would say no harm, but rather confusing to people that do not know big O, and useless to those who do.

Can you explain how it's wrong? Obviously, normally you want things to scale more slowly rather than faster, but other than applying to a different thing than usual, it seems accurate to me -- human drivers improve at a constant rate regardless of how many there are, and self-driving cars improve linearly with their number.

A function being in O(f(x)) means the function grows no more than a (multiplicative) constant factor faster than f(x). So O(1) is a subset of O(n). O is analogous to <=.

The semantics that GGP comment meant to convey are better represented by saying that autonomous cars' driving experience grows with Ω(n) (analogous to >=) or Ɵ(n) (analogous to =).

How so? There, n means "the number of [this type of] drivers", thus the GP is saying that "self driving cars improve in a way proportional to the number of self driving cars [that are collecting and sharing data], while humans improve their driving skills in a way independent of the number of humans driving."

(I think that's slightly wrong -- humans are probably O(log n) because large populations of humans invest in heuristics and teaching methods that enhance driver improvement rate, and for SDCs, the "lessons learned" between units are probably somewhat redundant, so it might scale as something like sqrt(n). But I don't see a problem with the O(n) usage.)

Just yesterday while driving to the gym, I witnessed the person behind me at a red light, violently gas it and swerve out into oncoming traffic lanes, drive past all the cars waiting for the light, and blow straight through the red light.

He went on to blow another red light before turning off onto a side street.

This kind of shit would never happen with a self driving car. The bar for self driving cars is incredibly low.

Well, that kind of stuff would still happen since in that case the driver was deliberately driving recklessly.

In fact, people could start accounting on other people's self driving cars being cautions and deliberately betting on their safety features. e.g If most people were driving a Tesla I could more safety run red lights since their other cars will stop for me automatically.

Unless we are going to make self-driving ONLY cars that is.

> Unless we are going to make self-driving ONLY cars that is.

This is the dream.

EDIT: More specifically, I envision a world where to manually drive your car you have to past a MUCH more rigorous test than the current licensing process. Including background checks for infractions such as DUI, would ban / block you from obtaining a manual driving license.

You would of course retain your freedom to have a car that can drive itself.

Maybe your bar. A reasonable (minimal) bar is that they have to perform at the 50th percentile of drivers by safety. This guy sounds like the first percentile.

Depends what the curve is, doesn't it? If the worst 1% of drivers cause 90% of the accidents, then replacing all drivers with vehicles at the 2nd percentile would drop the accident rate.

> A 16-year-old doesn't get to inherit the experience of the 70-year-old they're replacing. A new self-driving car does inherit the experience of older self-driving cars. They also learn from each other in a way humans can't.

> Human drivers also improve at a rate of O(1). Self-driving cars can share their experience perfectly, so they improve at a rate of O(n). I get a day of driving experience per day. Each one of a million self-driving cars on the road will acquire a million days of driving experience. They won't drive traffic accidents down to zero in a week, but in a few years, if we were to deploy them en masse, they could.

This is all very nice, but if true, why are self-driving cars still unable to drive in snow, avoid crashing into trucks pulled over on the side of the road, or navigate a parking lot without nearly getting t-boned?

The rate of improvement might be O(N) (I strongly doubt it), but the problem may lie in the constant multiplier...

> why are self-driving cars still unable to drive in snow

Have you seen people driving on the first snowfall of the year?

> or navigate a parking lot without nearly getting t-boned?

Driving in a parking lot is tough. Lots of pedestrians, lots of other cars, and way more people driving in reverse than you'd see in most scenarios.

> Have you seen people driving on the first snowfall of the year?

Most of those people don't have winter tires.

> Driving in a parking lot is tough.

So when is the O(N) rate of growth going to get SDVs to the point of an attentive 16 year old?

> Most of those people don't have winter tires.

You'd think that they'd drive extra cautiously in snow in that case.

> So when is the O(N) rate of growth going to get SDVs to the point of an attentive 16 year old?

Does it need to beat an attentive 16 year old or does it just need to beat a distracted early to mid 30s parent with a small child or two in the car? Self driving cars only need to be better than the average human to reduce overall accidents. Having said that, I'd probably be more likely to hop into a car driven by a random person than a self driving car that I know is only marginally better than an average human.

> I believe the number one cause of death for Americans under the age of 30 is traffic accidents.

And it's possible that the number one cause of shortened lifespan for older Americans is also cars, by virtue of encouraging (perhaps enforcing, if they've sufficiently influenced your town's urban planning) a sedentary lifestyle, and all the health risks associated with one. I also wouldn't be surprised if they contribute to the USA's relatively high rate of mental health issues, by contributing to social isolation.

Self-driving cars won't help with those bits. If anything, they'll intensify them.

> Self-driving cars won't help with those bits. If anything, they'll intensify them.

One common reason people give for not biking is that they're afraid someone driving a car will kill them. Self-driving cars should reduce that risk.

Of course, they also should make commuting by car cheaper and more relaxing/less of a waste of time, so you're likely correct, but there is that competing force that may increase cycling rates.

> Self-driving cars can share their experience perfectly,

I'm not a machine learning expert, but I'm not convinced this assertion is well grounded. If you have two ML systems, I don't know of a simple, general, reliable way to transfer the "knowledge" stored in one to the other without disrupting the "knowledge" already within the other.

If you mean just copying the video feed, sure, but that's only one input variable. There's no guarantee of a linear relationship between hours of video feed and driving quality. Humans seem to hit an asymptotic lower bound in the collision probability even as they accumulate experience, and while that might not be true for self-driving cars, it isn't guaranteed a priori.

You're right, it's not perfect, and it might be sub-linear improvement, but it can practice on its predecessors' input in a way a human can't. (There are plenty of scifi stories where they found a way to do this for humans as well, so it's certainly conceivable, but seems a long way off, whereas "feed all existing training data into the new model" is something we have now.)

You're pretty much guaranteed that self-driving technology "skills" will collectively improve over time, at some positive rate, whereas we don't expect "Gen Z" to be more skilled drivers than previous generations.

This is all entirely skill-focused. We're not guaranteed that self-driving cars will actually get safer if we don't have the proper incentives to favor safety over speed or cost. Similrly, despite the stagnancy of human driving skill, I anticipate safety will improve with human drivers for non-skill-related reasons -- specifically, driver assistance technology, and a shift towards just getting a Lyft if you're impaired.

One (probably not so simple) way could be to build up a shared database of input-decision-outcome records, so when a car is faced with some unknown object on a particular street, or its own risk assessment is high (e.g. heuristics- "I am planning to decelerate quickly"), it could search for similar input patterns (e.g. color, shape, gait) or similar decision plans in the database, and bias the car's own decision accordingly to avoid an undesirable outcome. Then that decision would get posted to the database as well, linked to its parent. Previous outcomes in the database could be assessed by humans in some kind of legal framework to help the cars determine undesirable vs desirable outcomes.

> I believe the number one cause of death for Americans under the age of 30 is traffic accidents.

Traffic accidents are not included in the "accidental death" category, and haven't been for as long as I've been looking. From page 2 of the full report[0]:

"For example, Malignant neoplasms of the trachea, bronchus and lung (lung cancer) and Motor vehicle accidents are not rankable causes of death (see “Procedures for ranking causes of death”), although they can be identified using the standard mortality tabulation lists. If these causes were included in the current rankings, lung cancer would be placed among the 10 leading causes of death with a rank of 6th, whereas Motor vehicle accidents would rank 13th."

It's not even close to being #1, for any age group. More people die from suicide every year than from car accidents. Please also keep in mind that 1 out of every 6 traffic fatalities in the United States was a pedestrian.

[0]: https://www.cdc.gov/nchs/data/nvsr/nvsr68/nvsr68_06-508.pdf

Good catch. However, that paragraph goes on to say "However, each of these is incorporated into broader rankable categories, namely, Malignant neoplasms and Accidents (unintentional injuries), respectively." So they are included in accidental deaths -- they're just not broken out.

Being #13 overall definitely doesn't preclude them from being #1 in younger age brackets. Most of the deaths from cancer and heart disease are the older age brackets ("Diseases of the heart" are 23% of overall deaths, but only 2-4% for each age rank among minors; ~6% in the 20s-30s). "Accidents (unintentional injuries)" is the #1 listed cause for every male and female age bracket until the 45–54 bracket (pp 19-22) (this precludes infants under 1). The question is what portion of those were traffic accidents.

If we assume traffic accidents at the same rate among all age groups, then we have 36,000/169,936[0] = ~20%. That drops it to around fifth place for most of the younger brackets. (Personally, I'd guess it's higher for at least the 15-19 age bracket, but that's just a guess at this point.)

So, okay, fine, traffic accidents is roughly sixth most common cause of death for kids in the US, and thirteenth for all ages. The original question was why we'd prefer that if the cost is the same. Even while my numbers were way off, it's still a big benefit.

> Please also keep in mind that 1 out of every 6 traffic fatalities in the United States was a pedestrian.

I'm assuming a significant portion of pedestrians traffic fatalities are drivers killing pedestrian, not pedestrians...I don't even know what to put as the alternative. Stepping in front of a bicycle?

My position is not that human-driven cars kill human drivers -- it's that human-driven cars kill lots of people. If drivers only ended up harming themselves, I'd be slightly less concerned about it.

[0] Just grabbing most recent number from https://en.wikipedia.org/wiki/Motor_vehicle_fatality_rate_in...

Let me guess. Single male, able body, no kids. When you have kids, you'd appreciate having a car/rent a car/uber when grocery shopping for example. And the city is full of people with kids. We are not the minority here. Not everyone is interested in a scooter that is inherently unsafe at moderate speeds.

original comment said "why would I prefer a driverless taxi if an electric car with a human driver costs no more?" Which would not seem to be talking about a scooter.

Also the original comment ignores the people that would pay considerably more for a driver less taxi, just to avoid interacting with a human driver.

> Single male, able body, no kids

What does being male have to do with it?

Married with 3 kids. I think the multiple errors in your assumption is due to poor thinking on at least two counts.

I live in the city, as my question may have suggested, and my family can cycle and walk and take the tram and bus and train everywhere we need to go in the city.

Also, we do have a vehicle, it's a campervan. We don't need it to do the shopping because we're walking distance to some small stores and get deliveries of groceries and fresh, organic vegetables (yum).

The reference to the safety of scooters may require you to provide some evidence. The children are required by myself to wear a helmet and their health is better.

We'd be better off building our cities to encourage mass transit. Micromobility follows as a bonus.

Good luck redesigning US cities and suburbs.

> Consider cities... like London and New York. (snip) Note that I'm only asking about city travel

Generally speaking, I don't see any future where people no longer drive, because the cost of not-driving is so exorbitantly expensive that very few people will ever be able to afford it. EVs are cheap, and get cheaper every day. Dense urban "car-free" housing is already unaffordable for almost everyone, and only gets more expensive every day as development continues.

Unless we get extremely lucky and the housing market crashes hard, the future will probably be one where it's cheaper to have buy 6 to 12 brand-new zero-emission EVs, than it will be to afford an urban car-free 2-bedroom condo in any top-20 city. Given Urbanists political goals (and the fact that Urbanists are winning their goals), I suspect public car commuting and VMT to grow every single year, for 80% of the low and/or middle class population.

High density being expensive and getting those rents shows there is demand: we should be building more of it to fill demand, and in turn bring prices down.

In less dense areas the car shines above anything else. However in those dense areas where people have shown they want to live cars are too big and we need something else.

> we should be building more of it to fill demand, and in turn bring prices down.

That's not really how construction works in the US. All construction generates higher prices than existed before (this is explicit intentional policy, at both the local, state, and federal level).

There is no such thing as new housing in the US that lowers property values, therefore there is no such thing as a new building that lowers "prices". No one does this (except perhaps, non-profits and such). And even if you wanted to anyway, no bank would ever finance it.

That is correct only to the first order. New construction is always higher priced and luxury. However the new construction in turns pulls someone out of some older place elsewhere which then goes down a little in value and becomes more affordable.

The best time to build affordable housing is 20 years ago, the second best time is now.

Of course increasing supply lowers the equilibrium price, relative to the case where supply doesn't increase.

https://research.upjohn.org/cgi/viewcontent.cgi?article=1325...

> This suggests that new construction reduces demand and loosens the housing market in lowand middle-income areas, even in the short run.

So I have to take my 4 year old who developed a high fever to the emergency room on an electric bike/ scooter ?

Sure, why not. If it is a really life or death emergency you of course need to call an ambulance which I'll assume has license to do things illegal for you to get there faster. Most ER visits are important to soon, but few are now important.

By your methodology everyone visiting an emergency room on a typical night should call ambulance because the cars have been banned in the city ? I know for sure as heck they can walk or ride a fancy electric bike.

No, most people don't need an ambulance because they can walk in to the ER. That doesn't mean they shouldn't be in the ER, just that an extra hour walking there isn't going to matter. (in hindsight 80% shouldn't be there, but the symptoms of something that needs to be handled now and a trivial ache are generally the same so you go just in case which inflates the numbers)

You'll take her by the fastest method appropriate. No one, said have to, that was just you.

It doesn't take much to imagine a situation where the nearest emergency room is a mile away and she can still go on the bike's child seat. Or a taxi.

If you live in a city your car, even if you have one, is not necessarily the quickest way to get around.

[Speaking about the US, specifically] Micromobility in major cities doesn't really change the realities for the vast number of Americans who don't live in a place where this is realistic.

I grew up outside Palm Springs (so not just talking about the very rural part of the country, which is vast and still populated). It would have taken me almost an hour to bike to my high school from my house, but only took ~20 minutes in a car, and for two of the years, the route included dropping off two younger siblings. Skipping past the safety concerns, I think that two daily hour long bike rides is pretty unrealistic for kids under a certain age. This also ignores that, in high school, I literally didn't have a spare 1.5 hours with varsity sports, AP/IB classes, and all the other "necessary" extra-curricular stuff for college prep (which is a separate issue, but still exists).

"No longer driving at all" is really only for single and/or childless people, and mainly those living in a large city where the retailers and service providers you need to see in person are within a realistic biking range.

I'm not sure he meant "no longer drive" literally or if he meant that the number of drivers has curtailed enough that the problems autonomous vehicles were created to solve are no longer problems.

As you described, clearly there are good cases for driving. Cities are not one of them.

Maybe we'll call them something besides "cars" in the future, but motorized air conditioned boxes that take us directly from point to point are not going away anytime soon. I'm all in favor of more transportation options and especially, since I'm a cyclist, segregated bike lanes. Heck, when I lived in Boston, I loved being able to ride just about anywhere in the city. On a bicycle, cars were the bane of my existence. I'd love car free streets!

But I'm also a young human who's in pretty good shape. My dad, who uses a walker and has other mobility issues, can't bike and can't walk distances. Buses and trains are options, but he'd have to use them for any distance travel + a wheelchair if he has to go more than a block... not easy.

When I was in Boston, I had the privilege of living in an expensive apartment in the middle of town. I could walk, bike, take a bus, etc. But if you can't afford to live downtown, you're once again restricted to buses and trains.

Of course, we should build up complementary transportation systems but a car-free future isn't going to happen and we shouldn't want it to happen. For many people and use cases, cars can't be beat.

Self-driving got taken over by the "fake it til you make it" crowd. Uber, Cruise (go find Cruise's original demo video and marketing claims), Tesla, etc. Only Google/Waymo approached it as a hard problem, and only Waymo has something that's deployed, even in a limited way, without a "safety driver".

Somebody will probably make it work, but it may come from some boring old company like Ford or Continental or Fujitsu. A background in webcrap has negative value in this area, as it does in avionics.

Einride got driverless trucks running on a public road:

https://insideevs.com/news/350293/autonomous-einride-t-pod-p...

And Las Vegas had driverless mini-buses running around downtown for about a year. Not anymore because funding ran out.

https://lasvegassun.com/news/2018/apr/05/how-does-downtowns-...

"a short distance on a public road within an industrial area – between a warehouse and a terminal – where traffic speeds are typically low".

I honestly wouldn't be surprised if these 'Auto-Pilot' cars are actually being controlled remotely from some off-shore ops center.

In fact at least one company is looking into this, having the car self-navigate as much as possible, but when confronted with a difficult situation like construction or debris in the road it would call back to a remote operator who would take over until you are past the difficult situation. They also saved the operators response so when the next car hits the same obstacle and calls back the next operator can just repeat the same solution.

I thought it was a good idea, but somewhat more expensive since you have to keep a big call center running and it depends on having reliable and fast internet connectivity at all times. The alternative is what everybody else is doing where you have to have a full set of controls and a licensed driver behind the wheel ready to take over when the AI runs into a situation it can't handle.

interesting, but what a terrifying job that would be. dealing with life-or-death scenarios all day long. and the liability questions would become even more complex

I think it really becomes a question of what to do when the computer senses it is getting into a scenario that it can not handle with a reasonable confidence %.

You wouldn't flash hand-off the car from auto-drive to the operator without some basic safety measures also occurring. One of those will be slowing (or stopping if possible) the car until the remote operator takes over. You won't hand the human "the bag" (responsibility) after the computer gets you totally up shit creek and you have no chance to effect a positive outcome. It will hand over "the bag" when the "we are getting up shit creek incrementally" value crosses some accepted safe threshold.

(Didn't mean to ramble.)

Yes, the AI did come to a stop before ringing up the support desk for assistance. The example video showed a construction site where the entire road was blocked off and the way through was to drive on the shoulder past the construction. However, in the demo the road was completely empty otherwise, which I didn't think was terribly realistic. The memory feature in particular was only useful as a convenience to the next operator, it didn't help the AI become any more self-sufficient.

I don't know enough about ML to comment with any authority, but I think this approach makes sense. Your call-center ops people would provide real-time QA (i.e. make sure the machines aren't killing people) as well as provide training data for your model. If this were in fact the approach, I'd be much more receptive to the promise of 'Autonomous Vehicles'. I for one, refuse to believe the hype that real-time distributed computer vision applications are 'nearly solved'. Consider the amount of variables (pedestrians, pot-holes, un-marked lanes, obstructed signage, other drivers etc etc) at play here...

Similar to how military drones are already operated.

You don't even need the autonomous driving part at all. Just being able to remote drive a truck would disrupt the industry, as you can pay your drivers less and get better results giving them six hour shifts in front of a monitor and letting them go home to their families when it is over. Then you can lower the headcount by slaving trucks so one driver is handling a train of 5 or 6 on a highway. All doable now with existing technology, with the extra risk being how reliable your comms are.

The article misses an important point. It’s not only a problem of technology, it’s also a problem of risk acceptance. While we accept to be killed by our own mistakes at the wheel, society (we) will not accept that a machine makes the same, even if the failure rate is on average /better/. This is the fundamental issue Elon Musk has. The driverless bar for safety is way higher than very clever driver assistance or even rocket science.

>While we accept to be killed by our own mistakes at the wheel, society (we) will not accept that a machine makes the same

Ideally we should not be accepting either. It is outrageous how many people we allow to be killed regularly for slightly more comfortable transport.

Some number of deaths will always occur due to the tradeoffs we make in transportation. We can minimize this number, but even "no cars, no bikes, no trains, no planes, no going outside" would cause deaths due to the lack of things like ambulance or fire services. (and of course if you allow those, someone small number of people will die in crashes)

I think there's an implicit bargain made when you get in a car that a person is driving: you're both risking your lives on this journey. A self-driving car is risking possibly millions of dollars and bad publicity, but not a life. It's hard to trust a computer like that.

Which is why the recennt spike in pedestrian and bicycle deaths is even more alarming.

Ultimately, you dont have to trust a computer. You have to trust the corporation that built that computer. That is quite different IMO. How am I supposed to entrust my life to a machine built by a corporation that just recently dropped the slogan "we wont do evil"? How am I going to entrust my life to a machine that has been built by a company that, like all the others, is maily interested in maximising profit margins? Why should I believe they care about my health as an individual? As you said, if I am passenger of a human driver, I can at least hope they dont want to die. If the machine crashes itself, who cares?

You trust corporations all the time to make products, construct buildings, etc. to not kill you when used more or less correctly in accordance with what most people would consider common sense because said companies know they will be sued, fined, etc. if they don’t make safe things. It’s actually unclear what this means in the context of autonomous vehicles that will presumably have some non zero accident rate even if safe enough at some level.

> slightly more comfortable transport.

Slightly more comfortable than what? Horses? Bicycles? Walking?

There were 158,493 registered deaths in Australia in 2018, [1] of which 1137 were due to motor vehicle accidents [1].

Motor vehicle related deaths don't even rate in the top 20 causes of death. [also 1]

The rate of motor vehicle related deaths in Australia peaked in 1970 at 30.4 per 100,000 population per year, and has been on a steady decline since and now sits at 4.6 per 100,000 per year.[2]

This seems reasonable to me.

1. https://www.abs.gov.au/AUSSTATS/abs@.nsf/0/47E19CA15036B04BC...

2. https://en.wikipedia.org/wiki/List_of_motor_vehicle_deaths_i...

How about the deaths not related to biology/age of your body?

It should really be compared with other accidents. Cyclist deaths, bathtub drownings, slipping on banana peels. Choking on food perhaps. Have the feeling it would rank pretty high in its cohort.

> Motor vehicle related deaths don't even rate in the top 20 causes of death.

According to [1] in the US road traffic accidents are the 17th-most-common cause of death across all age ranges.

But among 15-24 year olds? Number 1 cause of death. In fact you have to reach age 45 before road traffic accidents are even outside the top five.

If you really want to ramp up the rhetoric, compare road deaths to mass shootings and September 11th. Where's the trillion-dollar invasion of Detroit to overthrow the Drivetrain Of Evil?

[1] https://www.worldlifeexpectancy.com/usa-cause-of-death-by-ag...

Right. And you are just looking at deaths. The injury rates are much higher. And crash injuries can be debilitating. Certainly the #1 threat to health in the U.S.

And about the other causes of death, I want to see how driving correlates with obesity and heart disease. Then throw in depression and mental anguish that's endemic in the rush hour commute. Sitting for 3 hours in a car versus walking a little more and taking the bus sounds like a really bad trade for your health.

> compare road deaths to mass shootings

Well, if we are going to do that, we need to compare the utility of those mass shootings compared to the utility of a car.

Ok, lets not forget the 40,000 seriously injured or disabled every year in Australia.

> Slightly more comfortable than what? Horses? Bicycles? Walking?

Buses, trains, trams, scooters, many other ways.

71% of Aussies live in major cities, and most have no need for a car. Yet we are one of the most car addicted people in the world. Billions wasted every year on this. Ridiculous.

It is a shame that while the the first link has a table of Years of Potential Life Lost, it only counts things that made it into the Top 20 (so it just rearranges things - obviously suicide shoots to the #1 and dementia falls from #2 to way down the list). Road deaths kill on average younger people than most of the items on that Top 20 list.

Cars are way more comfortable and more convenient than public transit. Working Parisians who take transit spend 113 minutes a day commuting: https://www.thelocal.fr/20160418/parisians-spend-23-days-a-y...

DC, where most people drive, averages about 70 minutes.

Commuters in Dallas, where everyone drives, spend just 54 minutes a day commuting: https://www.chron.com/news/houston-texas/houston/article/Hou...

your implied conclusion is that its better to drive doesn't stand up.

You _can't_ drive to central paris. London you're looking at a £15 charge and a time penalty of ~15-45 minutes.

Public transport to and from work can be pretty awesome. There’s a lot of stuff you can do, that you wouldn’t be able to do while driving.

Maybe you can, but I get car sick when I try to do anything on the bus. I also spend twice as much time on the bus vs driving - all wasted time because either way I'm staring at the road out the window.

Most jobs aren’t in central Paris.

Not sure why you quote average commute times of random cities and what that has to do with comfort of cars v. public transit.

But personally, I prefer to take the subway and read and relax for half an hour before and after work, instead of having to drive in rush hour.

The subway where? Nowhere dense enough to have good subway service is a place where being in a subway car during rush hour is relaxing. I used to take the 1/2/3 in New York to work. It was hellacious. The only thing worse was the 4/5/6.

Speaking from personal experience here, my commute time tolerance is roughly double if I'm commuting by transit versus driving. In other words, a 60 minute trip by transit is roughly as painful as a 30 minute drive. Other people I've talked to have also expressed similar feelings.

I've found pretty much the opposite. People don't consider public transit unless its clearly the better option.

Agreed. I'm not some sort of high powered investment banker who measures minutes wasted in thousands of dollars. I'm just a normal guy, trying to take it easy. I'm not in any huge rush to get anywhere if it's not a hospital, and I'd rather "waste" a few minutes getting driven around town by somebody else in somebody else's vehicle than assume the responsibility and risk of driving myself around.

Sometimes transit or walking simply isn't an option, and in those cases I drive. But given the choice, transit might take a bit longer but it's much lower stress. If a kid jumps out in front of the bus and gets run over, that would ruin my day but not my life. If the same happens when I'm driving a car, that's life altering.

There are lower-stake considerations too: the money I spend on bus fares would doubtlessly otherwise be spend on parking; or I would need to schedule enough time to hunt for free parking which just adds more uncertainty and risk to my day.

Ideally, I would like a pony. Realistically there is no alternative option, so there is no question.

> for slightly more comfortable transport

In some large metro areas, maybe. Outside of that bubble millions of people depend on a car to even be able to work.

Large metro areas effectively subsidize the rural areas. Those millions of people have enjoyed amazingly subsidized civil engineering to have that car commute. If those commutes built-in the correct costs in land usage, gov budget impact, and environmental impact, those commutes would be untenable.

You missed the phrases "flyover states" and insulting people as rednecks and hillbillies.

You're accusing me of a bias that does not exist. If you want to live in the country with a different taste of personal freedom, please be my guest. However, it comes with a cost that has been ignored for too long.

Personally, I view a city that is safe from cars as the greater freedom. It means my children can walk to school/friends/store without the very real danger of being killed. Cars are the number one cause of death for children.

Without populated rural areas to support them, people in cities would start cannibalizing each other within weeks. You seem to have a lot of disdain for people that make your urban lifestyle possible.

OP is referring to people who live in a rural area but commute to/from a city. Sounds like you are talking about farmers.

Such people are participants in the communities they live in. They pay taxes and patronize local businesses. Often the spouses of commuters work in the local community for those local businesses or the local government. The spouse of the city commuter he despises might well be the best English teacher at the local highschool the farmer sends his children too. The children of a commuter might be the friends of the children of a farmer, and often as teenagers will work for some of those farmers. Beyond economic participation, commuters and their families participate in local clubs and churches which benefit those farmers as much as anybody.

By singling out an economically and socially productive segment of rural communities to attack, he is in fact attacking those communities as a whole.

(Incidentally, it's not just a matter of food. Where do you think the concrete, asphalt and timber cities are made out of comes from? I've not seen many quarries or forests in major cities. Rural communities as a whole, not just farmers specifically, make urban life possible.)

Large metro areas also include an incredibly large proportion of suburban homes where people absolutely continue to rely on cars.

A suburban area that likely expanded because of 1950's highways that bored through the downtowns of cities, displacing thousands. Streetcar Suburbs, predating car oriented development, are way more walkable and not car dependent.

That may be true in many cases. But I distinctly remember some of the highways that were built in Portland, and they bulldozed existing suburban neighborhoods to do it. This city was basically one big suburb well before someone had the bright idea to build I-5 through downtown. Heck, almost all of the [residential] tall buildings in the downtown area were built well after the freeway was built.

We have an urban growth boundary now which tries to limit further sprawl, but it can't change the fact that the Portland metro area is 2 million people most of whom live in single-family homes, or maybe low density suburban apartment complexes.

I think the fundamental issue Musk has is that his cars can’t drive safer than even a below average driver.

Oh wow, it can stay in a lane and match speed! So can every car manufacturer. They’ve got a level 2+ system and cars won’t be safer until level 4.

Do you have data that supports what you think?

https://www.wired.com/story/tesla-autopilot-why-crash-radar/

That article describes a failure of Tesla’s system that had terrible consequences. It gives no data comparing with human performance.

That article describes a systematic design defect of Tesla’s system. (It’s a defect that all camera based systems have. They are designed to ignore stationary objects because they can’t adequately distinguish background objects from obstacles.) Car and Driver tested this last year, and the Tesla ran into the stationary dummy nearly every time. This is something humans have almost no problem dealing with.

Assuming what you say is true, is the Tesla system worse than a human driver overall in its operating domain?

Humans have weaknesses too. What do the data say on their relative performance? I’m not taking a position here, just asking for data.

Look, "autopilot" works on highways, where there is a median. Everything goes the same direction, and there is a minimal amount of clutter.

Its about as easy as driving gets. This is why a level 2 "autopilot" can parade around the place as something that is both innovate and safe.

None of the data compares what happens at junctions, which is where the majority of crashes happen.

It cannot work in suburban sprawl, It can't cope with poor weather, it can only drive on highways.

Not sure why you state this. I've used autopilot in a Model 3 on city streets and suburbs, it works pretty great.

Similarly, I have used autopilot in rain so bad that I was scared to drive. It performed perfectly and was much less stressful for me.

I was fully engaged and ready to take over, but never needed to.

I'll echo that. Heavy rain, I could barely see the lines, but the car seemed to find them just fine and was driving straight and smooth.

Until traffic on the other side of the median barrier hit a big deep puddle, absolutely covering my car with a thick sheet of water. Autopilot immediately started screaming at me to take control. I already had my hands on the wheel, but it still scared me because I couldn't see either! After the water washed a way in a couple seconds, I was able to re-enable AP and keep going.

It has been a problem, but it is definitely not an intrinsic defect of camera based solutions. The stopped truck example has been examined in depth with rooted Tesla's. The Tesla actually detects the truck, but fails to fully recognize the scenario and attempts to drive under it! See also, greentheonly's work on Twitter on researching this.

Ignoring some stopped objects is only a temporary limitation.

I said it was a systematic defect camera based systems share, not that it was intrinsic to camera based systems. Camera based systems have much less information about object positioning than LIDAR based systems. They filter out non-moving objects because the vision processing algorithms are not sophisticated enough to distinguish between the background and obstacles. If they stopped for stationary objects in the road, they’d routinely stop because they mistook the background for a stationary object.

That may or may not be a temporary limitation. It depends on whether vision processing becomes reliable enough where you can almost always distinguish an object in the road from an object in the background.

> I said it was a systematic defect camera based systems share, not that it was intrinsic to camera based systems.

And I'm saying that even that is incorrect. Here is a Tesla failing to stop: https://twitter.com/greentheonly/status/1134987596508143617 - The driver clearly noted that aiming for a less clear spot (dangling wires) triggered a stop.

Here is another example, in which greentheonly has used root to overlay what the car sees: https://twitter.com/greentheonly/status/1135011184439169024 - It sees the truck but is treating some of the space under it as drivable space.

Another view, with more details on what the car considers drivable vs non-drivable: https://twitter.com/greentheonly/status/1134490014321131521

Also, a clear case where it did stop: https://twitter.com/greentheonly/status/1134489977801334784

The Tesla system has limitations around object identification, but absolutely does NOT just filter out non-moving objects.

Sure, https://en.wikipedia.org/wiki/List_of_self-driving_car_fatal...

That link does not provide data either way. In fact it explicitly declares its inability to do so:

“Cars driven under traditional human control are currently involved in approximately 1.18 fatalities for every 100,000,000 mi (160,000,000 km) driven.[3] According to many automotive safety experts, much more data is yet required before any such clear and demonstrably higher levels of safety can be convincingly provided”

That's not how this works. If the cars are safer, there should be independent data that shows this. As it is, we get one-sentence "safety reports" from the manufacturer with a single data point. They're not safer by default because the manufacturer says so.

The claim was that Tesla’s system is not safer, and that only level 4 would increase safety. I asked if data supported those two independent claims.

I’m ok if we assume that things are unsafe before we have evidence they are safe, but that’s not the same as declaring they are unsafe without evidence.

The current situation is somewhere between the two: There have been some crashes (e.g. Boca Raton) that strongly suggest there is a dangerous systemic problem.

In such cases, demanding statistical confirmation that it is a significant problem (as NASA did with the O-ring erosion problem) is not the right response. It reminds me of a case in the building of the 688 class submarines: a cracked weld, that had supposedly passed multiple inspections, was found in part of the torpedo-handling equipment. At that point, all welds, including pressure-hull welds, became suspect.

https://www.goodreads.com/book/show/31730.Running_Critical

Unless you are Boeing.

The quarterly safety numbers that Tesla releases are difficult to draw conclusions from, because they do not represent a controlled experiment over the same road conditions. However here are the numbers for millions of miles per crash (higher is better).

With AutoPilot, Without AutoPilot but With Active Safety Measures, Without AutoPilot or Active Safety Measures, and NHTSA Average Vehicle

  19Q3 - 4.34 2.70 1.82 0.50
  19Q2 - 3.27 2.19 1.41 0.50
  19Q1 - 2.87 1.76 1.26 0.44
  18Q4 - 2.91 1.58 1.25 0.44
  18Q3 - 3.34 1.92 2.02 0.49

Those numbers are useless for concluding anything about what would happen if you remove the human completely, there are videos that show how the human intervened and saved the car from certain crash and I this situation should be counted as "Deaths" for the AI only column.

Driving assistants would improve safety certainly, but removing the driver and letting the assistant drive will need real world like tests, for now maybe you could have an independent group monitoring a number of Teslas and checking each time a human intervened and give a score on how bad the AI was.

Actually I’m not convinced assistants will improve safety.

Level 4 for example could be a blood bath, where cars drive all the easy situations, and hand over emergency control to out of practice drivers barely paying attention at a moments notice.

I believe our driving skills were better when we had manual transmission, and were more involved in overall operation. Hard to determine if safer, b/c of advances in crumple zones, airbags, etc as manual declined.

The sudden control swap you describe is level 3.

Level 4 requires that the vehicle safely stop in situations it can't handle.

It is possible , we would need data,

The extremely sad part is that we could reduce the number of deaths until true AI arrives using simple tech that could detect if the driver is drunk or tired, and controversially forcing the speed limits(the car knows that this road is max 90Km/h and that the driver has less then 1 year experience so it will not let you get over 70 (local laws apply) unless you activate emergency mode - this would trigger some bad consequences if it was no emergency)

We do have some data. E.g., the Volvo XC90 hasn't killed any passengers, ever, thanks to collision detection and emergency braking systems: https://www.telegraph.co.uk/news/2018/04/15/britains-safest-...

Good to see improvement, but is this controlled for the fact that autopilot is mostly on highways? I'm not sure what the crash rate per mile is on the highway vs driving in a city.

I would bet a body part on it being way lower, especially if we do not limit it to fatal crashes only (I read that statistic quoted above as "all reported crashes", not just those with fatalities).

I had two crashes in my lifetime. Both occurred when driving backwards out of a parking lot. No fatalities, limited damage. Good luck finding such situations while driving on a highway...

I don't know about per-mile, but there are more crashes in the city per hour.

(that's data from the highest crash rate country where you were able see so many crashes that you develop a feel)

I use autopilot even on small roads even if it's not designed for it. After V10 it also handles roads without a centre line just fine.

However I do pay attention and keep my hands on the wheel, I think autopilot + human is much safer than either one, but since the latest updates I rarely have to take control if not going in roundabouts or junctions.

I also had one instance a few weeks ago where I tried to enter another lane manually and the car took control from me and put me back in my lane as there was someone in my blind spot without their headlights on so I didn't see it.

That sounds promising. OTOH, my car tried to drive into a tractor-trailer a couple weeks ago. That was a momentarily interesting experience. It seemed to become confused by the onramp merge joining my travel lane (pretty normal for autopilot to screw this up), but then it "realized" the mistake and dove for the other lane. I haven't been as happy with v10 as I was with v9.

Yes, I know especially model 3 owners are less happy with V10 than V9 in this regards. I've had the opposite experience with the X.

The OP is simply stating the base case: cars without drivers shouldn't be safer than cars with drivers. The industry claims that their cars are safer without drivers because AI. The industry must provide the evidence for its very strong claim.

The fact that a majority of the public think it's the opposite way around, that skeptical opinions about self-driving cars must provide evidence, is the industry's succes- the triumph of marketing over reason.

Of course the industry has no such data. Research has shown that this data would take many decades to obtain:

https://www.rand.org/pubs/research_reports/RR1478.html

Personally the only justification I've heard for self-driving cars being safer is "computers are faster than people". Yes- but at what?

> Personally the only justification I've heard for self-driving cars being safer is "computers are faster than people". Yes- but at what?

Faster at making the right decision - and the wrong decision too.

Computers can "watch" every direction at once.

But can they remember that they just saw something that is not visible any more because it went behind another object?

They'll also happily be fooled by adversarial patches (optical illusions) in ways that wouldn't fool humans. https://www.vox.com/future-perfect/2019/4/8/18297410/ai-tesl...

And humans are happily fooled by conditions (optical illusions, rain) in ways that wouldn't fool machines.

https://xkcd.com/1958/

If you want to murder people on the roadway, there are lots of ways to do this that are easier than fancy adversarial attacks (which only fool one sensor type, at best... and are then patched).

Computers don't get board and start daydreaming (when I win the lottery where will the pool go in my mansion). Computers don't get distracted reading (ie facebook) while driving. Computers can have redundancy so the equivalent of a heart attack won't be endanger everybody (this assumes of course the redundancy works - computers have a lot of failure modes of their own)

There is plenty of opportunity for computers to be better than humans. It is a hard problem though.

OTOH, humans can tell that a merging lane is a merging lane. Hard problem indeed :).

Today that is a factor. When (if...) computers figure that out though they never forget to look to see if the lane is merging.

At the risk of showing just how ignorant I am of current state-of-the-art AI, it feels like the biggest block with self-driving cars is how short their memory is and how little high-level logic is implemented. The computer doesn't feel like it's driving the car as an activity so much as simply responding to inputs as they come into sensor range, sometimes correlated, frequently not.

Autopilot isn't even designed to be used without a human driver. Which confirms that even Tesla doesn't think its better than a human.

It can't even handle intersections.

>I think the fundamental issue Musk has is that his cars can’t drive safer than even a below average driver.

No one in the public has enough data to say one way or another whether that is true and by presenting it as fact you are falling into the same trap that OP is warning about.

Well it's certainly true outside the limited ODD of autopilot, that much seems undeniable. (?)

Drivers immediately seem to offload their responsibility into this capability.

Anecdotes:

"I might be too drunk to drive, but I'm sober enough for my Tesla to drive."

And every Tesla driver I see engaged in two-handed cell phone operation with no hands on the wheel.

> I think the fundamental issue Musk has is that his cars can’t drive safer than even a below average driver.

Agreed.

> Oh wow, it can stay in a lane and match speed! So can every car manufacturer. They’ve got a level 2+ system and cars won’t be safer until level 4.

I don't find this to be true at all. ProPilot is terrible compared to AutoPilot and many of the other manufacturers have the same limitations (or worse).

> ProPilot is terrible compared to AutoPilot

This is frightening. I strictly limit where I will let AutoPilot operate because it is not at all relaxing unless you're on the open road with light traffic. If this is the state of the art, FSD is decades away.

I'm curious... which version of AutoPilot do you have?

> ProPilot is terrible compared to AutoPilot

I second this - recently rented a fully-loaded Nissan Rouge and drove for around 700 highway miles - the lane centering feature was just awful compared to my Model 3's. It would often times give up on even slight curves, or fight me when I would try to regain control after it misjudged a curve. After a while I gave up on it and just used the automatic cruise control, which worked fairly well.

> The driverless bar for safety is way higher than very clever driver assistance or even rocket science.

I agree. This isn't about beating the national safety average, which includes drunk driving accidents, inclement weather, motorcycles, old cars and unwieldy trucks. It's about providing a safer alternative in scenarios where you have no other. And, it probably needs to be affordable too.

There is a lot of other low-hanging fruit to which we apply vision technology (like early cancer detection!), it's too bad investors are climbing over each other to throw money down the long, narrow corridor that is self driving.

> it's too bad investors are climbing over each other to throw money down the long, narrow corridor that is self driving

I don't consider that bad. Anyone is free to make bad investments. I do consider it bad that they haven't yet had to book that many losses, due to the hype still going strong and them finding enough other idiots to join the party. I'm longing for the day on which bad investment decisions (such as in autonomous vehicles, at least if you expect them to be just around the corner to save your business model) seriously wipe out huge investments again - as it should be.

> I don't consider that bad. Anyone is free to make bad investments.

You don't find it unfortunate when capital is allocated towards failing business models rather than tech that has real promise?

Not as long as this situation does not persist for too long. The market is supposed to correct such inefficiencies by dividing such failing investors from their money and conversely rewarding investors in successful and sustainable business models with that money.

The actual problem is not that investment errors are made. Such errors will always be made, as long as it's impossible to predict the future. The problem is that it appears as if right now, we are in a situation in which the negative consequences of such errors can seemingly be postponed virtually forever into the future. Or at least that has worked great for a decade now (first cracks are appearing lately, but it's far from enough yet).

The driverless bar for safety is way higher than very clever driver assistance or even rocket science.

One bad driver kills themselves and maybe a handful of others. One bad software update kills... everyone. So, the demand that the software be near-perfect is not at all unreasonable. It’s strange not to acknowledge this fundamental difference.

Problems like that have technical solutions. You phase in updates over time. A hundred vehicles a day for the first week after release, then a thousand a day, then ten thousand a day and so on. That way if there is a serious bug, you find it when it's running on hundreds of cars rather than hundreds of millions.

Just like 346 people flying Lion Air and Ethiopian Airlines became unwilling parts of a "technical solution". No thanks.

Unless you think we should stop having computers in planes, what are you proposing as a solution to that? More to the point, what does it have to do with systemic risk or reducing it by staggering updates?

The comment was criticizing your flippant attitude towards treating humans like beta testers, with the attitude that it's OK to kill a few, just not too many, and dismissal of goatinaboat's initial concern. I don't have a specific answer, because I agree that it's a problem far different than rolling out updates of a new browser version on everyone's laptop. "Technical solutions" acceptable in that scenario simply aren't good enough when the "OS" is speeding down the highway at 75 miles per hour.

Here was goatinaboat's initial concern:

> One bad software update kills... everyone.

That's a fair concern. If you push a software update out to a hundred million cars at the same time which causes some of them to crash into oncoming traffic at high speed, you could have a large number of fatalities by the time you identify that there is a problem with the update.

But phased deployments do mitigate that concern, because it provides time to notice the problem before the bad update is more widely deployed.

That doesn't imply "treating humans like beta testers" because systemic risk is a real concern even when you have thorough internal testing. Even a thousand test vehicles will not encounter the full set of conditions that hundreds of millions will. You would expect to see a flaw that occurs for one vehicle out of 2000 occur 0 times on a fleet of 1000 test vehicles but 50,000 times on an installed base of a hundred million. Having a total of 50,000 fatal traffic collisions in one day is totally unacceptable, but having 10 on a bad day would mark a significant improvement from the status quo.

It's one thing to say "mitigate," which you do here, acknowledging that significant risk still exists. That's very different than the wording of your first comment, "technical solutions." A solution implies the risk is negated.

Aside from the suggestion of beta testing on human lives, you are assuming that your phase-deployment mechanism always works correctly. Even Google has previously botched phased deployments and slow rollouts.

Or that a nice phased rollout over the summer doesn’t fail as soon as leaves fall or it rains

The thing about systemic risk is that it's systemic. You don't have to wait six months for leaves to fall because when it's spring in North America it's autumn in Australia. One place going without rain for months may be common but every place going without rain for months would be unprecedented.

> Aside from the suggestion of beta testing on human lives

There was no such suggestion. You can do thorough internal beta testing and still not uncover every bug that will surface in production, because encountering every possible scenario that will occur in a complex system following mass deployment is only possible following mass deployment, and it is still worth reducing the impact of those bugs.

> you are assuming that your phase-deployment mechanism always works correctly.

It can work correctly 30% of the time and still be 30% better than not having it, because then a serious problem has to coincide with a failure of the phased deployment in order to have widespread consequences. This is why safety-critical systems are layered, so that you don't need them all to work all of the time as long as having all of them fail at the same time is sufficiently rare.

I don't understand why so many otherwise smart people fell for FSD hype. I'm not talking about the many con artists - I'm talking about the tech-savvy people who genuinely bought into FSD hype.

How is it not obvious that FSD most likely requires artificial general intelligence and is, thus, probably somewhere between "many decades away" and "humans will never achieve it"?

If I had to have an opinion, this would be it.

I don't see why self-driving vehicles can't be dependant on additional self-driving-vehicle-related-signage or sensors embedded in / around / near roads, etc etc.

Humans are fairly dependant on additional signage, maps, and navigation tools (Google / Apple Maps).

I would say safety dictates that the car would still have to be safe when such external signals are either missing, wrong or flat out hostile. At that point I don't really see the value they provide.

Make it self driving on specifically marked highways and switch to manual on normal roads. That would free up hours for many people and might even improve traffic congestion.

And where is this going to come from? Our public infrastructure is crumbling after decades of insufficient spending and maintenance as it is. It will take a significant social shift to reverse this trend and stop robbing the future to support short-term political agendas of neoliberal austerity and privatization that merely enrich the rich further.

If we can accomplish such a shift, there are already much more worthy ends to put the money and political will towards: increased public transportation, rebuilding of streets to safely support non-vehicle traffic, public ev infrastructure, etc. Spending massive amounts to subsidize the few rich who could afford autonomous cars would just be further regressive policy.

It takes more than a tech-savvy person to understand the difference between a difficult problem and a Hard Problem. It might be a thought experiment to pose during a technical interview, though you might encounter highly optimistic people that are otherwise very talented engineers.

Doesn't seem obvious to me. FSD using affordable sensors seems like an unsolved problem, and I haven't heard a convincing proof that it is or isn't solvable.

Did high accuracy OCR without AGI seem possible before it was developed?

Yes, because it was clear from the beginning that it would be just a pattern matching exercise. No reasoning necessary, no actual, deep understanding of the written text.

It was always known that pattern matching would be what ML algorithms would be great at - since the 60s or do. Reasoning about and "understanding" things - well, not so much. Some people hoped that this behavior would at some point just "emerge". But they basically resorted to this because they did not see any promising path to these goals.

Where is the high accuracy OCR? All the OCR software I've seen so far is less accurate than me, especially with handwriting, damaged material, odd viewing angles, etc.

I agree and think a lot of smart people still don't really comprehend how non-linear the task is. In other words, they believe if a self-driving car handles itself properly in 95% of situations, the final 5% is right around the corner, when in fact the final 5% (often the difference between life and death) may, in all likelihood, take longer to get right than the previous 95%. It's just that much harder to be perfect than to be pretty good.

Furthermore, even 99.99999% reliability is not good enough, and that's already incredibly hard to achieve. Five 9s of reliability still means 5 minutes of unreliable behaviour per year, which is unacceptable in the context of driving.

Just being pedantic, but those are seven 9s, not five. For five 9s, the downtime is indeed 5 minutes per year, for seven 9s it's about 3 seconds.

Still an interesting question: is 3 seconds of erroneous/unreliable behaviour per year acceptable from an autonomous vehicle?

Even 5 minutes per year is totally acceptable. 5 minutes is 300,000 miliseconds. Split over the course of 365 days at 86400 seconds per day, we are looking at roughly 1 milisecond of erratic behaviour every 105 seconds. Not enough to crash a car

Perhaps that'd be OK, if the 5 minutes per year is evenly spread across the entire year (including time the vehicle isn't even being used?!)

OTOH, if it means that the vehicle is liable to have a 25-second "blackout" in the middle of your morning commute roughly once a month, it seems very different.

25 second blackouts are too much. However a 1 second blackout every 5 minutes might not be a problem in most cases: you can keep going the direction you are going and other cars that are not blacked out will avoid yours while the computer recovers. Of course this depends on the blackout - if the car is still steering itself but not looking for problems (a much easier problem) it works, if the car isn't steering itself it could run into a wall.

My practical experience with system failures is that they tend to not be evenly spread out in any kind of predictable pattern. Instead they tend to be acute and concentrated at a few points.

I agree that my assumption that failures will be evenly spread out is too optimistic. But on the other hand, if failures are concentrated on a few points, they should be that much easier to detect and correct.

You literally just proved theieindjri's point. (Assuming you're a smart person) these edge cases, even if they are clustered, even if they happen repeatedly, are NOT easy to fix. Partly because there are so many of them, partly because they only arise with a specific combination of events/circumstances that are difficult to replicate, partly because navigating them safely relies on facets of the human brain that AI doesn't perform, partly because the consequences of failure are so severe. "That much easier to detect and correct" vastly underestimates the difficulty of providing safe, automated driving.

I would be interested to meet a human driver who didn't have 3 seconds of erroneous/unreliable behaviour per year.

Depends what's happening. If those 3 seconds consist of randomly veering into something on a highway, that would be... Not good.

> Furthermore, even 99.99999% reliability is not good enough, and that's already incredibly hard to achieve. Five 9s of reliability still means 5 minutes of unreliable behaviour per year, which is unacceptable in the context of driving.

40,000 traffic accident deaths per year in the US.

< 400,000,000 drivers.

So more than 1 fatality per year per 10,000 drivers. There's a 99.99% change that you won't kill someone while driving in a given year.

That's just fatalities.

> In 2017, there were an estimated 2,775,608 police-reported motor vehicle crashes in which at least one passenger vehicle (i.e., a passenger car or a light truck) was towed from the crash scene in the United States, which resulted in an estimated 1,686,240 known passen-ger vehicle occupant injuries.[0]

So 0.6% per driver per year.

So far today I've seen one driver going the wrong way down a one-way road, and one run a red light. (Plenty entering the intersection on yellow and leaving on red, but at least one where it was blatant). I was on the road for maybe thirty minutes.

Do you think the average driver spends less than five minutes per year driving unsafely?

[0] https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/...

Ok, point taken, I take my previous comment back. I guess we just have to wait until the risk posed by autonomous driving systems becomes socially acceptable. Right now it just seems like a scary unpredictable risk.

It takes a while to internalize what's difficult about the problem. Partly I think it's that humans take for granted their ability to see, which requires no conscious effort but expends 70% of the brain's computational resources at any time. It's not clear AGI is required... computer vision is not known to definitely be AI-complete. It's a very difficult problem but could potentially be solved in the next several years.

> How is it not obvious that FSD most likely requires artificial general intelligence

Because it doesn't. The issue is that all the self-driving car companies ignore how people actually drive.

People memorize their local driving area--it's why teenagers are so incredibly dangerous and why if you put a new stop sign into an area you cause chaos for several months. A self-driving car needs to do the same--the advantage is that a self-driving car can download that "memorization" for any new area it goes into.

Between memorized zones, people mostly "keep station"--they get onto a freeway and keep a relative position with respect to the lanes and the other cars on the road until they get back off.

When something goes very wrong, people panic and random chance takes over the fate of the vast majority of people.

None of this represents general or artificial intelligence. However, solving these tasks requires a lot of very grubby, non-scalable work over a lot of grubby data.

Except that when people travel away from their local area and even rent a car they're not familiar with, they don't go randomly crashing into things. They probably have to pay more conscious attention than on a daily commute. But they don't go around crashing into things.

> Except that when people travel away from their local area and even rent a car they're not familiar with, they don't go randomly crashing into things.

Collision Damage Waivers and the fact that they are relatively expensive suggests the opposite.

Accidents near an airport are more than 50% of auto accidents I have had totally in my life. All of the airport accidents I was literally not moving when someone drove into me.

We vastly overestimate the competence of the average driver.

> The issue is that all the self-driving car companies ignore how people actually drive. People memorize their local driving area (...) A self-driving car needs to do the same

Self-driving car companies have been memorizing ("mapping") areas for years.

My explanation is that some writers in the technical press got taken in by the idea (possibly encouraged by exclusive stories and other favours by the industry) and wrote a series of articles about self-driving cars, how they will change driving for ever, save lives, etc. This started happening a few years ago, around the start of Google's self-driving project.

These writers, being tech writers and not AI experts, didn't understand the subject very well and basicaly just repeated the industry's claims without really understanding the state of the art.

Then people reading those articles who also were not AI experts and didn't have much of a technical background anwyay (except being interested in cars) got taken in, thiking that if the press says that self-driving cars are just around the corner, then they must be just around the corner.

Why did software folks in particular got taken in? Because most people who write software don't write AI software and even many who use it don't understand it. Computer people are used to thinking of computers like magical dream machines, so the common reaction to overinflated claims about AI in general is wonder before surprise and cheering before skepticism.

If it requires AGI, it is difficult to see how it would happen even in that event - ethics would demand "human" rights be considered, which would hopefully quickly rule out sentient vehicular slaves.

A combination of:

- People with a vested interest in FSD being RSN playing a game of topper with each other, investors, and the market

- Extremely rapid progress in supervised learning, especially vision, being extrapolated into thinking along the lines of "How much longer can getting to FSD really take if we've made this much progress in just a few years?"

- A bunch of young urbanites with dreams of never having to buy a car or even learn to drive bought into all the hybe because they so wanted to believe. (And no small part of the backlash is probably that many of the same group feel basically kicked in the teeth at this point as reality sets in.)

Frame shift drives? I would need one of those very badly.

> How is it not obvious that FSD most likely requires artificial general intelligence

Waymo car fully self driving three days ago:

https://www.youtube.com/watch?v=t_1YLHW65M8

Wow, I couldn't even properly identify the turn signals in 4 of the 6 turns. It was never as obvious as it should have been. Dirty windows, yes, but nevertheless. In an age of LED-lighting there's no excuse for not having excellent and properly colored lighting.

Now this felt like a student driver, overly busy with correctness, insensitive to the actual experience.

I think these cars should be marked, like student drivers, to signal their inevitable difference in behavior.

It's a matter of defining "full self-driving."

If you're accustomed to horses and see someone selling cars, it'd be reasonable to say, "Those are ridiculous. There are numerous places only my horses can reach. There will virtually never be a time cars fully replace horses." And you'd be right. It's still true--there are places horses can reach that cars cannot. But because of the value of cars, society has reoriented itself such that reaching those places doesn't matter to the vest majority of us.

I think self-driving cars will be the same. We don't need them to operate in every single road / circumstance ever. There may always be roads that self-driving cars can't drive on. But it'll become irrelevant because the vast majority of places people want to go will become navigable by self-driving cars, whether through the cars getting better or the roads becoming more easily navigable.

>even if the failure rate is on average /better/

Which has not been true thus far. That's the problem.

> The driverless bar for safety is way higher than very clever driver assistance or even rocket science.

Why shouldn't it be higher? Why shouldn't it be on the level of airline travel?

>Why shouldn't it be on the level of airline travel?

In all fairness, commercial airlines operate in an environment that's mostly highly regulated with respect to maintenance, training, traffic management, etc. But I agree with your general point. After all--except arguably with respect to some drug side effects--there isn't really any consumer-facing product that regularly kills people and we shrug out shoulders and say "Hey, stuff happens" and get on with our day.

Tons: alcohol being just one example.

Substances used to excess and technology like electricity used improperly are in a different category than consumer-facing products that can kill you randomly through no fault of your own.

Why?. Alcohol is a consumer product sold everywhere and is one of the main killers.

Tobacco is an even better example, in that even when used in the "recommended" way it's often fatal.

And there was a very large settlement related to tobacco.

Legal "drugs" and unhealthy foods are something of a special case. They've been around for a very long time and we've decided as a society we're mostly accepting of them and have greater or lesser restrictions on their purchase and use.

But we're pretty unaccepting of consumer products that are assumed to be safe and then blow up or malfunction in some manner that kills a user minding their own business and not doing anything wrong.

> Why shouldn't it be higher? Why shouldn't it be on the level of airline travel?

Who said it shouldn't?

OP. He was bemoaning the fact that we hold driverless cars to a higher standard than average drivers.

I think OP was just pointing out a fact, not bemoaning.

The same can be said about airplanes, which have a way longer history of automation.

Anything in an airplane is automated: takeoff, flying and landing. Pilots make mistakes all the time, but they still need to do things manually to keep their practice high.

People just do not want to go in an airplane without any pilots. In metro's it's not that obvious if someone is driving it or not. But airplanes, 'we' want a human pilot to be in control, else we feel unsafe. I think the same can be said about cars.

> Anything in an airplane is automated: takeoff, flying and landing

With airplanes there's a whole system in place to make sure that every airplane essentially gets its own private road. This simplifies the problem of automation immensely.

For airliners and other IFR traffic, that is almost (but not quite) true. In the US, the overwhelming majority of airspace below 18K feet and the vast majority of airports do not require any air-traffic-control interaction at all. It's "see and avoid" under VFR in those areas.

My favourite pinnacle moment in the history of plane automation

https://www.youtube.com/watch?v=LJkEYAy7Qc4

"Feeling" has nothing to do with the issue. Current autopilots are generally unable to cope with equipment failures. The pilots are primarily there to deal with unexpected failure modes.

Most of the failures have multiple backups, and most scenarios are also accounted for on how to handle them.

So maybe I'm wrong, but I don't see how autopilots couldn't be programmed to handle most kinds of failures (engine failure etc). Unless the failure is with the autopilot itself of course (which could have backup systems)

Most of the accidents with airplanes are caused by humans (https://www.boeing.com/commercial/aeromagazine/articles/qtr_...). So the real question is if automatic recovery would still be incapable of offsetting all the human errors.

It is impossible to program an autopilot to handle unexpected failure modes; there are essentially an infinite number of possible failures. Experienced human pilots can often figure out how to recover the situation based on applying fundamental principles. Computers won't be able to do that until we have something close to AGI.

For a prime example, see US Airways Flight 1549. The pilots Chesley "Sully" Sullenberger and Jeffrey Skiles intentionally disregarded part of the official procedure for handling a dual engine failure because they realized it wasn't suitable for their current situation.

FAA Part 121 flights are already so safe that trying to improve safety by taking humans out of the loop would be nonsensical. In some recent years literally more people in the US have died by falling out of bed than were killed in commercial aviation accidents.

Thanks for the explanation, you seem to know your stuff. In that case it doesn't make much sense indeed.

Is this really true though? Frequently I hear claims that safety standards are too strict and liability too great, but all of that could be addressed through suitable government policy.

Maybe these companies just want it both ways. They can then brag about how safe their AI is, but blame the ignorant masses for it not actually being deployed.

Yes, and to shrug off responsibility when something horrible happens.

At this point the ignorant masses are pretty astute to see beneath the hype there’s a disturbing amount of reckless optimism.

> While we accept to be killed by our own mistakes at the wheel, society (we) will not accept that a machine makes the same, even if the failure rate is on average /better/.

We seem to accept being killed by someone else, since that is how most of the 40,000 people are killed every single year in the US. I don't hear too many drivers complaining about that.

I'd say it's also because today we instinctively know FSD cars have not proved their skills in the same conditions a human driver does every day. Racking up millions of miles by following a line or the car in front on the highway is far from dealing with a crowded city center, bad weather, unexpected issues that have no clear rules of engagement, etc.

Tomorrow the situation will be different. But it's hard to trust such a broad safety claim knowing it's based on partial data and say "the failure rate [for self driving] is on average /better/". Comparing failure rates obtained is such different conditions is misleading.

Self driving cars are like a student driver with supervision. They currently drive in relatively controlled environments, or at least ones with very forgiving conditions, and are learning. And just like student drivers with supervision they are relatively safe. Until the moment they're exposed to the full set of driving conditions without supervision. We intimately and instinctively know and trust how the learning process work in humans. For cars not so much. Not yet anyway.

> We seem to accept being killed by someone else, since that is how most of the 40,000 people are killed every single year in the US. I don't hear too many drivers complaining about that.

As a driver, that's not really something you have control over, apart from driving defensively. Your AutoPilot Tesla can still get T-boned in an intersection by some guy running a red.

it seems the bar should be that high as that is what the hype has told us we would be getting. I'm OK with it being as good as the hype ...and when it isn't the companies that write the software and build the cars can pay for the damage

I see a lot of talk about how society will not accept machines that might make mistakes, but I don't think society has actually had a chance to make its opinions known yet. Society isn't the noisy voices and the click bait headlines, it is people who mostly quickly adapted to elevators and lifts and planes and trains, all of which are vehicles that have killed people while operating autonomously. I expect you will find the same worries about them too if you trawl the newspaper archives.

Yup, I'd say it is more that corporations aren't ready to take the legal risk of SDC.

If I get into a car accident, my insurance company gets the fun job of litigating. Either I or the other driver will be found liable.

If the car was driving, however, and it makes a mistake then the manufacturer could be held liable. And since car manufacturers have deep pockets, you can bet they'll get sued.

This is why manufacturers are pursuing perfection. Not because of a moral obligation, rather, a financial one.

I wonder if the first true autonomous cars will be in Europe or somewhere less litigious, and not in Arizona as most people seem to be expecting. Maybe international European or interstate Australian trucking will be the first success (or not success... )

Not only that, but it's the type of situations that will kill you in each case that make a difference too. From the various self-driving related death I've read about, most of them appeared to be situations that could have easily been avoided if a person had been driving and paying full attention. On the flip side, the kinds of situations that kill normal drivers are often ones that a self-driving car could trivially avoid.

So with a self-driving car, you have the mental hurdle of asking people to accept the risk of the car killing them in a situation they could likely have avoided had they been driving, even if it is statistically less likely than them killing themselves while driving.

Everyone said the first death via SDC would be a massive setback. After the Uber death in Arizona, have we seen that massive setback? I'd argue not, I think you are vastly over-estimating the risk aversion of the American public, which has allowed opioid epidemics, numerous pointless/corrupt foreign incursions, and allows the caging of children. Can we factually say we're a nation that highly values safety and human life?

The delay is mostly due to a lack of suitable EV's and suitable tech. If big business was ready, big business would ram this through as soon as it could generate a good profit or good PR.

There are a lot of reasons Arizona didn't register on the radar much.

1) Arizona pretty much invited this problem through publicly denying oversight or responsibility. This is more a problem for Arizonians than it is for self driving tech. If things were done above board, there would be more questions and concern.

2) Everyone is used to Ubers driving badly, nearly killing other people, and generally escaping the law. This wouldn't be tolerated from Waymo/Google.

3) Deceitful PR with the camera footage painted this as solely the pedestrian's fault and not an alarming sensor/software issue.

I don't understand why there is so much focus on automobiles instead of just about everything else. In every other instance - airplanes, boats, long-haul trucking - the problem in significantly less complex to solve.

The only tangible reason I see for such disproportionate focus is some sense of false futurism; aka, driver-less cars have become the 'flying cars' of our times.

Large aircraft are basically self-driving these days. Pilots can choose to do almost nothing at all from takoff to landing. However, the failure modes are all mostly catastrophic so pilots remain under the assumption that they can slightly increase the survivability rate when the automation fails. Modern drone systems are also have elements of this kind of automation and can fly complex flight paths and station keep and land and so on more or less on their own.

Some large boat systems have are also similarly "self-driving" and can navigate to way points and along various lanes without too much intervention. They need pilots and tugs for the small-nav stuff like canals and ports and so on, there's simply too many permutations of "things that can be hit and sunk" to rely on the automation. Smaller boats can also buy commercial "autonav" systems that link the engine, steering and such to GPS. However, the marine environment is hellish on equipment.

Trucking is being seriously looked at, with active test runs, but like big ships needs local pilots for the end-points and local nav. Again too many things in flux and too many things to hit.

Cars exist basically continuously at the terminal ends of the trucking problem and are trying to solve the local pilot problem. Humans, with our highly adaptive sensory-response systems can do it to various levels of skill, but even we have lots and lots of failure modes even without impairment.

One advantage of ocean-going ships for automation is that they can (and do) embark pilots for operation waters.

On the flipside, the machinery isn't designed to operate for weeks on end without any engineering crew.

Long-haul trucking would be a great use-case. Even if you need humans for local pilot guidance, you can queue them up to a "cell phone lot" and courier the last-mile drivers in between shifts of going back & forth. It would also mean drivers would be based more locally to points of loading, rather than having to be constantly on the road away from home.

>It would also mean drivers would be based more locally to points of loading, rather than having to be constantly on the road away from home.

Which would probably make it much easier to hire drivers, too. Lately, long-haul trucking has had a hard time recruiting people into the profession, because the pay isn't that great and the lifestyle completely destroys any relationship you have. Local trucking doesn't have this problem: a driver can drive for the day and be home for dinner.

Long-haul is more than just driving. Will smart trucks be able to put chains on and pull them off when necessary? Handle unloading of items of weighing process proves problematic? Call for pick up of items removed during weighing? Handle snow? Handle loss of sensors? Handle run away down hill on a steep incline. None of these are insurmountable, but all are well out of the current horizon. Heck, just refueling is hard.

Most of those can be solved by changes to the truck or delivering spot local drivers. For weighing the trucks could run different trailers with load cells to weigh itself to ensure they're under the weight limit for their route. For snow there are automatic chain systems used on buses today those might be adaptable. Run away ramps are usually marked so a combination of good mapping data for route planning and finding run offs and vision/driving they'll have anyways.

> Large aircraft are basically self-driving these days. Pilots can choose to do almost nothing at all from takoff to landing.

Sort of. The plane will continue to fly the pre-programmed route without pilot involvement, so the actual manipulation of the control surfaces is no longer their primary task. But there are systems to monitor, controllers to talk to, weather systems to navigate around, climbs/descents to find more favourable winds, fuel imbalances to correct, etc. And of course, baseball scores to pick up from ADF.

I'd liken it to how (most) software engineers are not writing in Assembly. Instead they're using higher-level languages and libraries to solve problems. Similarly a pilot is not hands-on flying for 7 hours over the Atlantic, but they're still operating the aircraft.

They could automate pretty much all of those away but even if they did I don't think flying is a good comparison for self-driving. Flying is pretty much the ideal environment for computer control because there are only a few actual obstacles to deal with; weather, terrain, and other planes and all of those can be handled with outside signals.

> airplanes

I remember when Mythbuster did their episode on driving an airplane in emergencies like they do in movies, they came to the conclusion that if it was going to happen, the airplane autopilot would be used instead or a civilian that never flew one. That was in 2007.

> long-haul trucking

I'm pretty sure that's a long term goal of Tesla with their new semi. I have no doubt many are on that too. The challenges are similar to the one for a car.

There's no so much focus on automobile, it's just that we are talking about it more than the others one. Hell the airplane one was done more than a decade ago and you still don't know about it.

Another drawback is that it's taking focus away from mass transit, at least in the US. Decent mass transit ameliorates such a wide range of issues in cities it's hard to understand why we don't put more effort into making it work.

Mass transit doesn't really work in US suburbs because everything is just too spread out; it's not economical, nor logistically effective, to build subways to subdivisions.

But even in built-up cities, Americans cling to their cars just like they cling to guns and religion. Look at all the uproar in Manhattan, NYC, where they're taking away street parking and putting in protected cycling lanes. People are upset at this "attack" on cars, and somehow think they deserve free places to put their personal vehicles. If you look at another highly built-up city for comparison, Tokyo, they don't have this problem. There is NO street parking; it's all paid lots that cost a fortune. In fact, you're not even allowed to buy a car there unless you can prove you have a place to park it! So most people get around with subways, trains, walking, cycling, and taxis.

The bottom line is that it's cultural: Americans just don't like public transit and don't want to fund it with their tax dollars, but they do want to subsidize car driving with their tax dollars.

Americans like cars so much because they're incredibly useful. And public transit sucks, which is sort of a chicken-and-egg problem. We don't want to invest in something which is unpleasant to use, so it remains unpleasant to use. In my city, we have the additional problem that the local transit agency only wants to spend money on light rail, which is just a smaller version of the "nobody builds subways to subdivisions" problem. I'd be more likely to use mass transit if the bus was clean, comfortable, and came a lot more often than once an hour.

Yeah, I always describe it as the US setup being in a relatively stable configuration currently and the mass transit solution being another very stable setup but the transition between the two being both painful, expensive, and time consuming.

It isn't so much the spread out, as that there are not good routes for mass transit in most suburbs. Transit needs long straight lines with destinations along them. When roads wind, and destinations are at the end of cul-de-sacs transit has not chance because riders are forced to spend a lot of time making no progress in the actual direction of their destination.

I don't necessarily see that as a drawback. If the technology for self-driving vehicles really matures, I think it could be a big boon for mass transit.

The reason is a combination of two factors:

(1) Transit will probably always need buses. Rail and subway are great at certain specific things (very dense areas), but they are expensive to build and have trouble achieving good coverage of an entire city.

(2) Buses tend to be big and hold ~50 passengers. Why? Mostly, I believe, because bus drivers have to be paid a salary. You're pushed into choosing large vehicles to try to amortize that cost across more passengers.

What if you could take away that constraint by eliminating the bus driver? Some of those 50-passenger buses could each be replaced with (say) three 16-passenger buses. You'd have a choice. You'd no longer be forced to try to consolidate by reducing the number of routes or the frequency on the routes. Both of these things (long walk to a stop, long wait for a bus) are major reasons why many people don't like riding the bus.

And if you can make transit more convenient for people, you can get more of a critical mass of people using it.

I think the difference between your other examples and personal automobiles is that aviation, shipping, and trucking are important industrial concerns that are worth paying someone to operate -- while the private driver is almost completely superfluous.

A ton of money has been dumped into automated long haul trucking already. I think you're just not aware as to what type of work is being donee.

Is it a cost structure issue? Ie, replacing the captain on a ship that holds 15,000 shipping containers, does that lower the cost of each container by a couple of dollars? Similarly, with airline pilots, how much of my ticket cost is pilot labor?

However, if you look at Uber, if you removed driver labor cost from my fare, does each ride now become 60-80% cheaper?

Whats the percentage of the cost of the driver for airplanes, boats and long-haul trucking, compared to the cost of each trip? Maybe for the trucking it is significant, but I'm not sure airplane ticket prices would go down significantly if we automated them.

But with cars, the price of the driver is actually pretty big, at least 50% of the cost of each ride, when taking a cab or Uber. So that can be a lot of money people are saving.

Plus the whole part where we might get rid of parking spaces in urban centers. This might actually change the way we build cities in the future.

With airplanes, it might open up new things that seem far fetched today. For example, commuting to work with a drone.

In the US, most people experience car commutes and being stuck in traffic all the time.

Fixing that in a way that requires no lifestyle changes is an instant ticket to the Forbes 500, so it’s easy to get cofounders, investors, and media visibility.

It is already solved for (very small) airplanes in Rwanda and Ghana: https://www.youtube.com/watch?v=bnoUBfLxZz0

It's easy to make polarized judgements, like "AI driving can only follow lanes or drive on wide broad sunny streets". I recently watched a lecture where a guy from Waymo talked and demoed some situations, check out the 3 seconds or so in the timestamp in this video which made me re-evaluate that pessimism a bit:

https://youtu.be/Q0nGo2-y0xY?t=865

Sure it's still nowhere near the general understanding necessary for completely autonomous driving but still much more impressive than simple lane-following.

Nice clip. It's good to see actual evidence and data, rather than the rampant speculation that always comes up with SDC topics.

That clip is far from convincing. Sure, the video showed an AI making the correct decision in a single complex circumstance. But that is assuming you have a correctly-modelled situation with accurate sensors. Even if you could guarantee that the AI could make the correct decision every single time, millions of times per year, you have the problem of potentially inaccurate and/or failing sensors which give the AI an incomplete model.

This is why airplane auto-pilots disengage when they detect failing sensors, giving humans enough time to take over. Unfortunately in the case of cars, there is not enough time to hand over to a human.

It isn't hard to have redundant sensors so that you can drive for a bit on less than all of them. If there is a failure you drive to a mechanic and get the senor fixed before there is a second sensor failure.

The odds of a dual sensor failure is low enough to still be substantially better than humans. Assuming of course that the dual sensor failure is dual unrelated failures. That means care in the wiring of power systems, fail safe modes, over design such that failure can't happen, and other engineering tricks.

The most pertinent point here being that although collecting lots of self driving car data is easy, the tail and unexpected events that you care about are surprisingly rare. And many of these events can also be entirely new (never seen before). Current AI SOTA techniques simply require too much data and cannot reason well enough to solve this.

Leaving Musk aside, amongst those who are serious about self driving cars, e.g. Waymo, there is still a clear emerging picture that current AI techniques are too weak to get there right now. The limitations of current sensors and compute hardware are not helping either. Going the other way and eliminating LIDAR isn't going to help.

I agree that ditching Lidar by Musk is the wrong move. From what I have read most if not all the self driving accidents that Tesla have been involved in could have been avoided if a Lidar type sensor was involoved. With tech available trying to use the same limits as human vision is stupid.

lidar isn't cheap or small (yet), and you need cameras either way, can add lidar to the sensor suite later if it gets cheaper

I agree with Musk and George Hotz that lidar is never going to get there. That's because in Michigan the terrain that lidar has mastered changes drastically every time it snows. For those of us in snowy climates limiting us to self-driving to seven months a year isn't going to cut it.

But this doesn't explain why self-driving cars won't succeed in Arizona. They don't need to take over the world at once, they just need to work somewhere and that would be a good start.

Lidar has no problem with snow, only when its _snowing_. However, if visibility is reduced, a pure visual approach is going to be even worse.

Musk made a mistake, and he's spending billions to prove that he never made it.

snow makes HD maps somewhat less useful, as the terrain changes shape. But then, you need a map that evolves. Sadly for musky, the only company that can do that using purely visual data is Scape Technologies. Not even lyft can do that.

>Lidar has no problem with snow, only when its _snowing_. However, if visibility is reduced, a pure visual approach is going to be even worse.

Good thing Teslas have radar...which penetrates falling snow.

>Musk made a mistake, and he's spending billions to prove that he never made it.

People drive fairly well with only visual sensors (eyes). That is proof of concept. Lidar seems like overkill.

That radar is utterly worthless for this sort of application, it has awful angular resolution. It can't tell the difference between a building next to the street and a truck parked in the middle of the street.

Just because lidar is not working in certain places and certain conditions does not mean that it's always useless. It just means that an AV cannot rely [only ]on lidar. Lidar could be used as a suplementary sensor - that is providing additional sensory data when appropriate and switching to non-lidar mode when conditions require. This is one of the things where AV can have a clear advantage over [even above ]average driver: a superior sensory system

If you switch to non-lidar mode for unfavorable condition, why don't just use the non-lidar for all conditions. A system that is safe for unfavorable conditions is likely to be the system that is safe for all conditions.

Because lidar allows faster travel in the modes where it works. When there is heavy rain/snow the only safe speed is very slow anyway. When roads are clear and dry I really want my car to go faster than the elderly walk (I walk faster than that normally), and lidar may be required for that mode.

Because one system might work better in one situation and the other in a different situation and you are going to encounter both situations?

Just like in cooking you can use meat cleaver and a paring knife or in a network you use switches and routers.

I think Musk has no choice but to ditch Lidar. He has promised to provide full self driving on cars already on the road, and retrofitting all of those cars with Lidar is incredibly expensive.

Plus LIDARs aren't aesthetically pleasing in their current incarnations.

Bologna. He had/has a choice, he chose snakeoil salesman. Tesla could have focused on affordable (or luxury) electric vehicles. The choice to add self-driving was all Musk.

> He has promised to provide full self driving on cars

I continue to be surprised he hasn't been taken to court yet for selling a $6K package that isn't anything but an empty promise.

> Current AI SOTA techniques simply require too much data and cannot reason well enough to solve this.

If they had someone working on using the tech to ID mushrooms growing along the road then they’d probably come away with some new techniques to make the cars safer. Trying to make a self-driving car algorithm by only thinking about driving-related things seems like it will just lead to overfitting.

And in fact any car that doesn’t have enough cognitive surplus to identify mushrooms while driving in normal conditions may just not be safe to drive period.

I agree LIDAR is necessary.

The goal is not to achieve perfection, just demonstrably more safe than an average driver.

I'm curious how we will evaluate average driver. Clearly, including drunk driving accidents in the mix is not a helpful metric for safety-conscious drivers.

”Clearly, including drunk driving accidents in the mix is not a helpful metric for safety-conscious drivers.”

Sure it is. Even if you never drive drunk, there are still other drunk drivers on the road who are a danger to you.

But a self-driving AI, even an imperfect one, never drives drunk.

You're imagining a scenario where everyone has a self-driving car and completely ignoring cost and social factors.

Who among us would admit, "yes, I am a bad driver, I need this more expensive thing to take control in the event of an accident"? Very few is my guess. In 2019, people use these for convenience, not because they think the system is a better driver. I bet that's true even for Waymo's taxis.

As for getting drunk drivers to use self-driving, that's as easy as asking them to either stop drinking or not drive. Then there is the cost. I wouldn't be surprised if the unemployed drink and drive more often.

I don't think total coverage is necessary. If 10% of the cars are self-driving, to a first approximation there will be 10% less drunk drivers.

Cost is definitely an issue, but reducing accidents and increasing convenience have a cash value which could pay for it if the research pans out.

What social factors do you have in mind?

> If 10% of the cars are self-driving, to a first approximation there will be 10% less drunk drivers.

It seems unlikely that people who buy/use early self-driving cars have the same percentage of drunk driving as the general population. Surely employment rates would be a factor, for example.

I'm not sure whether or not there is a correlation here. Anecdotally, it seems to me there are plenty of wealthy, employed people who buy new cars and crash them after drinking.

But even if it is true that those who drive cheaper, older cars are more likely to drive drunk? Today's new cars with high-tech self driving features are tomorrow's affordable used cars. You have to start somewhere.

If self driving cars exist courts will be more likely to take away drivers licenses from those who drive while drunk. Thus I'd expect 10% of the cars being self driving to result in something like 20% less drunk drivers because such drivers are more likely to get the self driving cars - not that they want them more than anyone else but because they are not given any choice.

In utilitarian terms, self-driving cars just need to be better than the drivers they replace. (Or even stronger: not much worse, because the added convenience is worth something.) Ie we should start to replace the worst drivers already.

In practice, the target seems to be not just the average driver, but the best human driver (or at least the 99%th percentile.) Given the trajectory of improvement we often see with AI, surpassing the average human and surpassing all humans shouldn't be too far apart.

I disagree. It needs to be absolutely perfect, not just safer than the average driver.

If only there was some way to compute entailments from unlikely combination of events :-) But seriously, Gary Marcus and Ernie Davis make a good case in their new book Rebooting AI that logic has to play a role in any real AI.

That's false. The main reason Waymo is held up is because the accepted standard of driving is unsafe, and it has data and simulations that prove it.

For example, an unprotected left turn during rush hour. The gap between cars that people drive into in that situation often does not leave any room for chance. For example, if another car speeds up or changes lanes, there can be an unavoidable accident. But the standard for human drivers is to take those chances during heavy traffic, and drivers that try to avoid that are not accepted.

The issue is that people are incorrectly going to hold Waymo's computer responsible for some unavoidable accidents in those types of scenarios, but at the same time expecting them to make those maneuvers even though the computer knows there isn't enough margin.

That's not a weakness of the AI. That's an inability of people to understand the nature of traffic and the relationship between AI and people.

There were scenarios in which this LIDAR technology would effectively blind detectors on other cars. I can understand how this puts the use case for this technology in question, since ideally, you'd want passive sensors, or at the very least sensors that don't put other people's lives at risk.

The "easiest" (read: least difficult) to automate would be long-haul, between cities. But, we have an "automated" system for that, it's called railroads, where one driver can steer dozens or hundreds of cars.

The hardest to automate is in-city driving, where the environment is much more complicated to parse, and there are many more kinds of fellow-users (pedestrians, bicyclists, etc.). So, naturally, that's where they went first, because it's not about what makes engineering sense, it's about what gets the "dumb money" to invest.

Perhaps a city is more complicated because of the environmental hazards you mention, but I'd imagine it's easier to outfit a city with self-driving friendly sensors or lanes than those long-haul routes. In the article, it says this is how China is approaching the problem.

I personally believe a society with autonomous cars and regular cars sharing normal roads is a pipe dream in both a technical and political sense.

cargo by rail is actually pretty expensive. A fleet of autonomous trucks driving interstates would definitely be more cost effective

Just to ask, haven't there been several reports of LTE/5G protocols being compromised, making the worst-case scenario one script-kiddie away from major disruptions?

Why would they even contemplate building on a compromised foundation or worse yet, accelerating a roll-out? Since most of the tech from last century was built on trust models/trusted environments, waiting for a clean rewrite sounds like a better plan .

I personally hope we see them build new 21st century cities and dedicated autonomous highways to grow driverless technology, instead of trying to overlay it on the existing horse and buggy infrastructure with time-period correct code/models.

The world doesn't need any more digital diarrhea on its plate.

Tesla is already selling "Full Self-Driving Capability" for $6000.

Excerpts:

Coming later this year:

Recognize and respond to traffic lights and stop signs. Automatic driving on city streets.

Fine print includes this:

The activation and use of these features are dependent on achieving reliability far in excess of human drivers as demonstrated by billions of miles of experience, as well as regulatory approval, which may take longer in some jurisdictions.

https://www.tesla.com/model3/design#autopilot

They've been selling it for years, I think since 2016. They've delivered it so far to nobody.

But they just delivered "Smart Summon" feature where the car drives itself to you in a parking lot. The rollout of this feature has been glitchy. But it gives them a lot of data which they will learn from and improve. Once they fix all the bugs in Smart Summon they will have made a major step forward. Self-driving safely in a parking lot is very hard--there may be kids walking about randomly, other cars moving in unpredictable ways, and so on. In fact, my guess is that self-driving in a parking lot is harder than self-driving in city streets, because there are more rules, more room and more predictability in city streets.

I wouldn't be surprised if Tesla delivers city-street self-driving in a year or two.

Parking lots are hard for the reasons you mention, but they are also easier because the speeds are slow. Hit a kid at 5mph and it is probably not harmful, and even if it would be you get a lot more cycles to recognize the problem and need to react vs if you were driving at high speed.

That reminds me of the canard I haven't seen in quite some time: "Tesla is going to win the self-driving race because they've got the most data!" Has this died because people have started to finally understand that the logic doesn't actually hold, or is it because Musk/Tesla have stopped promising that full self-driving is actually imminent?

It hasn't died, the logic holds, and Tesla is still promising self-driving is actually imminent.

In addition to the most data, Tesla also has the best self-driving computing hardware you can buy today. (See autonomy day https://www.youtube.com/watch?v=Ucp0TTmvqOE ) They also have top software people. I think Tesla may very well release full self-driving in the next couple of years.

Tesla claims that autopilot is involved in one accident for every 4.34 million miles driven. Seems like a useful number, of course the media reports that the average is 1 accident per 500,000 miles.

So 9 times safer... right?

Question is how many accidents do similar people (read that as rich enough to buy a $50k+ car) in similar conditions (highways) with similar cars (newish luxury cars) without autopilot?

No fair counting drunks, meth heads, crappy cars, or snow storms (where auto pilot wouldn't work anyways).

Surely some of those autopilot miles involved drunks, so they shouldn't be excluded entirely.

Indeed, but likely at a lower rate than the general public. In any case the general idea is to find as fair a comparison as possible with the Tesla published numbers.

I suspect $60k BMW owners driving on the highways in good weather are much safer per mile than random folks in random cars on random roads in random weather.

while I generally agree with your point, that new cars with great braking, enhanced safety and in perfect technical conditions are safer than average, I'm not entirely convinced if BMW is the right make for this kind of examples

edit: typo

Heh, ok, pick any luxury car priced similarly to the Tesla.

It's odd that so many of the comments on the self driving threads make this assumption that drivers of expensive cars are safer than drivers of inexpensive cars. It is just part of the classist bias that attaches virtue to wealth? Or is there some data that suggests that a Mercedes driver is a better driver than a Kia driver?

BMW drivers are stereotypical bad...

I don't know of data, but the high cost of a Mercedes over a Kia means that the average driver is likely to be older and young drivers are a known risk factor in accidents.

It's a combination of things. For instance drivers are statistically more likely to have accidents at 18 than 50. Drivers are more likely to have an expensive car than 50 than 18. Being less financially secure often means you are trying to keep an old junker with crappy brakes going when young.

Also risky behavior like speeding, DUI, driving aggressively, passing when unsafe, etc are generally more common when young.

Sure individuals vary, but just look at the cost of insurance to verify.

Do BMWs have a bad safety track record?

I believe the parent poster is referring to the stereotype that BMW drivers have an aggressive and selfish driving style, and that they never use their indicators/blinkers.

I think that is an incredibly misleading statistic. For it to be comparable to human driving, autopilot needs to be turned on all the time, no exceptions. I think most people only turn on autopilot under ideal road conditions, so you would need to compare the human driver accident rate under perfect highway road conditions.

Well that's not realistic... yet.

But it would be interesting to see how safe people are without autopilot in similar close to ideal conditions with a close to ideal car.

You hit the nail on the head, it isn't a fair comparison. I suspect autopilot is largely used under 'safe' and ideal conditions, and as a glorified lane-assist/cruise-control feature. You add complexity, and things deteriorate. This is why Tesla had to severely pull back on their initial claims after several high-profile crashes due to driver inattentiveness, and you need only to watch Tesla navigating a parking lot with it's 'smart summon' feature to see we have ways to go still.

And you're correct, driver demographics are important as well. How does Tesla's autopilot safety record stack up to the record of other drivers of luxury cars? Luxury cars will tend to have, for example, proper tires for the season. The one accident I was involved in resulted from the other driver not having winter tires and rear-ending me at a red-light. So is it the autopilot that is responsible for the improved safety stats, or 'dumb' companion features, like a built-in blind-spot warning, and a properly maintained vehicle.

Also, though human rate of accident is on the order of 1 per 500,000 miles, the fatality rate is on the order of 1 per hundreds of millions of miles -> how does Tesla compare to that? Reducing minor accidents is important, but it isn't the primary consideration.

It's clear, this is a misleading marketing number. Don't trust the Tesla autopilot to drive you around safely without your full attention - that's not me saying - that's Tesla saying it.

I use autopilot everday on my 50 mile one way commute. This statistic is so misleading. Autopilot is used while i ride in the HOV lane for 45 miles of that commute. Autopilot including Navigate on Autopilot can't get me to cut across 5 lanes of traffic in under 1 mile or do the city driving where i am 10x more likely to get in a car accident.

Autopilot has caused me more near-misses than it has prevented. And I only use it on the highway.

The statistic would be more useful if 1) only compared to similar driving conditions, and 2) driver is not allowed to intervene. As long as the human is there to save things when autopilot goes pear-shaped, it's impossible to say how safe autopilot really is.

https://news.ycombinator.com/item?id=21218694

Previous discussion.

So, do we, as modern people, consciously accept that in order to get faster to our destination some people must necessarily lose their lives because of the technology we use to transport us?

Given that our transport technology is not completely safe, we must accept that some people will die using it. Even a very low chance of lethal accidents means that some lethal accidents _must_ happen. Some people _will_ lose their lives. So in order for most of us to get to our destination by plane or car or train, etc, some people must never reach theirs.

Does the general public appreciate this or is it just the case that people get into their cars or board planes, etc, thinking "nah, it won't happen to me"?

> Does the general public appreciate this or is it just the case that people get into their cars or board planes, etc, thinking "nah, it won't happen to me"?

I think the average person recognizes the risk tradeoff. Though I suspect a lot more people worry about planes than worry about cars, even though planes are statistically safer. Guessing that is probably due to a combination of speed, gravity, and lack of direct control over the plane.

I don't get why we can't create a "full self driving suitable zone", with standarized road paint, signs, beacons, traffic lights which broadcast their state on a suitable standar, and maybe a car to car communication mesh.

Start small with something like the mortantown prt, even in segregated lanes, and grow from there.

I can't see a way to have full self driving capabilities, without helping the vehicles with some context information.

Infastructure is hard enough to get spending on when it is needed long term and immediate unfortunately - let alone for experimental applications.

That sort of spending would rely upon rational risk assessment instead of irrational fear so utterly unrealistic politically unfortunately.

The problem is paint fades, beacon lights burn out (or the power fails). You can detect and fix those problems, but there is still the time between the issue and when it is fixed that the system needs to work. I've driven in cities after heavy snowfall took out all the power and you couldn't see any lane markings - we slowed down and used extra space (4 lane road, but everybody was driving between lanes making it effectively a 2 lane road), but those who needed to get around could - we even got through intersections despite the stoplight being out.

That is the self driving car needs to handle cases where the infrastructure around it is dead before it can react to infrastructure helping it out. This is needed in any case to ensure "terrorist" situations are not too harmful.

>a "full self driving suitable zone", with standarized road paint, signs, beacons, traffic lights...

We've done this already. It's called "railroad tracks" and the self-driving cars on it are called "trains".

I remember 2014 being a year full of these incredible claims of fully autonomous vehicles in the near future. One article in particular I remember, about some large car company expecting 80% of cars on the road to be autonomous by 2018.

I remember this vividly because I was house hunting at the time and became truly conflicted what location to buy. Should I choose a smaller more much more expensive house nearer to the city center, or a nicer drastically cheaper house further away? If I’ll be essentially chauffeured to work in a few years, why not go for the bigger cheaper house. I eventually chose the more expensive house nearby.

Since then, the small yet constant worry that Autonomous cars are going to show me what a mistake I’ve made has disappeared, in parallel with any noticeable progress of autonomous vehicles hitting the roads.

Thank god I didn’t choose that house in the boonies.

See also an article on Chinese vs. Western thinking on autonomous vehicles https://www.economist.com/business/2019/10/12/chinese-firms-... in last week's Economist, and a recent example of an edge case taking non road based autonomous vehicles off one campus https://thespoon.tech/starships-robots-pulled-from-universit...

Heathrow has driverless pods:

https://londonist.com/2014/09/a-ride-on-heathrows-self-drivi...

I'm not sure how you could introduce that into a modern mixed traffic environment though. Imagine the howls from all the incumbents.

The solution is pretty simple. We need to design electric highways that have a self driving lane. In the self driving lane there is guidance in the road that assist self driving cars. Since the cars are self driving they can coast in 80km/h sharing the lane with trucks. Add electric supply and rails to lessen the tire friction and we are a long way to reduce carbon emissions from transportation.

For city centers cars should for now assist the driver but not self drive as the system are not yet safe enough todo so. From machine learning we know that image analysis has probabilities associated with them. Probably might not be safe enough in cities.

> The solution is pretty simple. We need to design electric highways that have a self driving lane.

So, build train lines.

Any product that requires refitting existing roads is pretty much dead in the water, there's a chicken and egg problem inherent in that solution because before there's a push for billions of dollars and decades to be spent refitting the highway system there has to be a big demand for that project/product and that will only happen with a proven viable self driving car. And even once it starts the vehicle will only be useful once enough roads have be refitted. So a company has to 1) convince the US government to dedicate years and a few billion dollars to build them the infrastructure to run their product then 2) protect that funding across a couple decades (including multiple party switches) and 3) not go broke in the process waiting for the infra to be in place for people to actually buy their product.

There is an anecdote I half remember, that I would like to find the original source of, because it's illustrative when explaining the issues. It goes something like:

In the 70s/80s a research group was working on a CAD system. After adding all sorts of features, the professor assigned a task to one of his students to make it respond to voice commands. The professor expected this to take only a few months.

It is difficult, even for experts, to have a good intuitive idea of where a good ML implementation is going to be easy or hard. You can be making good progress and then hit the "diminishing returns" part of the curve unexpectedly.

The classic AI anecdote is that the problem of object recognition in images was assigned as a summer project in 1966. It was still considered an open research problem 40 years later.

Ah, thank you. With a bit of further googling, I found this delightful document:

http://people.csail.mit.edu/brooks/idocs/AIM-100.pdf

Self driving cars only make sense to me as a replacement for taxi services.

That is, you don't own and don't maintain the car and you don't care where the car is. You call it when you need it and it drives away when you don't.

I can't see this realized until we avoid the uphill battle of not running over pedestrians and deciding which one to spare in case you cannot avoid running one over. One way to avoid this battle that I see is completely outlawing pedestrians on the road.

Then you have cyclists and similar. What do we do about those ?

Considering the above I don't think driverless cars solve a problem that we have.

Self driving cars mean I can daydream instead of staring at the road. I daydream anyway: I know it isn't safe but I can't help it.

Woh, An article that correctly identify machine learning as a sub-field of AI.

I asked this question before and received no answers, so I'll ask again: Is there a source that details the specific issues that self-driving cars could face, and potential methods of solving them?

https://blog.piekniewski.info/2016/11/28/more-thoughts-on-se...

Is pretty good. For a more recent take he has also written some blogs on the short comings of AI in general which has some self driving car content.

https://blog.piekniewski.info/2018/05/28/ai-winter-is-well-o...

https://blog.piekniewski.info/2018/10/29/ai-winter-update/

I'm a little confused by the delivery only cars like Nuro. Logistically how would the package be delivered once it arrives at my apartment...? Would I have to be home and physically grab my package from the vehicle? If there's a human assistant who needs to drop off my package to my doorstep then is there really any cost savings over a regular driver?

Yes, you'd have to be home to take things out of the vehicle. It's more the pizza/grocery delivery model than the "we left your Amazon package on the porch" one.

That seems pretty inconvenient. I'd predict a ton of missed deliveries. Even with Amazon/Whole Foods delivery they ring but will just drop it off at my door.

I'd expect them to be very clear about needing to be home, letting you pick a specific time, and a fee for a missed window.

I don't remember the source, but a few years ago I saw the results of a poll about self driving cars. The question was: "How much more would you pay for a car that could drive itself?" In the US, 50% of the responses said $0. It is not a feature that the market wants. That's why we don't have self driving cars.

But this is so obviously wrong. Every time someone takes a cab, they're paying for a "car that could drive itself". There's a huge market for this.

How much more often would you take a cab if the cost was dramatically lower?

Not to mention the existing market for professional drivers that aren't transporting people.

People don’t always know what they want. Before Steve Jobs unveiled the iPhone how many people knew they wanted a smart phone?

Anybody who had a palm pilot with a cellular radio strapped to the back :).

Touché! I do think Apple owes PalmOS a lot! I wish other platform vendors such as Microsoft had also copied from PalmOS.

I don't think that's the reason. There's a lot of money going into self driving cars and a lot of talented people working on them. The reason is that it's easy to make a self-driving car that can drive in predictable situations (first done in the 80s), but it's extremely hard to make one that handles unusual situations.

The AI part of autonomous driving could have easily been handled by creating a network of roads (that is, converting existing roads) that has the appropriate markings and that is easy navigable by machines.

But that requires cooperation between governments and/with the private sector, and in 2019, God forbid that.

That's called trains.

And isn't it interesting that even though we've had driverless trains for decades many of them still require a driver?

In London the DLR (Docklands Light Railway) is fully automatic. In Tokyo the trains are computer controlled but a driver takes over outside of peak hours, apparently for practice.

So perhaps we'll never get past a self-driving car requiring a human operator at least pretending to pay attention (so no real change there, then).

The bigger hurdle is the cost to redo roads.

2-6 million per mile of road. 4 million+ miles of road in US.

It’s not so much the roads though, it’s the other cars and other things/people/animals in the way

We just invested an ag robotics company started by senior engineers from some of the top driverless car companies because they felt that they were never going to actually get a driverless car to market. We welcome the influx of talent.

Does Waymo even use the kind of machine learning with neural networks that's currently en vogue?

why is this foam of words called an article . And why is this on the first page of HN .

How does one get past the registration/paywall?

Try opening with javascript disabled.

archive.is has it - but honestly there's not really any content. Several paragraphs to say it's harder and will take longer than some people promised, because "edge cases".

The answer is usually click the "web" link at the top. In this case it takes you to a google search for the article title, and then clicking the search result on Google opened the article fully for me, presumably it's the referrer.

Fixes most paywalls: https://github.com/iamadamdev

More often than not, switching User-Agent to Googlebot is enough

By paying

  a) Use Firefox
  b) Have 100+ tabs open
  c) Click on the link
  d) Quickly press the "x" button to the right 
     of the address bar to stop the page loading
     the paywall.

Works on my system :|

I'm interpreting this paywalled article title literally, so I'm assuming driverless still can't solve the traffic jam?

A solution to this could be far less complex than using AI or sensors (installed per car) to steer the user automatically. Consider curing congestion by connecting most car's existing cruise control to a city-wide traffic light network:

Metal strip sensors are embedded into the first couple dozen feet of road at the beginning of an intersection, and this length is determined by the local DoT according to expected throughput (sometimes there are 3 or more sensors to count how many cars are queued up). This is because over the past several decades, classical traffic control created the enforceable expectation that vehicles come to a complete halt at the color red.

There is generally a pattern constant throughout all cities: primary arteries take priority until the left turn sensor is activated on the street parallel to the oncoming street, but the perpendicular street sensor activation takes priority before this. These are usually lower speed neighborhood or shopping outlet access capillaries emptying out into the higher speed artery stream.

Think about the feeling of seeing a dozen cars and heavy trucks going 40 mph and suddenly having to stop for a yellow light because only one car has activated the perpendicular sensor. That's a lot of wasted kinetic energy for just one car to cross, but timers and sensors guarantee static waiting times for the single car - something that a 2 way stop cannot guarantee.

Noticing that a car has activated the sensor is how I manage to catch most lights green - knowledge of the programmed sequence and line of sight of each of the 8 sensor sections means letting gravity slow me down at just the right time.

When a red light has been modified to interface with the approaching vehicles smart phone GPS and cruise control - "herds" of traveling cars can have their speed set in order to accommodate cars from lower speed capillaries without ever coming to a complete stop.

Here in Reno, a walk up one of the hills surrounding the city at night can lend a birds eye insight into the "herding" concept (large sections of the city can be seen without obstruction, headlights make crowd dynamics clearly visible). Speed limits and fear of enforcement generally means vehicles will cluster up. This pack traveling at their known speed (especially when giving consensual control to what we can call "Herd Cruising") can get through green lights while reducing wait times for perpendicular traffic utilizing the original sensor system. This same traffic could bypass the light similarly if the system already knows about an approaching bulk of participating "Herd Cruising" vehicles, although capillaries would mean cars entering the road and less likely to have time to herd together.

This could increase fuel economy, decrease waiting/idling time, save brake pads, and reduce emissions.

Maybe we should focus on finding a way to ... clone our brain inside of mechanical box. I know plenty of ethical considerations there, but a human with all the life skills of a human would be a perfect AI and not need the normal learning that focus ML brings w/ it. Imagine if they just clone a driving instructor's brain, and placed copies of that inside the cars...

Imagine being given immortality but you have to work a day job as a car and you don't get time off ever.

I know..but you could have time off... I mean you're not driving 24/7...maybe when not needed for transport you live in VR space.

You could also be turned off and reset. Why waste the power? Who will know? Is it even a crime to kill a simulated person?

Yes. If you buy into the belief that we're living in a simulation. It's a crime if we murder, therefore it's a crime to murder a simulated person.

This is not technically feasible for the foreseeable future and ethically fraught, but a fun thought experiment.

Hey, SOMA is a game you would enjoy!

Interfacing a brain sounds quite complicated. Instead, the mechanical box could just be a human body since we have over 1000s of years interfacing that with various vehicle controls.

We could probably harvest these fully functional units in less economically productive parts of the world, like Africa.

Even if we did this, we still don't know how to interface with the brain. Brain Computer Interface (BCI) technology is still very bad. We can barely do anything.

I recall that using EEG, you can extract at most 4 degrees of freedom/conscious-control, beyond that we can't do anything. We also kinda suck at monitering any "human intent" using EEG because there's no clear biological signal for that either.

Most of the successes seem to be done with some direct-to-brain interface, but those aren't magical either, and no one wants to do that.

I unfortunately missed a bright grad student's talk on the state of BCI. I didn't get to hear what he had to say, but the only slide in his talk was the image of a garbage dumpster on fire.

Sounds like an episode of Black Mirror

Probably not possible. See: https://en.wikipedia.org/wiki/Embodied_cognition

Reply on news.ycombinator.com