Hacker News Reader

Comment by MegaButts | Hacker News Reader

Based on what I hear from those in the know, the perception is that it's decades away instead of years. And an even more complicated problem isn't whether or not they can make it work with considerable constraints (geofence, no highways, avoid several spots that are hard to handle, weather, assistance via remote operation when stuck), but how to do it profitably. The cars are expensive. The maintenance is expensive. The ongoing cleaning, charging, and prepping of vehicles is expensive. Constantly updating HD maps is expensive. Of course the engineers to keep everything running smoothly are expensive. And the small army of people you need to actually talk to and help customers - something Google has basically never done before - is expensive.

I love the idea of self-driving cars and I really can't wait until they're ready. But I'd say that's going to happen about as quickly as changing the infrastructure in this country. Personally I'd rather have better public transit anyway.

View on HN

Yes,

The thing about self-driving you can get to 90% or 99% or even 99.9% success and then find the cost of closing part of that extra sliver of non-success is the same as all the work you've done. Especially since testing those corners cases is hard since they are unusual. There's even some part of the sliver that even is "AI complete". Very small but you don't know where it is, etc.

But I'd say that's going to happen about as quickly as changing the infrastructure in this country.

Key point!

Automation has traditionally been much, much easier when you are engineering a situation where you control all the variables. Automated factories are common. Automated warehouses are coming. Automated dock systems exist. In these situation, there's a hard border between "inside" and "outside". Roads are generally regular but the border between insight and outside fuzzy. Google can run it's self-driving taxis in a suburb in Arizona where roads are as regular as could be imagined but expanding beyond that is much, much hard than one imagines.

Note: I recommend this blog which periodically tracks progress in self-driving cars as well as making intelligent, critical comments on current AI.

https://blog.piekniewski.info/

Good points. And I'd add that most things we've automated are things that have been routinized in a way where people conform to the system, rather than the other way around. E.g., a ton of software development has basically been, "take this paper form and make a computer do the paperwork". The hard part was already done for us when everybody got trained to be subservient to the bureaucratic process. So societally we have this limited sense of how quickly computers can take something over.

Real-world things like cars and domestic robots are much harder, because those domains are ones where people come first. As you say, key elements could well be AI-complete.

Even when you control all of the variables, there are plenty of situations where the automated system just runs out of runway, and kicks the can to a human. Can't read the label on this package? Push it to the exception lane, and have a human figure it out, and relabel it. Works great, but is useless for a car, because when the car hits the exception, it's 10 miles away on a residential street, and it's not going to make it to you anymore.

For me, the most important thing a car can be is reliable, and every form of unattended driving cannot hit that level.

Sometimes the taxi you called never arrives. It happens regularly.

Of course the real problem is economic. Competing with the taxi service that uses humans is hard, because human labor is cheap.

>people you need to actually talk to and help customers - something Google has basically never done before

As a google advertiser this made me laugh.

At the same time I don't think expense will be a setback for autonomous vehicles. The reason being that if vehicles are fully autonomous there would be no reason for any individuals to own them. It would make more sense to have them operate like a taxi service. This is because without the costs of having to pay for a driver, it would be cheaper to just order an autonomous vehicle whenever you need to go anywhere than it would be own your own vehicle.

Just from an efficiency standpoint it is terribly inefficient for each person to have a car. We're not driving most of the time and the only reason we have our car sitting in the driveway is we need it to be available should we need to drive. With autonomous cars, my car could on it's own drive across town to pick up other people when I'm not using it. The reason we can't do that today is because of the additional logistics and costs involved in transporting my unused car from my driveway to the driveway of the person across town who needs it. This problem would not exist in a world with autonomous vehicles.

To me, it seems likely that a world with autonomous cars would be a world where the majority of vehicles in the world are owned by services like uber.

>> At the same time I don't think expense will be a setback for autonomous vehicles. The reason being that if vehicles are fully autonomous there would be no reason for any individuals to own them. It would make more sense to have them operate like a taxi service. This is because without the costs of having to pay for a driver, it would be cheaper to just order an autonomous vehicle whenever you need to go anywhere than it would be own your own vehicle.

I have a better idea that we can implement right now without waiting for autonomous driving to become a real technological capability.

We could make big cars with many seats and have them driven around town by humans- just a single human could drive a single vehicle capable of transporting 50, maybe 100 others. We could have those big cars stopping at predetermined points around the city, so that people would know where to get on and off. And we could support that service by asking the users to pay a small fee upon boarding the big car. This would cover the costs of the driver, and the big car, and still leave some to pay for infrastructure, etc.

These big cars would probably have relatively restricted areas of operation, but, for the foreseeable future, so will autonomous cars (which have to be geofenced) and, like I say, we already have the technology for those big-cars-with-many-people-sitting-in-them-and-a-single-driver that I'm talking about.

And since a whole bunch of people would be carried by one or two of those big cars per day, instead of each riding a single (autonomous or not) car, we could drastically reduce the number of automobiles-per-person, and, therefore, the CO2 emissions, thereby reducing pollution and protecting the environment to boot.

So? Anyone want to invest on that?

Getting off topic here but actually I would love to invest in transit. And judging from the kickstarter in the link below so would many other people.

https://www.cbc.ca/news/canada/toronto/crowd-funded-bus-take...

Unfortunately in my city (Toronto) and I would assume in many others it's illegal to run a private bus service.

https://www.cbc.ca/news/canada/toronto/liberty-village-shutt...

So we're stuck with the government monopoly burdened with the enormous costs of public service transit unions with some transit union workers making six figures for things like being bus drivers or ticket collectors. Then you have bus routes that are not decided by numbers business purposes but by politics. Finally you have rude drivers and buses and trains that smell like urine. And you have buses and trains provided by special favored contractors that are technologically and mechanically less than what is available. My aging mother says it hurts her body to ride the bus because it doesn't provide a smooth comfortable ride.

So if we want funding into buses and trains we can start by dealing with the political quagmire that has made buses and trains inefficient and unfriendly to users. But the familiar government cry of "just keep throwing more money at it until the problem is solved" is likely not the best solution.

> transit union workers making six figures for things like being bus drivers or ticket collectors

You say that like it's a bad thing. We also have "tech workers making six figures for things like increasing engagement on social networks by 0.1%." What do you think about that?

> Finally you have rude drivers and buses and trains that smell like urine.

This would change right-quick if congestion pricing pushed wealthier people onto public transit.

I have utmost respect for anyone who can safely and responsibly drive a huge vehicle full of people, day in and day out, on cramped city streets with precision down to centimeters on some complicated turns and crossroads.

If anything, good bus drivers aren't paid enough.

Do you know what the disadvantages of your vehicles? they don't take riders to their house. They don't allow riders to bring along any oddity of luggage. They aren't available at the moment notice. They aren't personal escape room when one is needed. I'm sure you know what kind of vehicles providing such comfort.

So take the high moral ground all you want, be snarky all you want, but the real world is a strange and messy place. Car has it place, along with bike, scooter, planes, trains. Until teleporter arrives.

People who rely on public transit adapt to public transit.

For example, they become more savvy about how much stuff they own and what they carry around.

I adapt by taking Uber/Lyft instead. I got fed up with crazy people threatening and attacking me on public transportation. Thankfully, I'm fortunate enough that I can afford to be driven around. I hope that in the future such a pleasant experience can be made available to everyone in the form of autonomous vehicles.

>> I got fed up with crazy people threatening and attacking me on public transportation.

As usual, this seems to be very different between EU and US citizens. I've been using public transport of all possible types (trains, boats, busses, trams...) in more than half a dozen EU countries since I was a wee bairn and I've never been threatened or attacked on any of them. Like, not even close. The absolutely worst thing that has ever happened to me on public transport is some guy puking his guts out on a London bus on a Saturday night. And that was just once. And it was more like "what an idiot" than "oh god I'm scared".

Which tells me that this is nothing to do with public transportation, as a means of transport, but with - other issues, for example various social issues that are particular in the US.

Or people are just exaggerating the misery of taking public transport in the US? I struggle to believe it's really that bad as you say, for instance. Public transport descriptions by US citizens always sound a bit like The Warriors, to me, and a bit unrealistic, to be honest.

Taking a driver to their house sounds like a luxury, not a real necessity. Same for having the car available at a moment's notice, except for emergency situations for which there are emergency vehicles.

Carrying luggage on public transport is perfectly possible and for larger or more items it's always possible to hire a van.

I don't understand the "personal escape room" comment. Escape- from what? In a car? How?

And note that even if all of the above were true limitations of public transport, the benefits would still significantly outweigh them. Just the reduced emissions from multiple people using the same vehicle instead of each their personal vehicle, would be (are!) an enormous advantage.

The fact that everyone has their own car and drives it every day has caused serious problems, not least of which is environmental destruction on an unprecedented scale from the necessary infrastructure and emissions- and that's before we count the number of deaths from accidents and pollution. We need technologies and solutions that take more cars off the roads- not ones that put more on them.

Driving in the car is one of the few times one can be completely alone with one's thoughts, for many people. I think I'd lose my mind without being unplugged for an hour or so a day.

Hunting and fishing isn't bad either, but I've got to get up at 4:30 these days to get out there before the turkeys fly off their roosts...

I think you're right in that mass transit is the best solution to transportation from a social and environmental standpoint. But the reality is that most people prefer cars. It's unrealistic to force people to use the socially optimum solution rather than their prefered solution. Many governments have tried and many have failed to force people to use the socially beneficial solution over their preferred one. A good example of this is the failure of carpool lanes.

Any system that caters to humans has to account for what people want. The way to make people take transit is to make transit comparable to cars - in terms of speed, reliability etc. And to make it better now rather than holding out some promise that it will be better at some future date when enough people take transit. People will take transit now if its better now, not if it promises to be better in the future.

I mean at the end of the day we can't force or guilt people into taking our preferred solution. We have to create a solution that caters to them. People will do their part for society and the environment - recycling programs are proof of this. You just have to create a solution that's not too much of an inconvenience. People would take transit more if it was cleaner, faster and more reliable. Rather than pouring more money into it hoping it gets better, or goading more people to take it under the promise it will get better - I don't think its unreasonable to ask to make current systems more user friendly and then increase funding as service improves and ridership increases

> It's unrealistic to force people to use the socially optimum solution rather than their prefered solution.

It is perfectly realistic and has happened, like when schools were desegregated much to the dismay of white people in the south, who abandoned the democratic party as a result of it. The hard part is getting a law passed and not diluted along the way by people who refuse to accept inevitable change. Making transit comparable and better than cars is the goal of every transit agency in the world, the difficulty comes from getting people to vote for funding and approve construction projects. Los Angeles is pointed to as a slow meandering failure of a transit build out, but the only reason why it is a slow meandering failure is that the people who would benefit from transit the most do not vote, and the people who vote care about their view or have a preconceived and bigoted notion of who rides public transit, latching on to psuedoscience along the way to further their racist worldview into political action (underground subway construction was banned in LA from the 90s to a couple of years ago due to an unfounded fear of blowing up the city, for example, but the real reason was that the concerned citizens of beverly hills felt they might see more black or brown people on their stretch of sunset boulevard).

>> It's unrealistic to force people to use the socially optimum solution rather than their prefered solution.

That sounds ...wrong? You might as well say that it's unrealistic to force people to pay taxes. If the benefits of using public transport outweigh the disadvantages, people will just have to get used to the idea.

This has certainly worked in many other cases- for example, with rules about smoking indoors etc. I'm sure that smokers prefer to smoke indoors. But there's good reasons not to, and a very strong push with fines and all to not do it, so they just have to suck it up and conform. Sad face.

> I don't understand the "personal escape room" comment. Escape- from what? In a car? How?

My car is my second home, one that I can bring with me anywhere. It's why I prefer to drive long distances rather than fly. It even saves on hotel costs in the summer.

It's a place to escape from "out there". When I worked in an office, I would take my car to get lunch, then spend the whole hour absorbed in my thoughts or reading a book, away from my co-workers.

I'm a bit of a loner, but I can say that I've observed similar behaviour in others. After a many hour hike in the forest, getting back to the car in the evening is a relief. It's comfortable, familiar, and safe.

Sure, I could get used to public transit. But as long as I don't live in a dense city, I'll keep my car, and love it.

Door to door transport is an individual perk but a network wide con. Imagine if american airlines was tasked with bringing you to your door or hotel rather than just asking you to find your own way to the airport hub, or if there were no connecting flights at all. It's way simpler to just ask people to walk (or bike, or scooter, or skateboard) a half mile (10 minutes, I'd do it in a blizzard), than to serve all those random last mile connections. Available at a moments notice is also another con, better to wait for a ride every 10 minutes than have the network strained by serving minute to minute stops. Plus I don't think I've had an uber arrive in 10 minutes ever.

The last thing, with the luggage, well try going to Portland, one of the only cities I've been to where you can ride a train from your downtown hotel to within ~100 yards of airport security in 20 minutes. People rarely are checking more than one bag, and you can strap your carry on to the handle of your roller bag and it isn't a strain on the system, as hard as this might be to believe. In Los Angeles, people find room to set up little potato chip shops on the trains, and there are people who manage to transport all their worldly possessions with them on the subways in NYC. And if you really needed to haul a lot of junk, you can rent a uhaul box truck (or a tall van, or a pickup truck) for $20 and change plus gas for a local trip.

Buses are great but cars are 2x as quick to any destination more than a few km away (at least in Singapore). I don't own a car fwiw.

Buses really need to be grade separated from traffic, but most cities are reluctant to remove a lane out of a road to make it possible. I believe rio de janero has a huge BRT (bus rapid transit) network. Grade separated, and they are pretty much a subway with rubber wheels and it's waaay cheaper to build even an elevated roadway vs. a tunnel with rails that is more constrained with how sharp you can turn and what elevation changes your track needs to avoid. In Los Angeles, BRT busses are twice as long as well, all the more riders in one go.

You would need to split subway-like buses and less frequent suburban buses. Suburban buses need to navigate streets.

I wish they also did this for bicycles and e mobility devices

I doubt it. Relying on taxi services (driven by humans or computers) might work fine for childless urbanites. But I keep a lot of stuff in my car, and there's no way I'm going to waste time loading and unloading a taxi every day. Plus my family likes having a nice clean car. I'm never going to give up convenience and time savings just to save a little money.

I fully agree with you. Another argument against this ride sharing that most people conveniently forget are peak times. How many millions of vehicles are standing still during office hours only to power up when everyone needs to go home? How is a robotaxi service going to solve that? The number of cars on the roads at peak times won’t go down much imo.

Peak times are already constrained by road capacity.

That said just having smart buses picking crusing and picking up people when and where they are and dropping them off at their destination would be great. (After all usually a lot of people want to go from similar A to similar B places at the same time.)

> but how to do it profitably. The cars are expensive. The maintenance is expensive. The ongoing cleaning, charging, and prepping of vehicles is expensive. Constantly updating HD maps is expensive. Of course the engineers to keep everything running smoothly are expensive. And the small army of people you need to actually talk to and help customer

I believe that only applies to the waymo model where you go with the full time robo-taxi using top of the line hardware like lidar and which pretty much requires level 5.

The tesla model is to use cheaper hardware which they have proven is cheap enough that consumers will pay for. I believe they also only need it to be level 3 [0] for it to be a massive success (I know a ton of people that would pay the $6,000 for the FSD upgrade if it allowed them to watch a movie/work/text during their commute most of the time).

Granted, Tesla's tech has a lot to catch up but I wouldn't count their solution to the profitability problem out since they have a lot of more cars on the road collecting data/testing their algorithms, and I'm guessing they have a solution to the imperfect tech that ends up killing people problem (Selling their own auto insurance which they have already announced plans for and which will probably have special conditions for autopilot failures, and lots of cars running their quasi level 2 tech which they can use to lobby congressmen to limit liability for non-tesla drivers suability during autopilot failures).

But, as with all big tech predictions, who knows.

[0]: Level 3 ("eyes off"): The driver can safely turn their attention away from the driving tasks, e.g. the driver can text or watch a movie. The vehicle will handle situations that call for an immediate response, like emergency braking. The driver must still be prepared to intervene within some limited time, specified by the manufacturer, when called upon by the vehicle to do so: https://en.wikipedia.org/wiki/Self-driving_car#Levels_of_dri...

It doesn't seem like Tesla knows how to make it profitable either, to be honest. Their current plan for profitability appears to be the million robo-taxi fleet by 2020 rather than the auto manufacturing business.

I think it would seem insane for them to present this opportunity to make a ton of money (robotaxis), and not fully commit to it.

I mean that they are positioning themselves as if they truly believe this is going to happen.

What other options do they have? It seems like they lost the bet that they were going to become profitable with the Model 3. Now they need a new big bet, or it all falls down.

It does not seem like that to me.

They have to stop pretending they are a luxury brand and release a cheap tiny little box with knobs instead of an ipad that will compete with the leaf or any other subcompact. The price is just too high right now, and if they can't get it lower then it sucks that they entered this market too early to benefit from the economies of scale on EV tech that might come years later. Any other automaker has other revenue streams to wait decades even for this to happen, but Tesla has 10 months to live.

But there is no way that they can get the from-scratch design, engineering and production line of yet another model up to capacity in less than a year. And everyone knows this. OTOH, making bets on AI are much harder for investors to evaluate the reasonableness of.

> They have to stop pretending they are a luxury brand and release a cheap tiny little box with knobs instead of an ipad that will compete with the leaf or any other subcompact.

Believe it or not, the screens are cheaper. Other companies (most notably Volkswagen) are moving towards entirely digital interiors as well. They cite reasons for it being a better experience despite decades of research pointing out the importance of touch controls while driving to limit distractions, but most people believe it is because they are cheaper to build and service.

A screen is cheaper than a knob? I can see it easier to remove a screen vs. a knob, but how often does a knob go bad?

Pretty much every manufacturer is already selling tech similar to Tesla, they just don't call it autopilot but driving assistance. If that's our goal we could just stop researching self driving cars and put the money into EVs instead.

Thing missing from the discussion is the different safety and threat model fully autonomous cars have compared to human drivers.

Mass incidents are one of them.

* If one autonomous cars fails catastrophically on some road, all other cars with same model and/or software are likely to do the same. Huge pileup in fast tragic can become nightmare for the car maker/operator.

* If there is a software error in the overnight update or planned sabotage, hundreds of thousands of cars can get into accidents in very short time period until the failure is noticed.

* Simple loss of trust due to security incident, discovery of exploit, server configuration error, hacking in the network can lead to security groundings of millions of cars at once. If all Toyota's are grounded for a week, the economic impact around the world is huge.

Even if the autonomous driving works, designing the security and safety infrastructure that keeps them working will be real challenge. I'm sure it will be worked out but it's not going to happen in few years.

The amazing thing to me is that I don't anybody who thought it was coming in the next couple years who based their estimate on anything but hope and magic.

Of course, that's nothing new. We've all been on software projects where the launch date was the earliest day nobody could prove it was impossible to finish. But I'd like to think that investors dropping zillions on companies would have had better due diligence.

The amazing thing to me is that I don't [know] anybody who thought it was coming in the next couple years who based their estimate on anything but hope and magic.

I'd agree that a bit of the proper sort of analysis would show that this wasn't easily happening. But I think the kind of thinking that imagines this as possible is extremely easy to fall into.

We've all been on software projects where the launch date was the earliest day nobody could prove it was impossible to finish.

Sure and I'm pretty sure a lot of software engineers have given or agreed to completely unrealistic estimates after having had this experience. If anything, the "software crisis", the inability of software engineers to provide good estimates even after hard experience and warnings is an expression that over-optimistic thinking is natural for human beings in this situation.

My theory is that in situations where a human is asked about the difficulty of a mental task, they essentially can only fall back on a basic language-expression of said task and rationally try to calculate with that - the brilliance and the limitation of we language-generating, language-using creature is the ability to correctly distill a complex and messy task into fairly simple sequence of symbols. But that correctness actually exists in a complex and dynamic context, most of which we screen out. A command like "Drive to the store" doesn't take into account all the compensations needed to do the driving just as "Create an inventory system" fails to specify all the client-specific "gotchas" involved in such a task.

So going from our normal language-distillation process to a complete software system that doesn't have the continual dynamic compensation of our intelligence is an inherently difficult problem - even though we can probably only do it at all because of our "language facility".

> the inability of software engineers to provide good estimates

I don't think it's inability. I think they're embedded in a system that doesn't encourage or reward honesty or accuracy.

I agree it's easy for people to say, "Gosh, how hard can that be?" Which is where we get the classic, "I could build that in a week" estimate. But I don't agree that there's something inevitable about people going from a finger-in-the-air SWAG to a "we'll have self-driving cars in 2021" business plan.

Normally I'd say that it's a straight up failure of both management techniques and the professional standards of the engineers. Which is true. But I think the problem is that a lot of fundraising in the "Uber for X" era is only slightly less of a con than Theranos. So I think it's more correctly seen as a failure of founders and VCs.

Sure,

At some point, when you're talking about a concrete and complete plan, you likely go from naive self-deception to the willful deception of others.

Of course, one should consider how people's tendency to naive self-deception makes things easier for willful deceivers.

As a scheme for selling an impossible dreams takes shape, I suspect the actors have a strange and contradictory mindset. A rational conman is going to sell a smallish scam and vanish. The folks running Theranos certainly deceived many but they also rode the train far past the point where they could avoid being caught in the wreck (so they wound-up ruined, looking at jail time, etc). Maybe they had belief that if they kept the charade going long enough could make the original scheme work but they could also have a sort of thinking that simply doesn't look at the possibility of failure once they have decided to seek success.

I fear the unavoidable cost is going to ruin the model. The advantages of these things is that they are able to work as a unified network, keeping space, allowing space, preventing pileups from simply driving too close, etc.

What about all the 30 year old $2000 corollas you see on the road? Are these people expected to shell out $50k+ to get a self driving car? Absolutely not. There will always be people who can only afford a 10, 20, 30 year old car. What happens when they own a 30 year old self driving car, outdated for decades and based on how tech companies operate today, probably unpatched and unsupported for 20 years?

Self driving would really shine as a self driving bus, in a dedicated lane. Single occupant vehicles are irresponsible in cities that are choked with too many of them, but there is no alternative if all you can afford is a $2000 corolla and rail connections to work are nonexistant. But then again, a self driving bus is much less sexy to investors than a self driving car...

That sounds like one big company doesn't know how to do it faster by keeping doing the same things they're doing now.

>Constantly updating HD maps is expensive.

Do any of the players in this space use HD maps?

Waymo.

Seems like autonomous personal air travel would be a far, far simpler problem to solve than road-based driving? Far less edge cases, once you are in the air the only obstacles you really need to worry about are other vehicles, birds, leaves and the odd stray plastic bag. With VTOL the whole thing could be an order of magnitude easier than driving. By the same principle, sea travel would also have less edge cases, at least in calm conditions. Is it the case that people aren’t seeing markets for those applications of autonomy or is there some other reason why there isn’t the same hype in those areas?

Vehicle cost; fuel cost; takeoff and landing; handling of edge cases (there's a reason that commercial airlines still have pilots despite autopilot being very effective in 90% of conditions).

how expensive would it be creating areas of road that only allows autonoums vehicles, something like carpool lanes we have now.

Wouldn't that cut problem size massively.

I can see that happening on major freeways, the way we have carpool lanes now, but not on a general scale.

A: Air travel generally is a far simpler problem to solve, because there's national standardization of the air navigation system, the airport paint and signage, and very clear rules of separation with very few edge cases. And yet we do not have anything close to self-flying planes no matter how much money you throw at it. And we're not even close to getting there. Whereas with cars, none of that is true, municipalities get wedged into violating their own equivalent of human interface guidelines all the time. Worn paint and signs for years, busted cross walk signals, confusing intersections, human driven cars that consistently do not follow the rules but in inconsistent ways.

B. Personal air travel has a significant regulatory burden in the transition from ground to air. The ground is city + state regulated. And immediately once airborne it's FAA regulated, but not under any kind of Air Traffic Control, as almost all airspace below 1200' above ground is uncontrolled. So ATC has nothing to say about it, and no central mechanism for negotiating conflicts.

Where are you allowed to takeoff and land is easily figured out today: airports only. And that's because it takes all kinds of things into account like obstruction clearance, noise abatement, anticipating engine failures and crashes. We can't even automate commercial flights - it might seem like it's more correct to say we don't automate commercial flights, but we could. That's not really true. The amount of changes to the air traffic control system, and on-board equipment for airplanes, is presently so cost prohibitive that it is effectively a "cannot be done".

TBH this is the pareto principle at work.

Sure you can get a car roughly driving around the city autonomously (there are basic self driving cars back in the 1950s that drove on a wire track) but getting it perfect is where the vast majority of effort is going to be expended.

> I love the idea of self-driving cars and I really can't wait until they're ready.

What if I were to tell you that you can buy a self driving car right now, and that there are 100s of thousands them in consumer hands, right now?

Tesla sells them. And they aren't that expensive.

If you can press a button, and take your hands off the wheel, that is a self driving car. And you can buy them today.

They are still working on the do not kill the driver by crashing into stationary objects feature though it seems.

Sorry, nope. If your car crashes after you press that button and take your hands off the wheel, Tesla will point to the line in their terms of service which says, "keep hands on the wheel, driver must pay attention at all times."

Tesla (and Elon Musk in particular) are of course happy to encourage the misconception that they already sell a self-driving car.

You can call it whatever you want. But at the end of the day, this definitely fits the vast majority of people's definition of "self driving".

You press a button, and it stays in lane, at the right speed. Most people are happy with that feature set.

That is the MVP, that provides most of the benefits to most people, even if it isn't perfect.

It kills people... it's not ready.

Most people think of "self-driving" as allowing them not to pay attention to what the car is doing. Autopilot is emphatically not that, even in the limited highway settings where you can actually use it.

That you can get from any number of other makes, except that they do not slam you into firetrucks, and don't call it self-driving when it obviously isn't.

Can I get in the car, set destination, then sit in the back and watch a movie with my feet up? Can I call my car to me when I'm in LA and the car in NYC?[1]

If not, it's not self driving as generally understood, it's just driver or lane assist. A perception intentionally created by Tesla promoting their driver assist as giving a car that could effectively wander off on its own. Maybe take a part-time job as autonomous taxi. Long having offered full self driving in the paid options list.

[1] https://twitter.com/elonmusk/status/686279251293777920

You don't need those features that you describe in order to be a self driving car.

For most people, the MVP of a self driving car is merely "able to take your hands off the wheel".

That's such b.s. Tesla explicitly says you can't take your hand off the wheel. So what are you even talking about?

"I'm going to drive to the store" means a complete sequence of events, start>unpark>integrate merge with traffic on route>follow route>navigate parking lot>park.

Nothing other than a human does that today. That is what is meant by either "self-driving car" or "fully autonomous". All exceptions to what a human can do is just making excuses, and it's b.s. Driving is a broad term that encompasses all responsibilities for getting from A to B.

Does a Tesla lane change automatically to follow a route? Does it enter turn lanes and turn onto new streets? Does it turn into parking lots? Nope. Human prompted lane changing, and staying in a lane, is not "driving", it's a vertical feature that partially offloads some task from a human.

Reply on news.ycombinator.com