Hacker News Reader

Comment by roymurdock | Hacker News Reader

I have a $100 bet going with a coworker at the office.

We defined self-driving cars: When I can pull out my phone in any of at least 5 major metropolitan areas around the world, order a driverless car, have it pick me up and deliver me to a specified location within the city in a timely fashion, with as little risk as getting in a taxi.

He is sub 20 years. I'm thinking more like 40-50 mark.

My thesis? We can solve the technical challenges, but so much physical, legal, and regulatory infrastructure needs to change to make this viable that it will be more than 2 decades until we see self-driving cars as a reality across multiple cities. Of course, I would be more than happy to lose the $100 as it would make everyone's lives better that much sooner. But I'm skeptical.

View on HN

> He is sub 20 years. I'm thinking more like 40-50 mark.

Perspective:

* 60 years ago, we didn't have highways.

* 45 years ago, seat belts weren't required.

* 40 years ago, we expected engines to last 10s of thousands of miles. We didn't really understand crumple zones. Open containers and even drinking while driving.

* 30 years, we added fuel injection, computers, and airbags.

* 20 years ago, the hybrid car, better fuel efficiency, and side airbags

Imagine what we'll know tomorrow.

Perspective: nearly 50 years ago we did have Category IIIa instrument landing systems in commercial jet aircraft capable of landing without pilot input. We had reliable gyroscopic autopilot capable of keeping aircraft on a set course long before that.

Vast sums of money have been sunk into developing exceptionally reliable avionics systems since then.

Passenger carrying aircraft still all have a pilot or two.

Pilots on modern large aircraft are more like managers that tell the aircraft what to do, and do fly them to keep with minimum legal requirements (minimy numbers of landinds etc...)

And in the US, don't pilots still have to do the landing part?

And to complete the thought: the main reason why human pilots are still on all those aircraft is because... the automated/computerized systems still can't really make decisions that match what the humans can do. So the computer is really great at maintaining heading, altitude and speed for a set period of time, but really bad at deciding which heading, which altitude and which speed, and for how long (the best you can do that I'm aware of is have a human lay out the desired route and punch that into the FMC, and then all bets are off if conditions change along the way, and the humans will have to decide what to tell the plane to do again).

And those decisions themselves could be easily automated, but people choose not to. There's a checklist for making those decisions. If you can write a checklist, you can write code. And for when there's not a checklist, then you can pick randomly, because that's what the human is essentially doing.

A human can notice when new evidence becomes relevant. If something isn't already on a computer's checklist when it becomes relevant, the human has a better chance of deciding what to do.

EDIT: I understand that there is research going into making computers recognize "novel" situations. My comment only applies to algorithms that contain "checklists".

No, then you need to get meta. There's a checklist for identifying "novel" information and reacting. By checklist I mean algorithm.

Sure, it can be tough to tease out the algorithm from the minds of current pilots. Luckily in aviation we already have training manuals. In other domains it'd be tougher to know when we're done extracting knowledge from humans.

How close we are to the autopilot which would safely land a plane with engines off on the Hudson?

a) A human pilot who does combines exceptional luck with unusual skill

b) Said human pilot was aided considerably by the airplane's autopilot system

c) I'm not sure why you think software would find this particularly challenging?

It's not the landing that's difficult, it's the "write an algorithm which correctly identifies which specific parameters should cause the aircraft to choose a river near the airport as its landing site, and get approval for it" that's the difficult part.

How often does it happen, though? You don't need a human on every flight in order to deal with that. Build a few ground-based emergency remote-control centers across the continent, and train the autopilots (or the air traffic controllers) to call them when things get bad.

I'm not convinced anyone in the commercial aviation industry would agree with you.

Some on-ground duty pilot sitting in a simulator who's suddenly placed in control of an aircraft that's notified an anomaly seconds before landing has absolutely zero chance of being able to make the same judgement call as a human (co)pilot who's been sitting at the controls for the last few hours. And the likelihood of human input being required is highly correlated with the possibility the aircraft has lost connectivity and/or some of the equipment that's supposed to be sending back data is malfunctioning (and yeah, you could build in more redundancy, but that still doesn't eliminate the problem and probably costs more than the pilot)

I agree in a sense, but one thing to keep in mind is that all automation requires a large investment, especially in systems that cannot fail.

If the cost of that automation (programmers, testing, etc) is greater than the cost of training pilots, then pilots are going to remain.

Usually, there ends up being a middle ground - you automate the low-hanging fruit, and leave the last percentage points to the humans.

The exact opposite has happened to cars. We've delegated route building to our GPS devices, and we maintain heading and speed until the computer tells us to do something different.

So you're agreeing with GP poster that the legal objections are what will prevent self-driving cars on a large scale?

"60 years ago, we didn't have highways."

Not quite. Here's the US highway system in 1955.[1] Construction of the US Interstate Highway System started in 1956.

[1] http://www.mappery.com/map-of/Pre-Interstate-US-Highway-Syst...

He probably meant Interstates, and 1956 is 60 years ago, which is why I'm guessing that. Interstate 5 (Seattle portion) was only completed in 1969[0] so depending on where you are, it could be significantly less.

[0] - http://www.historylink.org/index.cfm?DisplayPage=output.cfm&...

Good point. I-80 wasn't complete coast-to-coast until 1986. There were gaps in the Western sections for years. That was the first coast-to-coast Interstate, the old Lincoln Highway route. I-40, the "Route 66" route, was completed even later.

All these were some sort of win win, not hard sells, low resistance. SDV are rotating the notion of driving and the economy around it, car dealers, oil companies. It won't be the same game at all in my view.

> 60 years ago, we didn't have highways.

What?

> 30 years, we added fuel injection, computers, and airbags.

But we knew about them for a lot longer than that. Fuel injection was used in aircraft during WWII.

>so much physical, legal, and regulatory infrastructure needs to change to make this viable that it will be more than 2 decades until we see self-driving cars as a reality across multiple cities.

It could be a lot more straightforward than you suspect. The three things we have going for us:

1. Car accidents are almost always resolved by deals made between insurance companies (this will be doubly true with unpiloted cars, which will presumably not commit criminal acts)

2. The same few large companies are likely to deal in insurance of piloted and autonomous vehicles, so there's not a huge incentive to go sue-crazy

3. If autonomous vehicles cause significantly fewer accidents and companies can charge the same amount for insurance covering them, that is a huge win for some very large companies with the power to get the ball rolling.

Increased vehicle safety will be a big win for these companies in the short term, although in the long term some price discounts might be factored into the cost of autonomous vehicle insurance.

I have a bet going on with my wife that our 1-year-old daughter will never take driving lessons when she grows up.

I'm hoping there's a viable SDC on the market when my eldest daughter turns 16 in 5 years

Do you live in the suburbs? This could become true, but more because of the increasing urbanization trend making public transit more attractive than cars in more places.

We haven't exactly spelled out the terms precisely, but the spirit of the wager is that self-driving cars will be involved somehow. Certainly public transit could be part of a mix of transportation options, but I wouldn't win if self-driving cars don't come to fruition yet she doesn't get a driving license because she can subsist entirely on public transit.

I have the same plan for my four year old!

Call me a technology optimist. I'd bet 10 years.

You want to get in on this bet? We could open up a pool on the prediction markets.

Probably not. I don't think it's sufficiently well-defined for me to trust an arbiter to resolve the bet satisfactorily. Also, time value of money.

http://longbets.org/ would help facilitate this, though the minimum is 200 dollars.

I'd agree. I think in 10 years we're starting to have the conversation about when human driving should be banned in metro areas. At that point it'll be obvious that the technology has arrived, and the safety record will be there, but not everyone will have the technology. There will be the problem of telling someone with a jalopy that they need to park that outside of the metro and pay for an Uber to their destination which will probably snarl that up for at least a decade or more... At some point, though, as rich people are zipping along in cars that never get pulled over for speeding or having taillights out, the "Uber tax" is going to be calculated to be less expensive than the "PoPo tax" on the working poor.

I'm rather skeptical on that happening anywhere near 10 years. Even if the tech was ready to go full-scale tomorrow, just the process of tooling up to make tens of millions of these cars and actually selling them to people, letting them trickle down to lower-income used car buyers and letting the oldest manual cars trickle out of the market could take decades.

Once the first ones are available on the open market, I'd expect at least 10 years of them being very expensive rich peoples' toys while the remaining edge cases and kinks in the tech are worked out, manufacturing scales up, insurance and legal issues get sorted out, various road maintenance departments learn to deal with their needs, etc.

One reasonable scenario for rapid deployment of self-driving cars - is shared vehicles, driving a few people at a time over a shared but optimal route.

It's possible that such vehicles will have 5x-10x total occupancy per day of regular vehicles(which sit idle at most times). And it could imagine how the right financial incentives would shift a large share of the manufacturing capacity towards such cars. And considering the car replacement rate at probably something like 20 years - maybe 10 years is a possibility of starting to have that conversation.

That's an good possibility for getting self-driving cars to market sooner and getting more people in them all right, but I'm thinking we have to get a lot of manual cars out of the market before anyone starts taking the idea of restricting them seriously.

Rush hour patterns imply that a lot of that idle time would likely remain idle. EDIT: Fixed typo

I think 10 years is just a bit optimistic. I think it'll take about 10 years from the time we have the software/hardware at a point where it is ready. I think that could be 5 years. I think it'll be another 15-20 years after that where the majority of cars on the road are self driving. Baring any crazy legislation, which I don't think would happen, in the US anyway.

I argued that it'd be 10 years when we'd start the discussion, and that it'd be 20 years from now before metros start to go driverless.

Yeah, I apologize, I meant to respond to the GP and accidentally responded to you. I don't want to clutter the thread with duplicate posts though, so I'll just leave it as is.

Even if it were possible to go buy a fully automated car today, there's no way enough of the fleet would have changed over in as short a time as ten years to make such a ban practical. And it's not possible to buy an automated car today; nor are the research prototypes Google and others have been working with up to a level of sophistication where simply productizing them would yield a fully automated car which could be used as a generic replacement for current automobiles.

I could believe that the first generation of fully automated cars might be available for purchase in ten years.

> there's no way enough of the fleet would have changed over in as short a time as ten years to make such a ban practical.

Cars sit idle 95% of the time. That means you only need to flip 5-10% of the total fleet. Tesla could do it by themselves building 2-4 million cars/year, with the amount of cars needed per year declining the higher the utilization level you can achieve.

Yes, but they all sit idle the same 95% of the time. Everyone needs a ride at rush hour.

> Yes, but they all sit idle the same 95% of the time. Everyone needs a ride at rush hour.

Disincentivize rush hour using congestion pricing.

You could divide rush hours maybe to 2 shifts of an SDC. And there's always the possibility of shared rides.

I don't understand how replacing a small minority of the fleet with automated cars leads to a situation where it is reasonable to talk about banning human drivers.

You would always consider banning humans from an activity if it was overly dangerous compared to having a machine perform the operation.

40K people a year die in the US from vehicle accidents. Why would you not ban humans from driving if machines were objectively superior? Where do you find the "right" to drive to be? Driving is a privilege granted by the government, not a right.

The comment I replied to was suggesting that vehicle automation would progress so rapidly that it would be practical to start working on such a ban as soon as ten years from now. I'm not disputing that such a ban might eventually become a good idea, but for it to become practical so soon, we'd have to be much, much further down the road to full automation than we are now.

If fully-automated vehicles become commercially available (which seems likely at some point, though not in the immediate future), we will likely see rapid adoption in the applications where they are very well suited, followed by a long, slow, gradual trend toward broader usage. The hardest cases will be the ones sticking around longest, and one can't simply legislate them out of existence. Just replacing a small percentage of the fleet won't come anywhere close! It's only when automated cars have become so dominant that only a small percentage of the fleet still depends on manual drivers that it will become practical to start talking about a ban on manual operation.

So I argued we'd start the conversation in 10 years and that it'd take at least another decade. I that is 20 years until we'd see the first ban against human drivers in metro areas.

To be there I do think we need the first generation of fully automated cars available to the general public in 5 years or less, which I think we're actually close to or we wouldn't be having this very discussion...

Ah, I didn't notice the second ten-year span in your prediction. I still think your perspective is wildly optimistic, but it makes more sense now.

I think you or the colleague are more likely to be correct than 40+ years. There are hugely capable companies driving this and more understanding than ever on what will be achievable and how much money will be on the line.

Self-driving, electric cars will make future taxi-equivalents affordable enough that it's considered for many instances. That makes for a massive market.

It's not as though current taxis have a sparkling reputation as drivers either. Not many will be sad to see the end of them.

I think you are seriously underestimating what the baby-boomers will do when they have to start giving up their drivers licenses because they are getting to old to drive.

>> but so much physical, legal, and regulatory infrastructure needs to change

Let's go with that. Say that tomorrow Google finished the development of the car. It works very well.

Than it just need to find a single city that would agree to a wide scale deployment. And I imagine there are plenty of reasons why a city might want to do that. Heck if Google cared, they could legally bribe politicians. And it doesn't even have to be a city in the west. It could be a city ruled by a dictator somewhere in africa. Or it could be a self driving car created by a chinese company, and china would be willing to take the risk to lead in this field.

So maybe the first city isn't as hard as it seems.

Now what happens after a car has proven itself in a city for a year or two ?Showing almost zero accidents, and all the other benefits of a self-driving car. How much will people everywhere would want it ? how much political support would it gather ? i imagine a lot.

Would such support and proof be enough for relatively rapid deployment of self-driving cars ?

Careful, that $100 probably won't be worth shit in 20 years let alone 40-50.

You're right that there's significant legal and regulatory hurdles that have to be cleared, but I think they're more of an issue for broad adoption. SF will have a driverless taxi service probably within the next 5 years. Once that's in place it's only a matter of time/strength of the taxi lobby that's going to stop it, (and Uber has shown that can be broken even in places like London).

If you're talking about _everyone_ in major economies having a driverless car, yeah, I think 40 years is realistic given the replacement rate for cars alone, nevermind the legal issues.

I know what you mean, but London is a poor example — the well known "black cab" service has been heavily regulated forever, but cheaper "minicabs" have been available to book (rather than hail on the street) since the 1960s, and subject to a little regulation¹ since 2001. Most operators were tiny companies with a few cars servicing a local area, until Addison Lee started running a London-wide service about 10-15 years ago — they had nicer cars, better availability and marketing and so on. Uber set up in London under the same regulations.

¹ Additional car inspection, criminal record check on the driver, etc.

https://en.wikipedia.org/wiki/Taxicabs_of_the_United_Kingdom...

Humans overestimate tech advancement for 3yrs in the future, but underestimate it 10 yrs in the future.

Food for thought.

While I agree in general - I can think of a lot of technological predictions from my childhood in the 70s and 80s (fusion power, widespread off planet colonies, general AI) that we still seem to be waiting for.

Reply on news.ycombinator.com