Waymo’s autonomous vehicles are driving 25,000 miles every day | Hacker News Reader

Waymo’s autonomous vehicles are driving 25,000 miles every day

techcrunch.com

153 pointssahin180 comments

As an HNer, I find what Waymo (and other autonomous vehicle companies) are doing is simply fabulous. Technology improvement is a big impetus for human progress. Looking back over the past century, inventions like transistor, rockets, internet, satellites, nuclear power and aviation significantly improved our quality of life.

Autonomous vehicles belong to the same league. I am very happy that a non trivial amount of resources are devoted to them and it's not just a next SV fad like some ICO / Juicero / photo sharing app. Rooting for the success of this technology.

This isn't Reddit, no need to prefix an otherwise normal comment with internet credentials.

"As a mother..."

AA {Role} IW2...

Cars destroy the environment (with the infrastructure needed to accommodate them) and are one of the primary drivers of climate change. We need technology to make them obsolete, not to make more of them that can drive themselves (or not).

Thanks for playing Devil's Advocate. This technology applies to mobile transportation in general and it doesn't necessarily have to effect greenhouse gas emissions. Additionally, I believe the infrastructure for producing cars applies to such a broad category in regards to driving climate change. The manufacturing industry needs to independently solve this problem.

For the record, I'm not playing Devil's Advocate. I believe self-driving cars are a terrible idea that is only piling more environmental damage on top of what ordinary cards already do.

>> This technology applies to mobile transportation in general and it doesn't necessarily have to effect greenhouse gas emissions.

Well, I don't know that there is any way to drive a fleet of some millions or billions of cars (once self-driving cars have been adopted widely, that is) that doesn't rely on burning fossil fuels, either directly or indirectly (to charge batteries etc). So how are they not going to affect greenhouse gas emmissions?

So, you would like to just get rid of cars? Do you have any idea the level of human suffering that would cause? If you're not proposing something as silly as eliminating them entirely, then you should want to make them more efficient.

Won't autonomous driving eliminate something like 80 to 90 percent of private vehicle ownership, which is well on the way to obsolete?

Once people are less tied to private vehicle ownership and open to any effective form of transportation, then adoption of new technology could become easier.

>Won't autonomous driving eliminate something like 80 to 90 percent of private vehicle ownership, which is well on the way to obsolete?

We do not know that this will happen, people won't like using public cars. Also the number of cars is not important as the number of kilometers those car run, the companies will try the best to maximize the number of kilometers the cars will run to maximize the profits so we may have people using cars even more instead of walking or public transport

> we may have people using cars even more instead of walking or public transport

You're aware buses are large automobiles right? Lyft Line with self-driving cars is effectively a more nimble, efficient, usable, and cheap bus network.

Yes, can you assume so good faith from my part and let me know what part of my argument is not clear or what you disagree with.

My point is that if we get cheap or convenient cars to drive us around then this would increase number of kilometers done by cars.

About buses you have say 30 people using a bus and the bus is driving 10 km, if people would use cars and say they will be 3 people in a car you get 10 cars x 10 km = 100km, so you get 9 more vehicles on the road , you get the more energy consumed since a bus is not consuming 10 times more then a car.

My apologies, I'm a bit annoyed by comments everywhere on this discussion that don't count shared taxis as public transport. I wrote in another comment that without congestion pricing or aggressive expansion of HOV lanes, the efficiency gains of self-driving cars would not be realized. Even without that we'd still have the increased road safety gains.

> About buses you have say 30 people using a bus

The problem is that it's difficult to match transport demand and supply perfectly because the bus is such a large unit of transport. The more common scenario is a close-to-empty (off-peak times, unpopular routes) or bursting-at-the-seams (peak times, popular routes) vehicle. Because the self-driving car is a smaller, more granular unit of transportation, you can deliver more targeted supply to the places where there's demand, at the exact times when there's demand.

My question is that you think the number of cars on the road (not in garages) will drop or rise ?

I think it will rise:

- the people that afford expensive cars now(Tesla,BMW) will continue to own their own cars

- people that can't drive now but have enough money will buy their own cars

- some of the less rich people that own a car will keep using it

- people that use buses would use the new taxis if their are cost effective

So I think we will get more cars so we will have more energy consumed and we need better roads, better electric infrastructures that includes clean energy generation.

About safety my personal opinion is that with current software and hardware we are far for self driving cars in general conditions, we could get this cars maybe to work safely on specific roads.

I am also of the opinion that what Tesla is doing is wrong, aka here is auto pilot, it drives for you but you MUST still be 100% aware of the driving because you must intervene at any time, what advantages has this system ?

-advantage 1 you don't use your hands and legs that much

-no other safety that is not included in the driving asist packages from non self driving cars

- autopilot makes many drivers behave illegally and don't pay attention, the PR shows the cazrs driving themselves where the small print in manual tells you to pay attention 100% of the time

I think I understand your point, I am also frustrated by bad drivers that text, speed and kill people, I want them out of the roads too but it will not be easy to replace human inteligence, driving is a complex process and software and hardware is not there yet (I was an optimist like you about an year ago until the cases with Tesla hitting trucks, police cars and concrete obstacles happened and I read about the fact they have to ignore inputs because there are too many and the hardware is not good enough to process it) maybe Google cars have better sensors,hardware and software and can bring the optimism back for me.

I think it can drop or stay the same if the costs of driving by yourself and owning a car rise sufficiently.

You can do this by implementing congestion pricing and heavily advantaging carpool vehicles on the road by giving them more dedicated lanes, or by restricting single-occupant vehicles to a single lane on highways.

In urban areas, renting a parking space for your own car is already quite expensive. It's an easy decision to give your car up if robo taxis are cheap and plentiful.

In suburbia the calculus is a bit trickier because there's still a bit of keeping up with the Joneses. But there are many who'd jump at the chance of getting rid of their garage and driveway and have more living or personal work space.

I agree with you about Tesla autopilot.

Yeah, though adding more taxes will be hard, people don't like more taxes and the car industry will fight hard to keep selling cars to individuals.

With some good regulation and investing in infrastructures we could reduce the car accidents and pollution today without having to wait decades until most cars will be self driving cars, but we don't have the regulation and investment unfortunately.

Probably only in areas where parking of personal vehicles is expensive and in situations where people don't care about customizing vehicles/storing things in them. Most of the cost of a vehicle is typically miles, not time, at least if you buy them and eventually sell/trade them in. I live near a major city(though not in it) and I can't imagine getting rid of a personally owned vehicle.

No one in this thread can imagine getting rid of their personal vehicle; each and every one thinks that everyone else is going to get rid of THEIR personal vehicle, freeing ME to drive fast anywhere with no traffic.

It's some kind of narcissistic tragedy of the intellectual commons.

That's sort of true. I'm not giving up my car until government goons with guns show up to forcibly take it.

Ok let me unburden you. I can't wait to get rid of my personal vehicle if a comparable SDC alternative exists.

I'd be very much in support of congestion taxes, carbon taxes, and expanded HOV lanes, however.

How about everyone that doesn't own a car and just uses ride sharing? I imagine the low hanging fruit for car owners to embrace self driving cars is rush hour traffic. Commuter busses are already a thing, now add in the convenience going straight from your house to work.

Once people stop commuting with cars, they may find out that owning a car is no longer with it

I don't really get this argument. People will still want to own their own car to use whenever and whereever they want. If one currently owns a car, I don't see that this would change just because one doesn't have to drive anymore.

Cutting out the middleman (i.e. the company that wants to profit off of the ownership of the self driving car) should always be cheaper given enough usage. It's the same currently with owning a car vs Uber, Lyft or taking cabs etc.

Humans destroy the environment, maybe we should get rid of them.

I don't see how greenhouse emissions is relevant to this conversation. Obviously, we'd prefer to all be driving around (human operated or otherwise) in vehicles powered 100% by clean energy, and certainly people are working on that, but that's independent of whether or not those vehicles are operated by humans.

>> Humans destroy the environment, maybe we should get rid of them.

I belive genocide is against the law in most civilised societies, but, be my guest.

<Not that it matters> Genocide isn't the right term here. Genocide is to wipe out a national group [0]. Rather we are looking for a word that means "to extinct a species" or alternatively "to extinct our own species". It looks like "extinguish" or "eradicate" is the closest word we have to the former [1]. I'm not aware of any word for the latter.

<It matters even less, but> You could argue that to eradicate humans you would have to commit genocide several times.

[0] https://www.etymonline.com/word/genocide

[1] https://english.stackexchange.com/questions/253505/is-there-...

>> I don't see how greenhouse emissions is relevant to this conversation.

More cars means more emissions, the industry wants to sell more cars by loading them with self-driving features, therefore the issue of emmissions is relevants.

I mean, that's kind of like saying "I don't see how coronary heart disease is relevant to this conversation that's about selling more hamburgers", or "I don't see how diabetes is relevant to this conversation that's abot selling more sugary drinks".

> the industry wants to sell more cars by loading them with self-driving features

Strangely as it may sound, the plan is exactly the opposite, to make personal cars a rarity.

> "I don't see how coronary heart disease is relevant to this conversation that's about ending world hunger"

Fixed that for you. People without access to cheap transport or food really don't have time to worry about emissions or coronaries.

Your plan includes banning personal car ownership? If it doesn't, then how do you plan to reduce it?

Aside for car fetishists, owning a car is an expensive and draining choice for most people. It stands to reason that, if the fantastic self driving promises materialize, people would chose not to own a personal car, just like they choose not to own a chicken coop or a cow - themselves prized possessions in times before refrigeration and supermarkets (and cars).

The promises might fail but that's the plan anyway, to make ride hailing so ubiquitous and frictionless that it's more convenient than ownership.

The GP did not say anything about greenhouse emissions:

> (with the infrastructure needed to accommodate them)

Laying pavement quite literally destroys the environment, and a _ton_ of it is needed to support personal automobile lifestyles (particularly in suburban sprawl); far more than the amount required to support freight, walkers, cyclists, and mass transit. Modern US car-centric streets are very inefficient in space.

I don't know how much of the environment you think needs to be protected from being "destroyed" by the laying of pavement, but a helpful comparison might be to look at the figures for England:

https://www.bbc.co.uk/news/uk-18623096

"the proportion of England's landscape which is built on is… 2.27%."

If you want to make them obsolete you need a rival technology that possesses all of the attributes that make people choose cars over mass transit, notably convenience, scheduling & destination flexibility, privacy, and cargo capacity.

Failing that, self-driving cars as a service preserve all of these while also being significantly better for the environment. You don't need parking lots, because utilization is close to 100% and they go back to a depot when not required. It's easier to electrify the fleet, because the service knows exactly how far the trip will be and can bring the car back to a charging station while it has no passenger. You can share rides when multiple passengers have destinations that are close without them needing to go to the exact same place. You can have pods that dock with each other in motion (benefitting from better aerodynamics) but preserve privacy and split off to different destinations when needed.

>> If you want to make them obsolete you need a rival technology that possesses all of the attributes that make people choose cars over mass transit, notably convenience, scheduling & destination flexibility, privacy, and cargo capacity.

Actually, since the goal should be the preservation of the environment that is supporting our continued existence, what is needed is mass transit, which is the least damaging form of (mechanised) transport we know of; not a technology to replace it with privately-owned and operated vehicles.

I mean, I understand that's never going to happen, but that's only because we are collectively incapable of making the right decision on this one.

Exactly. It'll never happen, so why waste time and energy on a fantasy when there's a solution in front of you that isn't perfect but has a good chance of working and making things better?

We could just ban all transportation with the exception of walking/running. Important people could get special passes from the government to use other transportation forms. If only 1% of people were allowed to move around with aided devices that would surely get us to the goal of preserving the environment much faster that just banning the car.

I think that driving pass would be so valuable that it could corrupt any and all social institutions. The people handing them out would be demigods, imune from all forms of authority in this world other than the Holly Eternal DMV itself.

A more workable system for a free market society would be to shell out an equal endowment, say 20 driving hours a month, to all citizens. They will then chose if to sell them to the rich people or spend them to impress that special date.

Before you know it, the poorer plebs will fight the richer ones to maintain the driving voucher system, for the sake of the revenue they earn out of it.

I think this guy was being sarcastic to prove a point

>If only 1% of people were allowed to move around with aided devices

A lot more than 1% of people would be unable to function in modern society if you took their car and told them to walk and run.

> what is needed is mass transit,

Why mass transit and not walking more or personal blimps shaped like giant Trump babies? To make them more than your opinion, you need to put some numbers on those statements.

And the numbers work much worse than you think: between the low energy efficiency and low occupancy at non-rush hours, the massive infrastructure and the longer routes due to non-optimal layout, mass transit is better than American gas guzzlers but nothing to write home about. Here's some numbers put up by people who share your opinion - and that do not even take all these effects into account: https://www.citylab.com/transportation/2012/11/can-we-please...

So for in-city transport, mass transit is a hard sell on environment alone. (Intercity transport like trains are great though, and there is a strong convergence there, self driving cars pair well with fast trains, giving the best of both worlds: flexibility, speed and low footprint)

Technologies like Waymo's make mass transit more practical, not less.

Imagine a city-managed fleet of buses and shuttles that have a significantly decreased cost-to-operate, that dynamically respond to demand, aren't bound by work rules, and that customize their routes as needed. That's revolutionary.

Self-driving vehicles make mass transit more usable and economical, if you expand the definition of mass transit.

Implement congestion taxes and dedicated lanes for high-occupancy vehicles (vehicle more than 50% occupied) and you have a fleet of highly adaptable, efficient, utilized, and available buses at a fraction of the current cost. You've eliminated some of the major problems of mass transit systems - last mile connectivity, financially marginal routes having to choose between poor service or losing money, lack of frequency at off-peak hours, relative lack of comfort, convenience, needing to know a timetable as a user.

> while also being significantly better for the environment

That's probably true on a per mile basis, but it's actually expected to be net worse for the environment because it will enable car access for a large percentage of the population who are currently homebound due to either being disabled or not being able to afford a car.

That said, because they don't need street parking, in the long run it may enable more bike lanes and bike usage.

>it will enable car access for a large percentage of the population who are currently homebound due to either being disabled or not being able to afford a car.

That sounds like a really horrible outcome. /s

Personally I think achieving utilization anywhere close to that level is a pipe dream.

I just don't see any world in which the only reasonably efficient form of self-driving vehicles is basically a fleet of (dynamically scheduled, self-routing, autonomous) bus rapid transit - and I'm assuming these would only beat out trains in a realistic dynamic system mainly because of the additional flexibility/ease of re-routing.

> You don't need parking lots, because utilization is close to 100% and they go back to a depot when not required.

It’s just dawned on me how many cars you could fit in a carpark for self driving cars. They can park properly, park each other in and park close together. I wouldn’t be surprised if you could get 4x more cars in the same place.

Nothing like 4x but there are robotic parking systems (and other types of stacking systems) in places like Manhattan where the additional expense and inconvenience make sense because of land costs.

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

Fwiw, the Jaguar i-Pace that Waymo will be using in addition to the minivan is all electric. Getting people into electric vehicles gives us the option to recharge them via power generated by renewables. That's a great way to reduce the rate of climate change, even if it's a small step. (The minivan helps if it really gets people to drive with more people per ride)

Autonomous cars could allow space currently devoted to idle car storage to be reused for denser development patterns, which would reduce the need for car-based transport in the first place.

While I hope the technology succeeds I don't think it is going to be as big of a change as people think. If you are upper middle class in a large city where labour is cheap, like in some Asian countries, driving is already close to free. And while that is somewhat nice from a quality of life perspective, it also means that rush hour traffic is horrible and sitting in a car isn't much more fun just because it is cheap. Basically a car is still a car.

The idea is that with autonomous driving nobody needs to own a car and traffic jams are severely reduced because robots don’t drive like assholes (or at least by design shouldn’t...)

Not having to own a car to travel by car will of course increase driving, or being driven really. A situation where there is only self-driving cars is probably decades away at he minimum. And that doesn't really solve inherent throughput or speed issues.

Not only "doesn't really solve", but "makes vastly worse".

Not necessarily

>... technology move people from their homes or work to existing public infrastructure hubs

could be a thing. Also maybe some sort of road pricing to stop them being clogged with empty cars going to get pizza.

Not having to own a car to travel by car will of course increase driving, or being driven really.

I'm not sure. When you own a car, you pay large fixed costs but your marginal cost per mile is low, whereas with rented cars it's the reverse.

Sure, people who don't own cars currently will probably ride more, but everyone who does have a car now will have an incentive to ride less.

Your visible marginal cost (i.e. mostly your fuel) may be low but the vast bulk of car costs are proportional to mileage--especially in areas where salt in various forms (winter roads, marine) isn't an issue.

It doesn't really matter if it's lower than the fixed costs, it matters whether it's lower than these hypothetical rented cars, and most likely, it is.

Even if self-driving cars are 50% of the fleet, these cars drive better and can collectively work around traffic jams, so even human-driven cars benefit from automation.

They may also lower the costs (time, effort, money) of driving. That will almost certainly lead to more driving and likely greater urban sprawl.

And mostly because they can drive when you're not in it. I'd be very curious to see city simulations where people just get off 200m from their job and wave their car goodbye as they can park somewhere else smoothly.

In the techies dream world, the car goes off and, gives someone else who starts work a little later than you a ride.

Yeah sure, in the fully shared/non-owned system that would be it. A fleet of non polluting transportation mules.

In America people will get in the car to drive 200m instead of walking.

Because that's always what happens when you make something better: people use less of it. What?

That's clearly not what I said.

Reducing driving accident deaths (>37,000/year in the US alone) will be a pretty amazing change.

37,000 lives per year wasted due to fatalities.

400,000 lives per year wasted mind numbingly staring at a road and keeping a car driving along it.

Currently glad I did not go into trauma surgery. Most of their work is car accidents.

Sure, I just don't think that is as much of a paradigm shift as gradual improvement over time. It is not self-driving so much as the deployment of self-driving that will make that change if that makes sense. If everyone drove XC90s on barrier separated roads with automatic traffic cameras that enforced things like speed and distance between cars, we would also be much safer. But it would still be largely the same thing as today.

Uber went from zero to essentially killing off taxis, despite the fact that "a car is a car".

Primarily because they don't pay as much for their fleet as a taxi service does.

What's coming next is going to be that, except you don't have to pay your drivers anything because you don't have drivers!

(yes, there will be huge technical overhead of running this system and maintaining the cars/hardware, but over time this will become mass produced, and become cheap).

Uber went from zero to killing off a pedestrian, despite having LIDAR tech.

Downvoted by Uber devs no doubt.

"Uber went from zero to essentially killing off taxis, despite the fact that "a car is a car".

Primarily because they don't pay as much for their fleet as a taxi service does."

Uber (and possibly other similar companies) is a huge charity for passengers funded with investment money. It's definitely premature to declare they've demonstrated or accomplished anything before they operate on a sustainable basis. I do think using a credit card an app is far superior to the way taxi businesses are traditionally run, regardless of price, but I also don't see why traditional taxis can't adopt the mechanism eventually. At some point, the bubble will burst and capital will be scarcer for a while, and things will realign.

However, I think you should consider how autonomous vehicles enable better utilization of existing cars on the road. If you own a car, what % of the time is it sitting in your driveway idle?

Rather consider how with autonomous driving, when paired with services like Uber/Lyft enable a car to be active 80-95% of the time - driving folks around. This means that your cost “of the trip” can be significantly less once you factor how his affects vehicular depreciation.

Peak demand at rush hour determines the number of vehicles necessary to achieve a given level of service. Self driving vehicles don't change the underlying mathematics of the job scheduling problem.

This would only be true if everyone commuted at exactly the same time or along the same route.

You basically just described 'traffic.'

The goal I think is a complete ban on human driving on public roads. With that in place, maybe we can make significant improvements in traffic. I agree that the last mile will always be a problem though.

While self-driving cars are cool, I’d imagine self-driving subways would be easier. In New York, the f train in Queens hits a standstill because of problems in the d line. The air train at JFK has no driver. Why can’t we do this for MTA?

While we are at it, transit in the city should be entirely tax funded. N

We also need self-driving buses. The buses in Queens sometimes don’t show up for over half an hour and three show up at the same time. Why?

Agreed. While self driving cars are 'cool', they are still massively impractical. We need a better, rethought transit system in general.

>While self-driving cars are cool, I’d imagine self-driving subways would be easier. In New York, the f train in Queens hits a standstill because of problems in the d line. The air train at JFK has no driver. Why can’t we do this for MTA?

Technical debt. The system is 100 years old and MTA can't contain its contractors. At the current pace work to install communications-based train control (CBTC) on every line (it's on the L) will cost $20 billion and conclude in 175 years.

Oh, and rules require having two drivers on the self-driving train anyway, who occasionally have to drive to keep proficient.

https://www.theatlantic.com/technology/archive/2015/11/why-d...

The 3 buses can go faster through their route if they move together as on average each one has to stop less times. It's possibly a case of a misguided metric (average route time or delay) driving behavior to undesirable local optima.

Even so, on my way to/from work I'd estimate at least 50% of cars are still parked. That's 50% of space that could be gained.

That is what I am considering when I say that it is already close to free. Those city, at least for those people, essentially already have self-driving cars because it is so relatively cheap to be driven.

Optimistically, self-driving vehicles may allow the west to finally discover the joys of share taxis, which would increase people/vehicle and (hopefully) decrease congestion.

"“Our experiments show that with as few as 5 percent of vehicles being automated and carefully controlled, we can eliminate stop-and-go waves caused by human driving behavior,” Daniel Work, assistant professor at the University of Illinois at Urbana-Champaign and a lead researcher in the study, said in a statement."

https://www.rdmag.com/news/2017/05/study-shows-self-driving-...

Per the article: > “Before we carried out these experiments, I did not know how straightforward it could be to positively affect the flow of traffic,” Jonathan Sprinkle, the Litton Industries John M. Leonis Distinguished Associate Professor in Electrical and Computer Engineering at the University of Arizona, said in a statement. “I assumed we would need sophisticated control techniques, but what we showed was that controllers which are staples of undergraduate control theory will do the trick.”

If it really is that simple shouldn't it be possible to put this into an app that if 1/20 drivers had installed would also solve the problem?

An app can't convey near the granularity of control to the driver compared to what a fully automated solution could do, and you run the risk of making the situation worse due to the driver splitting focus between paying attention to the road and following the app instructions.

To some extent we already have "the app" (google maps) and I wouldn't be surprised if it's running for close to 1/20 of drivers actively in traffic, and a whole lot more passively collecting and reporting location info that is trivial to map to major roads. To that argument Waymo somewhat belongs to Maps (or Maps to Waymo if you'd like).

> An app can't convey near the granularity of control to the driver compared to what a fully automated solution could do

Yes I agree, however it might not require that level of granularity, that's what I was wondering given that the researchers were surprised by it working as easily as it did in their experiment. Since, That's why I am wondering, maybe it's something a human could be taught or directed to do by an app.

Open questions to me: Is central coordination required which observes the current state of the system and doles out specific instructions depending on that? If it's simpler is possible to craft a simple set of heuristics we could be taught to follow that would make things better if enough people knew them and followed them? Could they perhaps be implemented with simple road signaling targeting know problem area where jams frequently occur?

I'm pretty sure central coordination is not required for at least some significant benefit.

If you drive at the average travel speed steadily instead of filling gaps in front of you as soon as they open, that smooths traffic out for the car behind you as well. Traffic waves of stops and starts can still form back a bit behind you, but it's unlikely to happen immediately behind you. If even a small fraction of drivers drove like this, there would be much less opportunity for spontaneous jams to form.

I saw an impressive video demonstrating this a few years ago, but just spent a few minutes failing to find it. It showed a guy driving at the average speed of the car in front, so that the gap opened and closed as traffic in front slowed down, sped up, stopped, etc. There are articles about doing this, and animations, but what made this a compelling video was where it was filmed: a long straight concave-up road full of cars that was coming down a hill, so you could look out the back and see an long line of cars moving smoothly and not jammed, while in front there was a long line of cars stopping and starting chaotically.

I think this simulation[1] explains why even a small change in traffic patterns can have huge consequences to the stop-and-go behavior.

(experiment holding one of the buses for just a couple of seconds...)

[1] http://setosa.io/bus/

Autonomous vehicles will make deliveries become almost free. This in turn will make food delivery explode, to the point of kitchen becoming an option in houses. In general, owning stuff won't be as important as before because you'll be able to simply borrow anything you need.

Do they really belong to the same league? I’m sure this will be downvoted into oblivion because everyone on here is working on these projects but I don’t think self driving cars should be uttered in the same breath as any of the aforementioned technologies. It just doesn’t matter all that much.

Maybe I’m being dense but can soemone spell out the glorious moment we are all waiting for here after spending all this money?

- Trucking is cheaper because we fire the drivers?

- There are cheap Ubers taking people from ramen to sushi in SF but clogging the streets?

- There safer roads because there are no bad drivers?

Ugh, this is not a transportation revolution folks. Where are the personal transportation devices? The hypersonic planes? The hyper loops?

Well optimized, high volume autonomous taxi services could be revolutionary: imagine if cars being parked went from being the norm to being an rarity. Instead of 90% of cars in cities being parked, there could be drastically higher usage rates. The average apartment complex wouldn’t need to be 50% parking lot; cities wouldn’t be rows of parallel parking. I can’t find the reference, but a surprisingly high amount of useable space in densely populated areas in the US is parking lots. It wouldn’t happen overnight, but if cars were optimizing for time spent moving people, a city could be much more space efficient.

Wouldn’t it use more gas to park further away?

They won't park at all. They will drop you off and then go onto the next person.

Autonomous cars could be electric more easily - they could charge themselves.

So if it prevents thousands of people from dying it somehow isn’t a big enough impact? Or tens of thousands from being injured? Or makes co2 emissions 10-30% lower? Or makes delivery of goods same amount more efficient? Are you overlooking all these gigantic benefits to social and economic life intentionally? I don’t get it...

This is a good question, and doesn't deserve a snarky response.

This is the long-term good-case scenario (using USA-focused numbers, but you can extrapolate to the rest of the world):

-Car accidents will be reduced to nearly zero, saving about 34,000 lives and 250,000 injuries each year. The NHTSA estimates that this costs $871 billion per year[0].

-Because car accidents will be much less frequent, cars can be built lighter, saving the energy that's currently spent on armor. And in the scenario where people hail rides rather than own cars, the car that transports you will be the right size for you - i.e. a one-person car for most trips - which saves a lot of energy too. The energy savings help with intra-city air pollution, climate change, and decreased costs.

-Self-driving cars free up the 50 billion hours that Americans spend driving each year. At the average wage, that's worth about one trillion dollars.

-A lot of valuable land is used for parking (14% of Los Angeles, for example[1]). Hailed robot cars wouldn't need to park, and self-owned cars can drop their owners at their expensive destinations in a central business district, and drive a few minutes away to park in a garage where they pack together much more closely than humans can park them. The freed-up space can be used for housing or other valuable uses.

-Hopefully, self-driving cars massively reduce congestion. Due to their much-faster-than-human reaction time, they can drive very close behind each other at high speeds. Two single-person cars can fit side by side in a lane. Street-side parking space can be turned into more lanes. Reduced crashes result in fewer traffic jams and lane closures. (Personally, I worry that increased car usage might exceed these effects.)

Most of these figures come from Brad Templeton's webpages about robot cars[2].

[0] http://thehill.com/policy/transportation/207567-nhtsa-car-cr...

[1] https://www.citylab.com/transportation/2015/12/parking-los-a...

[2] https://www.templetons.com/brad/robocars/robot-cars.html

With the possible exception of MAD geopolitics, I can't point to instances where nuclear power has significantly shapes my day to day experience in the US. Certainly much less than hydropower. Sure Voyager and other nuclear powered space craft have provided infotainment for many years, but terrestrial electricity is fungible for the end user and nuclear power plants have always correlated with more expensive watts when I've had utility service that owned them.

I feel like MAD is pretty significant by itself. Might not have been as peaceful otherwise.

A serious question here, and not trying to ruffle any SV'ers feathers, but why do you see autonomous vehicles to be as valuable as...say transistors, or satellites? In terms of long term usefulness for the 99% who care much more about turning up at work on time and/or transporting things efficiently from one place to another.

Autonomous vehicles could optimize routes and driving to get things to places faster, reduce stress while driving, and create safer roads. If the tech was widespread enough, there could be autonomous lanes that can drive at very fast speeds too.

My favorite explanation of why autonomous vehicles will be so game changing is this imagined future conversation:

[30 years in the future...]

My son: Wait...you mean to tell me that people used to operate multi-ton machines, with a minimum of training, in a range of different conditions, and relied solely on human reflexes and human attention span to avoid collisions?

Me: Yes

My son: And nobody was killed?!?

Me: Oh no! People died ALL THE TIME!

---

Look at any source you like, and it's as clear as day: the LEADING cause of death for individuals in their prime productive years (late teens through late 30s) is the automobile! If autonomous cars are even a fraction safer than human drivers, the impact it will have on demographics and population productivity will be MASSIVE.

"the LEADING cause of death for individuals in their prime productive years"

No matter how unlikely it is to die in your prime, something will be the leading cause. Since it is very unlikely to die in your prime in modern first world countries, why focus on that? Are we going to go on a crusade to eliminate bathtubs next? I may be biased, because I'm now too old to die young...

Not eliminate them, the focus is on improving them, and yeah, we should definitively improve bathtubs so people die less in them.

I’m glad this logic wasn’t applied to planes.

You are thinking of scheduled flights. But general aviation is 80 times more dangerous. So we have refrained from applying an extreme safety culture to all flying. Obviously because there is a trade-off between people being able to do it and safety.

Human drivers average about 1 fatality per 100 million vehicle miles traveled. [1] So we are still a long ways from being able to access safety in comparison to humans. [1] https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/...

Number of fatalities per million miles driven is only one of the metrics. Hours spent per lifetime driving is arguably a more important metric. I can't wait for everything (cities, highways etc.) to be redesigned for self-driving cars but looks like its not going happen in my life time :(

If you have to sell self driving cars as "more people will die but some people will have more free time" it's not going to be as easy as if they are unarguably safer for everyone.

I don't think that's true. That's exactly how we were sold on cars in the first place -- you're more likely to die when you drive than when you walk, but you can get places faster to save time. Same can be said about e.g. caffeine. Humans are notoriously bad at estimating and considering tail risks.

Care to expand on what you mean by risk vs. return of caffeine?

People drink coffee (consume caffeine) because they wish to be more alert/are sleepy. This is a consequence of not getting enough sleep. Lack of sleep has many documented negative effects, but people trade sleep (health) for productivity/etc. via caffeine intake.

Cars didn't replace walking, they replaced horses.

I think you are begging the question - are cars more dangerous than horses? Just from reading random history, it seems like a lot of famous people were killed by riding accidents. And even today when the vast majority of people don't ride horses, I bet you can name a famous person who had a terrible accident.

Horse riding was much less popular than you think. What was much more popular was horse-powered carriages, which are much more convenient than riding, and also less dangerous. Horse riding was pretty much a hobby, a hobby of the rich at that. Try reading some books from the era. For example, in Steinbeck’s “East of Eden” there is a quite funny moment when Hamilton sons were going for dances, when two older brothers already took family buggies, the younger didn’t choose to ride on a horse, but did something quite different, which made his mother very unhappy.

Do you have any specific reason for believing the transition from horses to internal combustion engines increased casualties in the long run? Or is it just a gut feeling?

I mean, it is already hard enough to redesign infrastructure for trains, trolleys, buses, etc...

> I can't wait for [cities] to be redesigned for self-driving cars

Personally, I hope it never happens. The US is still suffering from when its cities were redesigned for cars, vs. public transit and pedestrians.

Self-driving cars can and will play a very important role in transportation, but cities should be designed around people, not cars.

Edit: I'm by no means criticizing your hope, and I do agree with you that cities should be redesigned to "accommodate" self driving cars, such as "buffer garages" (where autonomous cars are parked in a FIFO way, not in the way cars are parked today), auto-charging stations, smart curbs for pick-up/drop-off, etc.

One of the biggest issues is that the road infrastructure and the city layout is already there. Sure cities may be better if they were walkable but most people live across the full metropolitan areas.

Taking the SF Bay area an example: Sure I can take the Caltrain or the Bart to get to nearly any city but it'll always be the long way around.

Think San Mateo to Union City. Crossing a bridge it's direct. The road infra structure is already there and the cities are built around it. Sure there's probably some cryptic bus route but the difference right now is (with a quick google) 35 mins vs 1.5 hr.

I think self driving cars have the potential to build on our car centric ecosystems. Look at how Uber/Lyft have adapted into Uber Pool and Lyft Line. They've done more to change people's driving habits than any changes to the public transportation system. I'm excited to see how self driving cars solve this by dropping the price point further and exploring larger pools.

>Think San Mateo to Union City. Crossing a bridge it's direct. The road infra structure is already there and the cities are built around it. Sure there's probably some cryptic bus route but the difference right now is (with a quick google) 35 mins vs 1.5 hr.

that is the use case best suited for flying tub https://www.youtube.com/watch?v=EQK9m_OBVgY

> I think self driving cars have the potential to build on our car centric ecosystems.

But the problem is that car-centric systems are not scalable and sustainable.

Sure, self-driving cars would do well in the Bay Area, but they would only make traffic worse, as now people are willing to spend a longer time in traffic, since they are no longer driving.

Self-driving cars actually remove one of the negative feedback mechanisms of traffic.

Isn’t a self driving car fleet essentially the same as self driving public transit? Self driving cars can pick the most efficient routes to get everyone where they need to go, in a way that would not be possible if a human driver is required.

> Self driving cars can pick the most efficient routes to get everyone where they need to go, in a way that would not be possible if a human driver is required.

What's so special about efficient routing with a driver-less car? There's no route that a human driver couldn't follow just as well if assisted by GPS software.

Self driving cars aren't any more space efficient than human-driven cars. 300 cars on the road needs a lot of road, regardless of who's driving.

Putting people on a train is going to be far more efficient.

Will the train come close to my home so I can unload my shopping bags?

If the city is dense enough, the supermarket will be less than a block or two away.

I guess you saw the before and after in netherlands where they replaced some streets with a trolley line and filled the remaining space with grass, trees and walkways ?

Seems like you could estimate fatality rate from crash rate? From your source, about 1 in 200 crashes are fatal.

Yeah, but humans crash differently than robots. The fatality rate for human crashes does not apply to robot crashes.

There's more to safety statistics than just the mortality rate. We should be able to look at frequency and severity of non-fatal accidents too. Those happen a lot more often.

For example, accidents that occur at 35 MPH or less are much less likely to result in a fatality or major injury, due to the amount of kinetic energy that a human body can safely dissipate. So if Google cars have even slightly better braking or speed control, you're going to see an improvement. Looking at the average speed at which accidents occur would be useful information.

>> There's more to safety statistics than just the mortality rate. We should be able to look at frequency and severity of non-fatal accidents too. Those happen a lot more often.

When the industry or the press discuss the safety of self-driving cars, they almost always do it in terms of fatality rates. Not general accident rates.

Huh I thought there was that huge fixation on just "disengagements" or whatever a year back.

"Disengagements" are a measure of how close we are to a self driving car. I.e. one that doesn't need a human sitting in the drivers seat to handle these disengagements, not a measure of safety.

Right. Only Waymo has disengagement numbers that are even vaguely decent.

The CA DMV data has enough info that you can distinguish between "driver had to take over to avoid crash" vs. "vehicle was unable to proceed due to problem ahead and needed help". The first number needs to get very low before self-driving can work. The second, Waymo proposes to handle by having a very limited remote control capability, so remote support can get a vehicle past a strange problem.

"they almost always do it in terms of fatality rates"

That's an element of right-wing policy created to fight safety regulation. The position is that since the severity of non-fatal workplace injuries is hard to measure, only fatal accidents should be recorded.[1][2] Then there's too little data to make policy.

[1] https://edworkforce.house.gov/uploadedfiles/testimony_michae... [2] https://www.oshalawblog.com/2017/04/articles/its-official-os...

I wonder if there will be a psychological factor too. If SDV really manage to get better, during the transition, will the remaining human drivers improve to avoid feeling beaten by a machine ?

I personally think, during a transition to an SDV majority, human drivers will get worse.

Mainly because good driving is often synonymous with defensive driving. And you don't need to drive defensively if there's a large number of cars on the road which won't unexpectedly change lanes, turn right from the left lane or wildly change speed for no reason. If anything, human drivers might start doing unsafe maneuvers more as they see SDVs around them react quicker and safer to their bad driving. Why bother getting in the correct lane when you can pick the less congested one and rely on SDVs to move out of your way when you barge past?

There's also the issue of people switching from SDV to manually operated vehicles. What if someone uses an autonomous car for commuting on weekdays and gets used to the car acting for them, then uses a normal car on the weekends? They might expect the car to emergency brake for them when a pedestrian steps out into the road, and in the split second it takes them to realise the car will not do that for them they've already traveled too far to stop in time.

Also, some people are more reckless than others. A responsible driver might have a much lower chance of fatality than the average

That sounds like common sense, but do you happen to know if it is actually true?

i’m sure insurance companies know whether it’s true or not

Waymo have driven 5 billion in simulation. I think simulation is going to be the major change that automous vehicles can bring.

Assuming simulation is proven to be good enough for training cars, then having the ability to simulate crashes so that cars can respond in an optimal way is going to make a massive difference.

Waymo uses simulation to fuzz test their algorithms. When real driving records a troublesome situation, they load that data into a simulator and run it over and over with variations to see that the results are safe. Remember the problem with machine learning - most of the time it's right, and sometimes it's really, really wrong for no visible reason.

People almost always tell me autonomous vehicles are much safer than Human drivers, but that has not yet been proven [1]. Luckily Tesla, Waymo, etc. are starting to put major amounts of mileage into the data.

https://orfe.princeton.edu/~alaink/SmartDrivingCars/Papers/R...

well, that’s not really the right question; given enough time, they’re always going to be safer than human drivers

the question is at what point we decide that it’s “good enough” to be a life saving compromise

humans are really terrible drivers compared to sensors. the number of inputs, precision, and quick decision making are all much better for driving than humans

Waymo has tested in 25 cities, they aren't currently in 25 cities. Of the 600 Pacifica's they have, only ~160 are deployed. So to do 25,000 miles per day, that's an average of 156 miles per vehicle.

That does seem like a lot, I suppose they theoretically could be driving 24/7/365 (the wear and tear would be serious, you'd think).

The idea of moving people to public transit sounds good, but I wonder if there's enough spare capacity in most transit systems to serve lots of new riders...

There is spare capacity in most transit systems to accommodate new riders, just not at the SLA that prospective new riders are accustomed to.

If your suburban bus stop is a 10 minute walk from your home, and the bus comes by once every half-hour, and then you have to transfer to an every 20-minute bus, that is sufficiently offset to make the transfer take time... You aren't going to have a good time on your commute.

That seems pretty reasonable - 8 hours/day at 20mph, 5 hours/day at 30mph, and 2.5 hours/day at 60mph. Given a mix of highway, rural, and city driving, 156 miles/vehicle sounds about right.

And only 6.5 MPH if they're running 24 hours/day

To me, the very existence of a self driving vehicle is an amazing technical achievement against a really hard challenge. I suppose that explains why the math makes sense. 25,000 aggregate miles/day for a fleet of 600 vehicles is below 42 miles/day/vehicle. Each vehicle is providing a level of service below 2 miles per hour. Again, as a geek I think it's amazing but not convinced it is going to scale to the point that will justify those governors pushing public policy changes that impact the general public...at least for now.

The number of miles is not a sufficient metric; the question is, how much of the automotive phase space (weather, traffic conditions, terrain, road surface type, time of day, etc) has been explored? Running over the same mile a million times does not add anything to the self-driving model.

Interestingly, going with that logic disengagements per mile will at some point correlate with the most advanced AV companies, not the least.

I guess the response would be to target some SLO budget and, once disengagements get too high, refocus on existing cases to get it down to an acceptable level.

I'm still waiting for readings on n-SDV interactions too. I can sense the emergent resonance failure coming.

Waymo is supposedly planning on launching a ride-hailing service. Perhaps they can start giving rides to customers while continuing to test with a live driver (to take over if necessary). The revenue from the ride-hailing service could offset the cost of testing, and allow them to operate indefinitely until full autonomy is ready and legal.

Unfortunately the liability is too high for them to start taking passengers without thorough.

The idea is pretty clever. People paying to be guinea pigs.

Sure if people signed off, it would reduce some of the liability but that's still the golden question: who's responsible if something happens?

IANAL but seems pretty obvious. If I'm just sitting in some company's product and it causes harm to me or someone else, it's 100% on the company. Of course, anyone can be sued for anything but the idea that I could be criminally negligent because an autonomous vehicle I'm sitting in (whether car or train) killed someone seems absurd.

Waymo has insurance from Munich Re. Do not think liability problem.

https://www.insurancejournal.com/news/national/2017/12/20/47... Startup Trov, Munich Re to Insure Riders of Waymo's Self-Driving ...

You can insure just about anything. At some point Waymo will be able to show the risk is low enough that insurance will be economical, if it’s not already.

ride sharing isn't exactly very profitable at the moment. I don't think the revenue would offset the additional engineering/product costs of just running the rideshare service, much less the cost of testing.

>Waymo Early Riders Can Hail Actual Driverless Minivans Now http://fortune.com/2018/03/13/waymo-driverless-minivans-phoe...

5B miles in simulation. I wonder do they source the situations for simulation from Youtube from crazy/fail/accident driving videos.

I can't imagine these poor video streams are useful inputs for the algorithm...

Maybe simulations are when the human is controlling the vehicle and the algorithm is only allowed to .. simulate what it's actions would be if it were in control?

If they've been recording everything, they can run the previous inputs on new algorithms and check the results, and also simulate sensor failures and what not.

Running the system on recorded inputs to see how it compares with human output is good too.

When they encounter interesting situations while testing on real roads, they then take that situation over to the simulator and train their algorithms on 1000s of subtle variations of the situation, which Waymo calls 'fuzzing'.

Simulating effectively is non-trivial. In spite of testing on a comparable scale to Waymo, Uber made very little progress in 2 years because they couldn't get their simulator to work right.

> I can't imagine these poor video streams are useful inputs for the algorithm...

Store a rolling last X minutes of full sensor data into memory, then dump it to disk if the supervising driver hits the "that was a fuck-up" button.

They don't use the video data in the simulations, they use the processed "Where everything is" stream. That makes it more robust to sensor tech upgrades, but less to errors to do with bad sensing or sensor failure.

Computationally cheaper and simpler, no doubt, too.

The source the simulations from sections of road that their cars had trouble with.

https://www.theatlantic.com/technology/archive/2017/08/insid...

That would be entertaining to watch. Possibly also useful to check the cars don't freak in weird situations. They kind of had a real one in the Chandler crash https://www.youtube.com/watch?v=vxqBS2-4puw

Are Waymo using AI and humans to build a ginormous curated decision tree that can be audited or are they doing something more black-box / pure AI / something else?

I always imagined a big decision tree that could get diff’d after each AI training run and tuned by an army of humans.

Why do you assume there's any "AI" in it at all?

Because this would fairly obviously be impossible without ML techniques.

There's some info on the systems here https://techcircle.vccircle.com/2018/05/09/how-google-s-waym...

>“While perception is the most mature area for deep learning, we also use deep nets for everything from prediction to planning to mapping and simulation.” ...Dmitri Dolgov, CTO and VP, engineering, Waymo

etc

> ...the company is also working to apply its self-driving system to three other areas, including logistics (so trucking)...

This whole 'disenfranchised underclass' thing that America is struggling with will get considerably worse when automation turns all the truck drivers and their families out onto the street.

When automation arrive to that level, that could happen, but it's far from unavoidable.

If automation happens is because it makes the economy more productive. If before happening the economy could sustain all the people, logic tell us that now could do it even more easily.

At the end, it's a political decision.

Most people used to be farmers, too, but prohibiting industrialization was not the right call.

Driving trucks is a dangerous occupation that's terrible for your health, and often takes you away from your family for long periods of time. In a few generations when it's long gone as an occupation, no one will miss it. Getting from now to then is unfortunately going to be a rough transition for some people, but stopping that transition entirely isn't the right call.

> Most people used to be farmers, too, but prohibiting industrialization was not the right call.

The industrial revolution caused enormous suffering for ex-farmers. Many of them lost a life of autonomy to be forced to move to cities and work in dangerous conditions for starvation wages. It caused significant social instability.

https://en.wikipedia.org/wiki/Industrial_Revolution#Standard...

And I think that is the place it will happen first. At least outside America where most roads are far too complicated for driverless cars.

Motorways, on the other hand are relatively simple. And we have service stations that driverless lorries could drive themselves to, to be picked up by human drivers for the final leg. And there's an obvious commercial need. And there's an obvious safety need (who hasn't seen a swerving lorry?)

1042 miles per hour is a fast car

>> Waymo has “driven” more than 5 billion miles in its simulation, according to the company.

Impressive. That's roughly equivalent to about 0 miles driven in real-world conditions.

I'm assuming sarcasm. What other way would you have them test besides current ways that they are doing now?

Reply on news.ycombinator.com