- Cruise permit is for robo-taxi service, available to public (fully driverless, nobody in the car)

- Waymo permit is for robo-taxi service, available to public (human safety driver behind the wheel at all times)

- Nuro permit is for robo-delivery, available to public (no human passengers)

> Cruise has had state authority to test autonomous vehicles on public roads with a safety driver since 2015 and authority to test autonomous vehicles without a driver since October 2020.

> Waymo has had state authority to test autonomous vehicles on public roads with a safety driver since 2014 and received a driverless testing permit in October 2018.

More info at the DMV's website (although it's a lot to digest): https://www.dmv.ca.gov/portal/vehicle-industry-services/auto...

https://youtube.com/watch?v=zdKCQKBvH-A&feature=share

On a scale of 1 to 10, how confident are you that your service won't kill someone?

>Fatalities from Crashes

>In 2016, over 33,000 traffic crashes resulting in fatalities, major injuries or minor injuries were reported on Bay Area roadways.

10% of the crashes resulted in more than minor injuries, so 3.3K serious crashes.

Drivers in the USA are - sorry - absolutely terrible, the driving license requirements are ridiculous.

You may know this already but here in the US, even though driving is technically a “privilege”, in reality it is a requirement in most locales due to lack of consistent and reliable alternatives. That’s part of why getting a drivers license is fairly easy.

It’s also why your “I’ll have my son shred my license after I turn 70” wouldn’t work unless you were living in a place like NYC or Chicago, or were totally okay with having little social life (dependent on your kids’ and friends’ ability and willingness to drive you to places).

F1 cars can break 7g’s cornering where street legal cars have trouble hitting 1/3 of that. They do 0-120 mph faster than most cars can hit 0-60.

I'm all for self driving cars and I believe they are the future. I also believe they will be a decade or more in the coming and that we need at least one more level-up before AI sofware/hardware combos are solid enough that your typical self-driving solution will outperform an experienced driver.

(Or even better, what he predicts will be the distance driven per serious at-fault crash.)

Maybe he does care, but how much?

This is not to be flippant (i appreciated and used Justin.TV a ton back in the day..canceled Cable TV cause everything was there for one to watch ... 24/7 marathons of your favorite shows) but what I view of the harsh reality of putting these things on the road as Uber's self driving car already killed a pedestrian. Personally, I'm not sure I could in good conscious work for a robot car company (just was being recruited by such a company).

2. Where are Cruise's testimonials etc. for the presumably successful San Francisco non-commercial testing of this "fully driverless, nobody in the car" system you now have commercial deployment for?

https://en.wikipedia.org/wiki/Yes!_We_Have_No_Bananas

there's a way to ask a serious question to get a serious answer, and then there's your way.

There's a way of asking a question without being a total dick, and then there's your way.

Being direct doesn't mean being an asshat. I have plenty of experience of doing it the wrong way, and yes, I'm suggesting yours is not the right way.

Edit: "I'm sure it's in the budget."

Actually, I'm guessing this would be covered in the liability insurance. So it's probably a much smaller line item in the budget than you are so dramatically inferring it to be.

However companies do provision for payouts in potential cases whether private/ class action or government regulatory ones , they are legally required to do so and disclose it in their filings.

The actuaries at Insurance Co do that kind of analysis for all products for life/ medical to workplace/ product safety this would probably fall under.

The courts also have fairly well known tables/methods for calculating cost/value of life used to compute compensation for injury/death.

It seems like with self driving, it's like 45 companies just started making clones of the THERAC25 where they "are pretty sure it's fixed now, this is easy". It's the same damn ego.

I believe you, but skeptics won't be persuaded by "the founder said so on HN".

> - Cruise permit is for robo-taxi service, available to public (fully driverless, nobody in the car)

We need to have fewer executives talking about their awesome safe technology and more executives being forced to use their awesome safe technology.

It is extremely unfortunate that as the condition of the permit the DMV did not force dogfooding requiring that every executive of a company applying for driverless robo-taxi permit and executives' families, including their children, were mandated to give up their drivers licenses and all other modes of transportation ensuring that those executives and those executives families are the constant test subjects of the technology.

Same goes for the likes of Musk.

I am pretty sure they recently announced "Full self driving v 2.0 for real this time, seriously, guys" and did not drop out

"In this paper, “disengagement” refers to any event in which the AV’s automated driving mode is disengaged. This is broader than the definition used in certain California state regulations, where the term more narrowly refers to certain safety-relevant disengagements."

Most Waymo reports are "stopped in traffic, was rear-ended". There are two reports this year where the Waymo vehicle was responsible. In both cases, it was backing up while doing a three-point turn in a narrow street and bumped into something at 2MPH. So they need better sensor coverage in back.

[1] https://www.dmv.ca.gov/portal/vehicle-industry-services/auto...

https://en.wikipedia.org/wiki/Gartner_hype_cycle#/media/File...

They're generally against "create a self-driving startup in 2021, IPO by 2024 at the latest". That's just snake oil salesmanship.

At night only testing at super low speed. It's a good way to get started, but definitely not an autonomous taxi that title implies.

Source: actual ADV engineer

(aside from folks crossing the street in non-reflective dark clothing)

I think the worst time to drive - as a human - is right when the sun is coming up or going down, or under tree cover when you go into and out of shadow.

Fatal accidents do indeed spike around these times, although they seem to be worse in the evenings than mornings, and more pronounced in southern states than northern ones.

The overwhelming majority of people who fall from 30ft do not attempt to decelerate in the last second or so.

People who text their way into stopped traffic notwithstanding, 30mph is the kind of collision speed you see on roads where traffic moves low highway speeds. Edge cases notwithstanding (i.e. someone barreling through a red or whatever) collisions on roads where traffic moves 30-40 typically happen at speeds 10-20mph below that because people generally brake but too late to prevent a crash.

- How often is an accident while walking along the roadway getting sideswiped, instead of just a full-on crash

- The equivalent of being run over by a car doesn't happen when falling 30 feet. I wonder how much this accounts for getting hurt as a pedestrian.

This is just what I thought of, it is also a threat to cars, but that would be offset more by the velocity of the car that got hit than it would be by the velocity of a pedestrian. (For example, a car that is going 25mph and gets rear-ended by a car going 30mph)

Also there is often hitting the brakes before collision. So it will be slower sometimes.

If a driverless car just stops, I don't feel comfortable walking in front of it if there's no indication that it sees me.

How could that be solved?

For car-car collisions, not really. Yes, getting into a head-on 30-30 MPH collision is pretty bad, but you're more than likely going to walk away from it, especially if you're in the back seat.

https://one.nhtsa.gov/people/injury/research/pub/hs809012.ht...

Features to prevent/alert to collisions, sure, but I'd assume this is about hitting someone when going X speed; surely that's pretty similar now and then.

Not taking a position on the study. Just felt like sharing.

Recent trends in vehicle purchases also increase pedestrian harm, as the popularity of sedans has waned in favor of SUVs and trucks. These days 72% of vehicle sales are for SUVs and trucks[2], and the trend is expected to continue into the future.

[1] https://www.codot.gov/safety/traffic-safety-pulse/2019/march...

[2] https://www.nytimes.com/2020/05/21/business/suv-sales-best-s...

Hitting one's head on the windshield is/was the source of a lot of the pedestrian injury and fatalities and modern safety requirements try and prevent this outcome.

Would you rather get hit by a'21 CRV or a '95 LeBaron? CRV seems like the obvious choice.

Of course there's a lot more big trucks to hit you on the roads in 2021 but improvements in pretty much every smaller class of vehicle do a lot to balance that out.

Looking at the data I don't see any clear trend that pedestrians are at more risk today than in the past.

https://injuryfacts.nsc.org/motor-vehicle/road-users/pedestr...

One key trick other than the pop-up headlamps going away, was a gap between the bodywork and harder internal surfaces. As I understand it that goes something like this:

Think about a large steel panel such as a car bonnet (hood?), obviously it's no comfort blanket, smacking into that isn't a good idea, but it will bend and absorb lots of energy during impact. Now, think about an engine block, that's not going to bend at all. In a desire to give a more stream-lined look, older cars would mount that large panel almost touching the engine block and other large stiff elements, because why not. Well, dead pedestrians is why not. If you add a gap that gap absorbs lots of energy that otherwise is going to cause injuries to a pedestrian. I can't prove it, but I credit that for the difference between walking away with a damaged jacket and spending the rest of my working day in A&E being told I'm not dying so can't cut to the front of the queue, but I'm also not OK and so mustn't leave yet.

You're exactly correct that all that empty plastic in the top front of modern cars and crossovers is there to protect pedestrians (hood designs have been revised to facilitate this as well).

On the other hand the height and shape of the average car has changed to make them less pedestrian friendly.... but any one of these vehicles also sold in Europe includes a bunch of measures to lessen damage to pedestrians.

So I feel very comfortable saying a study from the 90s is not relevant because too much has changed on both sides of the equation.

Speed being "slow" or "fast" here is unrelated.

https://www.littlerock.gov/media/2484/the-relation-between-s...

There's got to be a way to make public transportation stop being the least attractive option in low to medium density urban settings.

Frequency and reliability are what actually drive ridership

That's why it's so shocking when these things happen and are accepted by people who actually resist reform because they view enforcing standards of cleanliness and safety as being "anti-poor" and suggest that even complaining about these issues is being classist instead of just advocating for basic human decency and higher quality public transport.

BART has police, they're shockingly bad, even for police.

I will never come to understand why the cloistered folks of Palo Alto hate poor people so much and will always resent them for it

Seriously, have you ever taken a BART train late at night where there was a guy tweaked out of his mind? Or sat in chairs that look like they were shit on 5 times and never properly cleaned? Why do you expect people to deal with that as if it's perfectly okay? Why do you expect people not to take alternative modes of transportation when that's what they have to deal with? This kind of anti-societal behavior is only accepted in the US, and nowhere else in the world.

[1]: https://www.sfgate.com/crime/article/BART-takeover-robbery-5...

My experience with the BART may have been unique, but from what I've read it probably is not.

and my point is that the ne'erdowells or whatever is correlative of poverty. the conditions present in the greater bay area are represented on bart and I guess if you're a cloistered tech bro and you've never been to oakland that's scary or something

Honestly, I don't understand people who try to normalize violence and argue on public forums that it's acceptable that public transit be an open sewer just because it gives them a faux edginess in liking "gritty" public spaces or sense of twisted moral superiority to those who do care about these things.

For example, public transportation that has locked “cabins” for people who want to feel additional security. For cleanliness, constant sanitization of seats and floors. Trash bins available on the train, with workers replacing them as needed during station time.

No one is kicking a poor person off MUNI but if a homeless person pees on the seat, it gets cleaned up immediately. People that feel unsafe can go into the cabin. Unarmed security guards can be placed at every station, not to hassle riders, but to give assistance to people that ask.

This is not some foreign concept — some Asian countries do this quite well.

A few years ago, I was helping my friend consider options for saving money. On the table was if they needed to keep their car. Most of their commute was via public transit. However we looked at the number of transits for which public transit was not viable per month (visiting friends on the other side of the metro area, running to stores), multiplied by the cost of using a ride share service. Comparing this number to the car payment, auto insurance, gas, and maintenance, it was no contest: even a few trips amortized the cost of the car and made it worth it.

Every trip has a cost in dollars, time, and externalities (causing congestion, environmental impact, etc). People will tend towards being efficient with those costs, judged by their relative importance to them personally.

The thing is, owning a car is mostly a fixed dollar cost per month. Once you pass the point where owning one is worth it for a few trips, the cost of additional trips is far less, while the time savings remain high. The only time I consider public transportation is when parking is a problem, as all other things considered driving is preferable.

Once self driving cars reach a point where they are closer to the amortized cost of a trip when owning a car, the tradeoffs start to change. Those who use their cars for few, but necessary, trips will be the first to ditch owning. I suspect many multi-car families will downsize to a single car. Once you get rid of your car, and are using self-driving cars for the trips where they are necessary, the question for every trip becomes "public transit, or self driving?" Different trips will have different ways they play out. However lot of people who are currently eating the cost of owning a car and might as well drive anyways will suddenly be able to save a few bucks by using public transit instead.

The final magic of all this is that if you're going anywhere more than a few minutes away, there are many other people going from approximately where you are, to approximately where you're going. The big killer in public transit speed is time. Time going to a common pickup point, time waiting for it to arrive, time spent not going on a direct route, time spent stopping for others to get on or off. With a sufficiently well used self driving car network, it could easily join a few rides together at a point nearby, and have self driving cars ready and waiting at the point near your destination. Two quick transfers, and that's it. If the self driving car app says "45 minutes, $20, or 50 minutes for $10", lots of people would save that $10. For destinations such as downtown, airport, sport stadiums, there's even only one transfer. Given car pool lanes and such, transferring to an express bus might actually be faster than driving directly.

However, there are a bunch of potential one-off situations that make owning the car (or at least having it on hand) important, too. For example, what can you do when you need an emergency vet visit? You're probably not going to be calling a rideshare or taking a bus for that. Those what-ifs are also a big reason I still have a car, even though I don't technically need one.

Once self driving cars remove the inconvenience of the last mile and difficulty of coordination, mixed modes of transport start to become really convenient. In your scenario, you might take a self driving cab from a suburb house to a small transfer station, transfer to an express bug/shuttle, which then takes you to a car rental. This hypothetical car rental wouldn't make much sense today, but consider a rental in a small mountain town, renting out off-road capable vehicles. From there you could go on your hike. Though this solution would be a little slower and less convenient than driving directly - on the larger scale for many people dealing with this 2 times a year would be better than the costs of owning a vehicle. Plus for many people it doesn't make sense to own a large rugged vehicle for driving in cramped downtown spaces - using the right vehicle for the job becomes an awesome option.

More popular trailheads for which the drive is fully paved won't be the first area serviced by self driving cars, but would be eventually.

The meta-point is this: Many people survive today just fine without owning a car at all, so it's reasonable to assume that self driving cars will expand the circle of people who's use cases will be fully covered by self driving cars. This will definitely not be everyone as soon as self driving cars come out.

As this circle grows, various niches of transportation will reach a critical mass and be developed. Another example is transportation between major cities. Currently you have inconvenient greyhound busses, or the rare train in the US. Then once you get to your destination you have to deal with transportation within that city. Today it's just easier to drive a few hours, assuming you already own a car.

Pet to the vet?

Musical instrument?

Shopping / incidental purchases?

People will not care, and I also do not think these automated cars are going to kill any pedestrians anyways.

I don't believe that.

One pregnant woman gets hi, and she and her unborn child die, and there will be so much anger and pressure on the politcians, everything will get outlawed.

Human drivers are individuals... John Doe killed someone, blame John Doe. Autonomous car kills someone, blame all autonomous cars!

Maybe not all autonomous cars will be banned when someone is killed, but I wouldn't be shocked if the same thing happened to whatever company that built the car in question that happened to Uber after they killed an innocent bystander.

Well that's exactly my point. The manufacturer and model of the self-driving car will be blamed and possibly banned, but not all self-driving cars will be. The first pedestrian death isn't going to be blamed on all autonomous cars.

Regular people are very wide bands, and there's a reason horror sci-fi movies where robots kill humankind are extremely popular.

I think we'll see a situation more like nicotine vapes. Instead of trying to crush the industry, big cigarette manufacturers just bought them.

I know Uber did, and now they don't have a self-driving car program anymore. not necessarily causation, but it can't have helped.

I do think it was causal: when the official report came out it was clear that their system was completely inappropriate for autonomous driving. https://www.jefftk.com/p/uber-self-driving-crash

There's been no confirmation whether Autopilot was enabled.

[1] https://apnews.com/article/technology-business-6127ae797c528...

    This entire incident occurred without any actual input or action taken by the driver of the Tesla vehicle, except that the driver had his hands on the steering wheel as measured by Tesla’s Autosteer system. Indeed, the Tesla Model X was equipped with an Event Data Recorder (EDR) which is intended to enable Tesla to collect data and record information from its vehicles and also provides information on various processes of the vehicle’s functioning systems when a crash occurs. The information regarding vehicle speed as extracted from the Tesla Model X provides proof of the foregoing facts[2]

Wow, driverless actually means no driver here. But how does that work? What if it makes a mistake? Is there some fallback pilot somewhere that can take over and control it remotely?

Waymo is allowed to test at all hours, in rain and light fog, within a limited geographic area.

[0] https://news.ycombinator.com/item?id=28710098

The permitted taxi service allows them to scale up and charge a profitable fare. In the testing phase, their fares are severely reduce.

Cruise can charge full price without a driver, but Waymo only has the permit for a backup driver at full proce.

"Hi, yeah, car's gone berserk, we're approaching a river, all doors are locked, can't get out, need assistance."

"Hi, thank you for calling. Please hold (your breath)."

What makes you believe the autonomous car is more likely than a human driver to do that?

2. life experience, software expertise, common sense.

https://medium.com/swlh/what-do-self-driving-cars-and-elevat...

At this point I’m pretty sure it’s not true. If past public sentiment was so against automatic elevators, there would be at least one newspaper clipping or digitized article.

1. https://news.ycombinator.com/item?id=25585122

Certainly not any recent (consecutive) 10-year period.

The US average is about 36,000 auto-caused fatalities per year: https://www.iihs.org/topics/fatality-statistics/detail/pedes...

>Waymo is authorized to use a fleet of light-duty autonomous vehicles for commercial services within parts of San Francisco and San Mateo counties. The vehicles are approved to operate on public roads with a speed limit of no more than 65 mph and can also operate in rain and light fog. Waymo has had state authority to test autonomous vehicles on public roads with a safety driver since 2014 and received a driverless testing permit in October 2018.

Why is this progress? Evaluating progress in self driving cars is extremely complex. Why do people trust that the California government is able to do so?

A milestone is a working product with many user videos on YouTube. This is a press release from the California DMV.

Does the same hold for the FAA? The FDA?

"you would not understand that DMV issuing these permits is actually a strong sign of progress" explain to me why you believe this. Videos on YouTube convince me something is real and exists. A government press release convinces me that some process was followed, but perhaps one having nothing to do with the solution to the problem.

Self driving is almost entirely a software problem. Testing that software is equally as complex. Do you believe the CA DMV has an army of good software engineers? Data scientists? If not, how did they evaluate it?

The fact that they're permitting it means it will be happening in practice, which will generate lots of data points, which (a) will help the companies improve their software and (b) assuming it's not a disaster, will get people used to seeing self-driving cars that really work, which (from a practical-politics perspective) will make it much easier to get permission from other regional governments, and to expand the permits.

Even then, I'm not sure how you can say its always just a software problem when no one has accomplished it yet.

Self driving may be a software problem, but that does not necessarily mean that we have the means to make that software a reality today. Given that fact that we know that intelligent beings make plenty of errors and that they are much smarter than your average self driving system, the enormous range of conditions and inputs and the relatively limited amount of computing power that can be thrown at this problem I highly doubt that it is a 'mere matter of software engineering' at this point, taking into account the state of the art.

Agreed that testing it is super complex, in fact, I think that testing is the key: without a fully representative test equivalent to a driving test none of this software should be on the street. That's what we agreed upon: pass the driving test and you are allowed to drive. American driving tests are ridiculously simple, so it should be easy to pass such a test and yet that hasn't even been made a mandatory requirement for certifying a self driving solution. Until then I don't think any of this stuff should be on public roads.

I don't think the CA DMV needs an army of sofware engineers or data scientists, they are first and foremost a regulatory body that can hire external experts to give them guidance, and I suspect that they genuinely believe that self driving systems from these manufacturers are now solid enough to be given wider latitude during this experimental phase. I personally disagree with the decision, but I do think it is progress, that a regulatory body would dare to attach their name to such a decision rather than to play it safe, this is a clear shift in perception. Time will tell whether or not it was the right decision, if it turns out it wasn't there will be a bigger set-back than this was a step forward.

The shortcomings of Comma.AI and Tesla FSD are very clear due to YouTube. If Cruise has a much safer system, that is great, I would just like to see more data.

DMV consists of bureaucrats that seemingly just rubber stamp the permits hoping for the best.

In that context, people actually using the technology, testing and evaluating its performance are like FAA/FDA and not DMV.

https://www.nhtsa.gov/technology-innovation/automated-vehicl...

You can't expect the people working there to not be driven to insanity.

Technical progress is the current limiting factor so I agree that it is more important to emphasize right now.

On the other hand, I disagree that user videos on YouTube is a sign of progress. 1. There is an obvious selection bias as only a minority of people with access to the software actually upload videos. 2. The videos are qualitative in nature and it is not feasible for a single human to go through all of them and form an objective opinion on the state of the art. So it’s prone to confirmation bias too.

I’d argue that there isn’t a common yardstick of technical progress for game theoretic reasons but I’d love to see one.

So Waymo then?

https://www.youtube.com/results?search_query=waymo

While FSD beta may still have issues, I see no reason to believe that this doesn't also have similar issues.

Obviously this is cherry picked, but you don’t think a car equipped with LIDAR, imaging radar, more cameras, more compute, and temporally dense HD maps that has focused on the same area for its entire existence would perform better than a Tesla?

As you know, Tesla's technology 1) lacks lidar and 2) is not targeted towards driverless operation (yet).

The AV industry isn't in a place where it's building light AGI that elegantly handles the driving task through a sophisticated understanding of the world. The systems out there are still fit to their fairly near-term objectives; it's ridiculous to assume that Tesla's camera-only, hands-on-the-wheel system tells you which issues (if any) are faced by systems with different objectives built on dramatically different hardware.

Sadly, that software is still a ways away. Don't fall for the trap believing that self driving is at all a regulatory problem. It is only a software one.

1. As a developer, the hardware and software to accomplish it.

2. As a rider, the improved UX (i.e. safe, comfortable, and fast).

3. As a current-driver, taking over the liability.

Based on the Comma 3 launch event, it seems clear that you and Comma.ai are hyper-focused on 1 and 2, but have no plan for 3 anytime soon. I think this is the big disconnect: Cruise and Waymo are focused on 1, 3, and 2 (in that order). DMV approval is entirely in the bucket of 'transferring liability', and this announcement is progress on that front.

The chicken and egg problem of 'publicly facing data driven technical viability to then get transfer of liability' and 'transfer of liability to then get data at scale' is a fundamental problem of self-driving. Comma.ai and Tesla are tackling this problem with 100% 'publicly facing data driven technical viability' and 0% 'transfer of liability'. Then one day Comma.ai and Tesla may also try to tackle the 'transfer of liability'. However, why limit the competition? Cruise and Waymo want to start with 'transfer of liability'. Its useful work that needs to eventually be done. Might as well have some companies start early.

Hopefully even if Cruise and/or Waymo mess up, the DMV will use those learnings to improve their validation strategies for future players looking to transfer liability (i.e. Comma.ai and Tesla when they choose to be ready).

Public transit is also based on an order of magnitude larger transport. You're talking about buses that are made to move 50-100 people or trains made to move hundreds... not a car that will be moving 1-3. The reason many routes get cut down isn't because there's no demand at all - it's because the demand is below the threshold to run the service profitably. If you can cut down on the cost to run the service by cutting the driver out, making the vehicle cheaper to run the route, and so forth... the route will come back.

I think this service will do fine. This is one the issues that SF has in a large part. Very hard to get around at night because public transit is dead before midnight.

Semantics. Profit, net zero, sustainable with tax dollars, etc.

Doesn't matter. You get the idea dude. It's about routes not being worthwhile because the damn transport option is too costly for just a few people. Autonomous car is gonna be more effective for these scenarios.

Even if you don't care about the revenue, having only a handful of people on the bus per run in the evening makes it hard to justify the cost of operating the line. Having other options to get home is great.

You're thinking about this the wrong way. There's immense value in getting _any_ commercial service out there, not just in PR but in a full end-to-end test of all the ops/logistical/product/etc that come from having to deal with customers in reality. A good chunk of a decade and billions of dollars have been poured into the research behind these services. I doubt anyone at Cruise or Waymo cares or even expects these limited services to be profitable, but that's entirely missing the point.

Great complement. Otherwise your comment becomes very confusing :-)

I doubt they're going straight for the meatiest market. A commercial service of any sort would be a huge coup for a self-driving car co (as the Tempe launch was for Waymo). In SF's case, it also provides a foothold in a market that's actually economically sustainable. Once the technology reaches the point where more complex and crowded scenes can be handled safely, expanding is relatively trivial compared to spinning up from scratch.

Waymo has had driver-free cars in Phoenix for over a year, so it would be strange for Cruise (who doesn't have a driver-free robo-taxi service already) to get a more permissive permit than Waymo.

in russia the rampant insurance fraud ended in everybody getting dashcams to defend themselves. waymo and cruise have way more than just a dashcam - if you can successfully commit insurance fraud while being recorded on a dozen cameras as well as lidar and infrared sensors, you probably deserve the settlement.

That doesn't mean enough people will hit it to matter.

Short answer: in the U.S., it will depend on the facts and circumstances. Common law has many drawbacks. But organic adaptability is one of its advantages.

While this is not exclusive of other liability, probably one or both of:

(1) The person hit, if they were breaking the law in a way which made it unreasonable to expect a driver to avoid hitting them,

(2) The manufacturer of the self-driving vehicle, under normal defective product liability principles.

The owner and/or, where different, operator of the vehicle, as well as other people in the chain of commerce may also be liable, especially when (2) applies.

You can't just stumble onto the highway where the normal traffic flow is well in excess of the speed limit and expect to get a payday because whoever hits you will be speeding.

That said, most of this goes out the window if you are committing a handful of crimes that society has decided exempt you from most of the protections of the legal system (terrorism, driving drunk, etc). If you are drunk and someone jumps in front of your car to commit insurance fraud it won't matter if you have 360deg 4k video of it, you will likely lose.

As is usually the case with the legal system, whoever has the least amount of political/economic power. So either the safety driver or the victim.

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

Possibly they give you a small credit for the trouble, and maybe ding the previous rider somehow.

A car with a spill in it, bodily or otherwise, is probably out of commission for the night. Depending on the type of stain and time before treatment, it could take the car out of the fleet for a few days. Eventually self-driving cars have bus seats? idk.

No problem, we can make the rider accountable. We have cameras in the car. First issue here is you can't trust users alone to provide accountability. Say I put my leftovers on the floor. You get in next and don't notice. The next rider however complains of a strong fish smell in the car. The car is taken out of service and your reputation is dinged since you were the last person in the car.

That's no good, so we'll need humans to review video and find culprits. Spinning up such a team will require software, managers, and oversight. Being alone with your date in a self-driving car is _exciting_ so you definitely don't want your employees to freely copy or share private videos of riders.

Like I said, this stuff is all solvable with centralized teams, but it cuts into the profit fantasy a bit. And it simply didn't exist when every car in your fleet came with a chaperone incentivized to take care of your/their car.

These are all solved problems.

Also, you think those engineers, capable of creating an autonomous car, somehow will find it too difficult to install liquid detectors all over the car and combine them with cameras inside to would allow for instantaneous, remote inspection of the car and imposing the penalties/taking the car offline – all of which are possible with existing technology?

How does the company prove this? Are there detailed criteria listed somewhere that these companies should then prove to meet? AFAIK, its all self reported. I mean, if the company is confident it can do it, I guess it's a good thing, but is it enough?

Waymo, Cruise, and who?

[0]https://www.dmv.ca.gov/portal/vehicle-industry-services/auto...

The value is now cutting the labor number by 1/3rd. Need a loader and unloader, no more driver. In food delivery this essentially removes your labor cost since the restaurant and consumer are the loader/unloader.

It must sure feel good to know that you would be an employee if anyone wanted to hire you when you're waiting in line for the dole.

How mad would you have to be to keep riding in human-driven taxis after the data shows AVs to be safer?

Have pilots.

> bicycles

You steer them.

At that point, would you choose to ride in one, or a human-driven taxi?

Freeway in LA in a self driving car? Absolutely not. Full stop. I've seen more variables play out on a given day than any training model can possibly consider with good support. If someone says they have good support with their model for an LA freeway I will laugh in their face. The freeway is chaos. You have people with scrap piled 12 feet over their pickup truck doing 40mph in the middle lane, people with BMWs and a cell phone in one hand doing 100 splitting lanes with no signals, outright debris on the road (my roomate hit a log in the middle of the 101 that took out another dozen or so cars one morning, how did a huge log even end up there?), not to mention construction conditions and less than stellar road markings during that making lane identification no small task for a self driving system.

On a little loop through something like a zoo, sprawling corporate campus, or as a vehicle to get me from my parked car to the front of the venue, absolutely I'll hop right into a self driving car. On the road with the current state of the road in my city? I'll let someone else take the literal hit today.

What we're discussing right now is riding in one on Bay Area streets where they're already proven their safe and orderly track record.

Computers are good at some digital things. They are not good at others.

> machine driven cars, which have already demonstrated far, far poorer safety than human driven ones

Citation needed.

Uber ran someone over, it turns out they were shockingly incompetent, and thankfully they don't have an AV program anymore.

Given some of my recent Uber drivers in the Bay Area, I reckon a blender would be a safer ride.

And obviously the launch is conditioned on regulators' belief that the system can handle reacting, to a degree of safety that's similar to human drivers: otherwise they wouldn't get the permit. The reaction time of safety drivers is not what this approval pivots on.

Though obviously the interesting question is: how did the company and the regulators get to this level of confidence in the system's safety

BTW, you seem to have been reading TFA as some sort of snark. It's a pretty well-established bit of internet jargon that's (at least in my experience) more winking than aggressive at this pt.

So they get to write the rules on how law enforcement interacts with their vehicles? Can I do that too?

This involves automatically pulling over and stopping the vehicle when a cop is behind the car and flashes their lights. The vehicle will also roll down its windows so the cops can access the interior of the vehicle and have easy access to passengers.

This also links in rider support automatically to communicate directly to law enforcement / rider support can also provide various commands to vehicle then.

They can also open a door, which triggers a an autonomos driving cutout. They can also contact these folks directly to gain access (ie, remote unlock etc).

To retrofit your vehicle so that law enforcement has this type of direct access may take some effort, but I doubt you'd have much objection - there currently is a real issue with drive offs and law enforcements ability to respond to those under vehicle chase rules. If you can set up your vehicle to override your drive-off efforts that will probably be welcomed, especially if it pulls over, stops, can unlock doors to provide LEO access to your person etc.

Separately, rider support can give permission for a vehicle search as the owner of the vehicle, so you'd want to register vehicle with a service that would give consent automatically and then provide instructions on where to find registration and insurance in vehicle. Waymo I beleive does it on the visors.

I'm not sure you fully grasp the implications of Law Enforcement Interaction Protocol efforts. In the future, the car may be able to drive you to a "safe" location for a car search and your arrest.

Police procedures should not be developed by way of an agreement between the state and a private company, who is in a position to error on the side of cooperation because they need a license. It is similar to the state granting a locksmith a license to practice under the condition that law enforcement be provided with copies to every key just in case they need them.

There are some serious Fourth Amendment considerations when traveling in a vehicle for hire without a driver. Is a Waymo closer to a private automobile with implied consent, or is it similar to a bus where you can deny consent to a search? Is the provider (or virtual driver) in a position to consent to a search of your property? Is the virtual assistant that opened the doors and consented to a search now in a position to be charged with transporting narcotics for distrubtion? Who's license gets dinged if the passenger has no seatbelt?

In other words, they WILL stop the vehicle, roll down the windows, unlock the doors etc for law enforcement, give permissions out for a search etc.

There is a separate question (unrelated to Waymo) around if that is enough for example for LEO to search the vehicle. My guess is the resolution will be that like a taxi, LEO can search the vehicle and waymo will build that implied consent into terms of service if needed. Hopefully patting down an individual will still require consent or cause or similar.

The OP seemed to imagine that these services are going to be some sort of lawless area, and they are not, they are going to track and record the start / finish / stops of all your trips, record you in the vehicle, record the space around the vehicle and will cooperate with LEO.

This is asking the manufacturers for a user manual, like how to disable auto-piloting etc.

I don't understand. Isn't "place both hands on the wheel, show me your license", etc precisely a "deliberate procedure"? It sounds like your complaint is that police are too procedural instead of engaging on a human level, but then you complain that theyre not _more_ procedural with humans?

With autonomous vehicles, they're asking how to interface with them politely. Like if they signal one to stop and it's not stopping, here is a phone number to call where someone can gracefully shut it down. As opposed to just deploying spike strips or ramming as is their default.

Whereas human drivers don't get the luxury of specifying any such protocols. And rather than reform their own procedures, police are attempting to compensate by creating protocols drivers are supposed to know and follow before an officer is even at the car.

As a motorist I'd love to be able to choose a protocol that police would follow when pulling me over. For example, call this phone number first rather than just walking up to me and risking yourself getting spooked and me getting killed.

This is quite an extraordinary assumption. Do you have any evidence of this that I missed? My assumption was closer to something like "here's how you communicate to an SDC that it should pull over; here's the decision path it will follow when pulling over; here's how to get in touch with the co".

I don't see any indication that an out-of-control car will be given free rein to rampage (in situations where a spike strip would be used on a human driver) while the PD waits on hold with the operating company. Honestly, this sounds like a paranoid fever dream.

Again, I think you could write a similar protocol. Pull over, roll down windows to allow cops easy access and visibility to your person, and unlock all the doors so they can open them. Maybe do a dash cam plus interior cam to get cops the footage they need that way?

Waymo also will automatically connect to a ride support operator who can do things like give permission for a vehicle search if needed. Not sure how you could copy that, maybe vehicle owners could put a sticker on their cars so regardless of what passenger wants the vehicle can be searched at any time?

And Tesla is rolling out FSD beta to cars based on drivers' safety score.

"Full Self Driving" does not exist. Every indication is that it never will.

I still can't believe no one's gone to prison over this. Tesla's "self driving" features are probably the biggest fraud in history with a body count to match.

Since most drivers only operate in certain fixed conditions often (e.g. driving to work or along certain routes during the day/night), people are going to get vastly different experiences of the same software.

This is why you get some people claiming FSD beta is a fraud and others claiming that it works well with 0 disengagements 99% of the time.

I think their tech is probably a LOT better than what you’re making it out to be but no where close to living up to its promise of full self driving.

I don't really trust anything Elon says anymore. I feel like he's not followed through on commitments about FSD far too many times. I'd personally never buy a TSLA until they get things much more in order (build quality/fit and finish, and actually deliver FSD)

Once the costs of ownership are well understood, that uncertainly will be resolved and the cars can be priced correctly. That will take time -- if I buy a 10 year old corolla with 100K miles I will know (after the usual mechanic inspection) what my expected maintenance costs will be, on average, for the next 5 years of ownership.

You can't really say that for an out-of-warantee Tesla because there isn't a large group of independent mechanics that can service it and because there is a lot of uncertainty about the battery life and service costs. Over time, if Teslas remain popular, and if the company doesn't DRM them out of existence, such networks will develop and we will accumulate actuarial data on repair costs that will reduce risk.

Until then, this market will remain primarily one of price insensitive or risk-insensitive customers -- e.g. wealthier people and possibly government or other large corporate fleet sales. That is exactly what you would expect for new technologies. But for most people, a car is a major proportion of their networth, one they need to get to work, and so they are going to be risk averse (which is exactly what they should be).

My next car will most likely be a 10 year old Lexus or Acura product. Even if it's a different manufacturer it will be an ICE vehicle and not an electric vehicle. Purchasing an out of warranty Tesla isn't remotely close to a viable option as it's not possible for me to price it. But used car prices drive new car prices because most purchasers are not planning on driving the car until it has zero value, they are counting on the used car market to pay for half of their car purchasing costs. Really for new technologies, Tesla should focus on leasing their cars if they want to reach out beyond the price insensitive market.

Where I live I see FJ cruisers, Porsches, and Bentleys every day, but I understand that my area is not typical. It's a wealthy area in which offroading is very popular. I also see Teslas and BMWs (more Mercedes than BMW and more BMW than Tesla).

I also understand that those cars tend to be more memorable, so it's foolish to pretend that FJ Cruisers, Porsches, or Bentleys are suitable for the vast majority of the public just because they seem suitable for many people in my area. In the same way, perhaps you should recalibrate your expectations as well when extrapolating based on what you see in your area.

Now what prevents a Porsche or Bentley from being suitable is that these are for niche performance or price insensitive markets. But what prevents something like the Tesla from being suitable is the charging network and repairability uncertainty. With a Bentley, you know you will pay through the nose for repairs which is why Bentleys depreciate so massively. But with Toyotas you know their reputation for durability, which is why used FJ Cruisers cost more than used Bentleys. But how much should a used Tesla go for? I honestly don't know -- and neither does anyone else right now. Unless you have a lot of money to play around with, that's too much of a risk. You don't spend 15-20% of your household net worth on something that you can't price.

I think you’re massively overthinking things. Most people look at the Model 3, see it costs the same as a Toyota Camry, want to help the environment and have Spotify in their car, skip the dealership, go to tesla.com, see dirt cheap financing available, and make the purchase. If they don’t it’s because they’re not comfortable giving up their gas car yet because that’s what they know. Not because they can’t “price it”. Further in the current market Teslas depreciate noticeably less than ICE vehicles because they don’t need maintenance.

You may not be able to price a Tesla, and that’s fine, but don’t get some impression that just because you can’t that other’s cant and moreover that it’s the sole factor contributing to why they allegedly don’t buy. Where I live nobody is “pricing” their $65k Truck, they just want a big fucking truck 30% of their household net worth be damned! You’re full of FUD.

Lastly looking at overall population numbers is kinda silly. We’re not going to wake up and everyone has replaced their perfectly fine car with a tesla. If you look at market share growth Tesla shot up in 2017 with the Model 3. And it wasn’t just a balloon, their growth has continued. They sold 500k cars last year, a little less than half of their previous total up to that year. Again: people are buying Tesla and more people by percentage are buying every year. A Tesla was a rarity 5 years ago but that’s no longer the case. The fact that we can observe this with only 1.5MM Teslas on the road is pretty cool.

Just to clarify your position, you want to hand over beta car driving software to people that do not first pass a test for safe driving?

I don't think that's accurate. Tesla collects a whole host of telemetry on all of their drivers. "Shadow Autopilot", etc., etc.

They show zero hesitation in releasing that telemetry to the media, or elsewhere, if there's the slightest hint it will put you under scrutiny rather than them. Even misleadingly so - witness a fatal accident where Tesla said "the driver was not paying attention - the vehicle had warned him about taking his hands off the steering wheel", neglecting the ever so slight detail that it had warned him, once, and that warning occurred a quarter hour before the crash.

Tesla collect all this data constantly, not just for entry into the FSD beta program.

Even if they arrive 1-2 years later than Waymo/Cruise, they can quickly dominate the driverless market.

This sounds like premature optimization. The real unsolved problem here is autonomous technology (software and hardware), not manufacturing equipped vehicles. No point having millions of vehicles if your self driving technology doesn't actually work.

> They do not rely on HD maps.

Precisely why FSD doesn't work and is still level 2. The AV players have shown fully autonomous driving needs HD maps and (inconspicuously) the only one not to graduate out of level 2 self driving is the one that's not using HD maps — Tesla. Creating/maintaining HD maps isn't nearly as hard as you think it is, but the benefits are incredible for safe driving.

It's tough to anchor Tesla's arrival based on the rest of the industry's progress. Cruise and Waymo are heavily dependent on Lidar; Tesla is solving a fairly different technological problem.

Tesla doesn't have competition because they aren't in the game at all. They do not make self-driving cars and probably never will.

When the robotic overlords take over, it will not be to thunderous applause nor loud protestations, but rather quietly beneath the surface silence of billions of people distracted by something more interesting or enticing.