DMV tells Cruise to reduce its driverless vehicle fleet in SF by 50% | Hacker News Reader

DMV tells Cruise to reduce its driverless vehicle fleet in SF by 50%

sfgate.com

38 pointsgwbas1c37 comments

They definitely aren’t ready and should move away from operating in SF altogether. They should operate out larger cities with wider roads or limit where they operate in SF like bus lines.

I think this compromise needs to happen because SF didn’t plan for robotaxis when the roads were built decades ago. Towns like Phoenix and Las Vegas have the roads to support it. So they should consider splitting operations that way to continue working.

This does nothing but make other towns think the taxis aren’t ready and will fight to keep them out of their towns.

It’s an optics issue at this point if they want to convince people the robots can drive better than humans

I think its certainly a matter of when, not if robo-cars are capable of driving as good as/better than human. Sooner, rather than later too. That said, totally agree.

SF seems like the worst place to pilot this, for so many reasons.

Definitely a matter of when, but that’s also the case for nuclear fusion. Trouble is these perennial 5-10 year away things seem to always be later than we expect.

It’s probably the best from their training standpoint. The city is quite small relative to others and their only hope is to essentially memorize San Francisco in the neural networks. IMO you could not take a cruise and drop it into any other city and have it work.

what stopped them from memorizing a different small city in the US? there are tens of thousands to choose from.

it seems to me like their decision to start with SF instead of something a hundred times smaller was hubris

No, test your silicon valley death machines among yourself, please.

Are current cars driven by humans also death machines?

People can be held accuntable. Robocars have a more murky record.

But you still think that anytime you drive you make the decision to sit in a murder machine and no matter what the reason is any accident is your fault and you should go to prison? even if the breaks fail due etc

I can't speak to Cruise, but I strongly believe Musk set Tesla back a decade or more with his demand to use only video cameras.

We have lidar, infrared lasers, polarized ultraviolent imaging, multi-microphone 3D audio sensing, triangulating RF antennas. Why limit yourself to the limits of human perception?

I’ve heard that lidar is somewhat unfavorable because it’s so susceptible to becoming useless due to misalignments from something as simple as a pothole. Radar and other options though are absolutely worth using, I think musk is just trying to cut costs. Their S/X are using radar with the new FSD HW4

What makes you believe that? None of the research supports "as good as" being anywhere on the horizon, let alone "better than".

Cruise has been operating in Phoenix for many years now.

They’re too deep, their networks are overfit to San Francisco at this point. Making it work in other cities would require the insane training hours to basically memorize that city.

That’s not how it works at all

Nobody is training ml models to “memorize the city”. They use cm-precision maps

Should have gone to Dallas and its surrounding suburbs.

As an aside: Why are robocars practicing on wide, sunny roads in the American southwest? Shouldn’t they test on the hard cases? Rain, Snow, alleyways, streets, that exist since medieval or earlier times, traffic which takes a laissez-faire approach to rules?

What's really stupid about this entire thing with robo taxes is they're trying to solve the hardest problem where there's the most competition with humans

In my opinion, the most underserved part of taxi type transport is rural and semi rural areas where the distances are large. The costs are high. The human salary cost is extremely high.

But the driving is actually the simplest cuz there's just a lot of highway driving and pretty simple downtowns

Similar to how pilots often ascend from freighters to passenger aircraft, self-driving systems should have started driving freight through the desert. (Or mining pits, for that matter.)

Passenger operation as well as operation around people should have come much later, and step by step. Error rates must be determined step by step, and reliability ratings must be earned, similar to ETOPS.

I can only see one reason why they would begin with taxi operations, the driving task with the most visible showing of servitude, and that reason has to do with US history.

Urban taxis are also the most straightforwardly profitable part of the driving industry. Rural taxis can't be as profitable because they're spending more time in the driving phase (lower margin) and the distance between fares is much higher. Worse, almost all of your customers have alternative transportation that doesn't need to drive to their location that they'd rather use and the possibility of fare sharing doesn't exist to reduce prices. It's a difficult, low margin market to compete in, a terrible choice to demo your expensive and unproven technology.

I wonder if there's a density benefit? eg it's more valuable to have 10 cars drive the same stretch of road every day than to have one car drive a lonely highway every 10 days

>have one car drive a lonely highway every 10 days

This is obviously an availability issue. If it's available and cheap enough it will drive that stretch all day, since there are private cars driving there all day. The only reason it's barely used now is that it costs a days salary to drive to and from the next town.

I can see that self-driving in rural areas might be less of a challenge, but what happens when such a vehicle hits a town? Does it drive anyway or does it stop in a pre-assigned parking area? That's a logistical and financial burden for both the company providing the vehicle and anyone using it.

While figuring out how to get self-driving cars to work, there are problems beyond just the self-driving that involve a lot of staff. Examples:

- Supervised training drives - Rescuing stuck cars - Maintaining & refueling the cars

I imagine those efforts are easier to scale when concentrated in a large, dense urban environment.

But also like 5 people who need a ride once in a month.

Cities have lots of people, many that don’t bother with car ownership.

Also discussed here: https://news.ycombinator.com/item?id=37184904

I wonder how observable a 50% decrease in self-driving cars actually is. Recently it’s felt like the Cruise (and some Waymo) cars are EVERYWHERE in downtown SF.

I'm selling bumper stickers that say "Self Driving Cars Have No Soul"

How terrifying that must have been for all of the humans involved! You would think that emergency services vehicles would be equipped to broadcast shutdown codes to nearby driverless vehicles when their sirens are on. It’s my understanding that some intersections are equipped with devices to keep lights from changing when emergency vehicles (fire, police, ems) are moving through.

Is Cruise offering to pay for the devices and retrofit them onto all vehicles? Or is the city supposed to pay for it as a gift to Cruise's bottom line?

It is not the responsibility of emergency services to make their vehicles predictable and tractable so as to make up for the shortcomings of self-driving cars. It is the responsibility of self-driving car operators to make their cars able to deal with the unpredictability and unforseen events that may happen on the road, particularly when first responders are handling an emergency.

That’s completely wrong. There’s a reason fire trucks are painted brightly, have flashing lights and loud sirens, and can interact with intersections to flip lights: our governments have determined that these features are useful to make them conspicuous. Adding a sonic broadcast is just another doodad on the roof wired to the sirens and supported by laws to force autonomous vehicle manufacturers implement support.

> There’s a reason fire trucks are painted brightly, have flashing lights and loud sirens

Exactly. And why are those not enough for self-driving cars?

Why aren't bright paint and flashing lights enough? Why do we need loud sirens too?

The answer is that communicating over multiple spectrums turns out to be useful. Adding one more isn't the Herculean effort folks commenting on this proposal seem to think it is.

Why isn't it in our national best interest for AV companies to standardize on an emergency shutdown/slowdown protocol?

A cryptographically signed digital broadcast is a lot safer than relying on light/sound detection. It works in a variety of weather conditions and is much harder to spoof.

For the record, emergency vehicles are highly predictable and tractable: they have sirens and lights and will predictably not follow traffic laws when those indicators are on. It beggars belief that self driving cars are unable to correctly respond to such conspicuous signals.

This is a half of a good idea, but frankly every road mile traveled, airline flight, walk to the grocery store creates some number of dead people - I find this reaction from SF to be concerning because we appear to be placing higher standards on AVs than we are on humans.

the burden should lie with Cruise. instead, in typical tech fashion, they “move fast and break things” except now it’s interfering with emergency services, blocking traffic randomly, etc.

Reply on news.ycombinator.com