Cruise is opening driverless cars to the public in San Francisco • Hacker News Board

Re: Cruise is opening driverless cars to the public in San Francisco

#451

Post by seanmcdirmid » Wed, Feb 02, 2022, 5:00 AM UTC

ck2 wrote: post #441

I still want to know who goes to prison when a pedestrian or cyclist is killed. Because no-one is arguing someone will be killed, only that "human drivers kill people too" or "machines will kill fewer, eventually". Will it be the passenger? The coder for bad code? The person who assembled the hardware for improperly tested malfunctioning sensors? The person who made the malfunctioning sensors? Who? Because someone is…

Unless the driver is drunk, has malice, or willfully breaking a bunch of other rules, I don’t think they go to prison now.

Re: Cruise is opening driverless cars to the public in San Francisco

#452

Post by kgin » Wed, Feb 02, 2022, 5:00 AM UTC

mwattsun wrote: post #203

> Today we are opening up our driverless cars in San Francisco to the public I doubt that means what they hope we think it means. So your employees have been taking your carefully designed ride and now you've let a couple of non-employees ride the course. If it means more than that, especially allowing random rides with random passengers, I'd be very surprised.

Get ready to be surprised. The routes are created by the car on the fly. There is no preset course. The service area is bounded by the current permit, but within those bounds, the car can go anywhere.

Re: Cruise is opening driverless cars to the public in San Francisco

#453

Post by kgin » Wed, Feb 02, 2022, 5:02 AM UTC

TulliusCicero wrote: post #404

dontreact wrote: post #341

Earlier quoted context omitted.

I saw a presentation at a conference that pretty strongly suggested they will have a remote driver take over at certain points (the car will assess it can’t continue and remote driver takes over)

A remote driver, or remote navigator/coach? The former is incredibly dangerous, the latter is what Waymo is doing.

Second one. Someone remote can help label and plan until the car has enough confidence to proceed.

Re: Cruise is opening driverless cars to the public in San Francisco

#454

Post by TulliusCicero » Wed, Feb 02, 2022, 5:02 AM UTC

kgin wrote: post #453

TulliusCicero wrote: post #404

Earlier quoted context omitted.

A remote driver, or remote navigator/coach? The former is incredibly dangerous, the latter is what Waymo is doing.

Second one. Someone remote can help label and plan until the car has enough confidence to proceed.

Right, which isn't a remote driver.

Re: Cruise is opening driverless cars to the public in San Francisco

#455

Post by jowday » Wed, Feb 02, 2022, 5:05 AM UTC

Retric wrote: post #445

jowday wrote: post #435

Earlier quoted context omitted.

> Cars really don’t need to identify what something is just it’s location and movement which is a vastly easier problem. A trash can rolling down the street can be treated just like an oil drum doing the same thing etc. You’d think that, until you encounter something like a turn restriction sign with a bizarre conditional restriction that it’s never seen before. At which point the car needs to OCR the text, parse the…

Or treat that turn restriction as applying 100% of the time.

And now we’re already making concessions about the car’s abilities.

There are 10 MPH speed limit signs on Market Street in SF that specify in incredibly small text “when behind trolleys”. Assuming we take your approach, the car will just always go down market at 10 MPH.

Imagine if it’s a negative turn restriction - IE, it’s permitting turns except for during certain hours and conditions. Now the car is treating it as always permitted and turning into traffic. An edge case, but something it’s going to encounter in the real world.

Re: Cruise is opening driverless cars to the public in San Francisco

#456

Post by exhaze » Wed, Feb 02, 2022, 5:07 AM UTC

mwattsun wrote: post #387

gxs wrote: post #361

Earlier quoted context omitted.

It might actually do a lot better than you and I. Remember, it's equipped with radar so it can probably see cars the next block over.

radar doesn't see around corners. At any rate, would have to see some evidence that it can handle the worst of The City before I took a ride, like the incessant construction, steel plates on the road covering new holes, taped off lane changes, resulting traffic jams where merging is happening and so on. It's a really difficult environment for a human and I've never heard the claim yet that self-driving was better tha…

Cruise publishes a lot of their AVs driving around SF for over an hour. In a video from their YouTube channel [1], at around 22:29, you can see exactly how they handle a steep 4-way intersection. You can also see on the above visualization what objects their system sees as it is performing the maneuver.

The evidence is easily accessible - next time I suggest you just research a little bit first instead of going straight to speculating.

[1] https://youtu.be/HiG__iqgYHM?t=1349

Re: Cruise is opening driverless cars to the public in San Francisco

#457

Post by JoshTriplett » Wed, Feb 02, 2022, 5:07 AM UTC

comex wrote: post #60

gitfan86 wrote: post #32

Earlier quoted context omitted.

What is success? Does 50 miles of geofenced and daily mapped streets mean Cruise won self driving? What if Waymo gets to 20k miles of geofenced roads and monthly mapped? What if Tesla gets to the point of one intervention/crash every 100k miles? 10M Miles?

The human accident rate is about one per 500K miles, so if they were able to get in that range, then yes, they would have succeeded; drivers would be able to stop paying attention to the road without putting themselves and others in danger. But the current FSD beta's intervention rate is more like one per 10 miles, judging from some quick googling. I see no particular reason to assume that incremental improvement can…

> The human accident rate is about one per 500K miles, so if they were able to get in that range, then yes, they would have succeeded; drivers would be able to stop paying attention to the road without putting themselves and others in danger.

Unfortunately, I expect that automation will be held to a higher standard than human drivers, rather than the same standard. When an accident happens, people want to know who to blame, and an unimpaired human driver gets somewhat more latitude for a genuine accident, while a piece of software is always going to be perceived to be at fault (which it may well be, even in a situation where a human wouldn't be considered to be). And conversely, people (somewhat validly) want to have more control: every driver thinks they're above average, and the software won't be as good as their accident rate, and if something happened at least they were in control when it happened.

I don't necessarily even think those are incorrect perspectives; we should hold software to a high standard, and not accept "just" being as good as human drivers when it could be much better. But at the same time, when software does become more reliable than human drivers, we should start switching over to make people safer, even while we continue making it better.

(Personally, I wish we had enough widespread coordination to just build underground automated-vehicles-only roads.)

Re: Cruise is opening driverless cars to the public in San Francisco

#458

Post by wutbrodo » Wed, Feb 02, 2022, 5:09 AM UTC

ummonk wrote: post #222

comex wrote: post #60

Earlier quoted context omitted.

The human accident rate is about one per 500K miles, so if they were able to get in that range, then yes, they would have succeeded; drivers would be able to stop paying attention to the road without putting themselves and others in danger. But the current FSD beta's intervention rate is more like one per 10 miles, judging from some quick googling. I see no particular reason to assume that incremental improvement can…

What is the appropriate point of comparison though? All human drivers? Sober human drivers? Sober cautious human drivers? Sober cautious human drivers with driver assistance technology (e.g. auto-braking and blind spot warning, or potentially even more sophisticated LiDAR tech)?

I don't think this question is even meaningfully-defined. There is no "the" point of comparison. The relevant point of comparison is whatever ride it's displacing.

The rideshare explosion has already had a measurable effect on drunk-driving deaths; to the extent that a theoretical lower-cost AV will make rideshare even more accessible, then its effect on drunk-driving reduction absolutely makes non-sober drivers a relevant comparison.

For an average young person who'd get in the car with one of their friends, or drives a bit recklessly themselves[1], an AV at sober-human-driver level would be valuable.

For a guy who needs his kids driven around, a "sober cautious human driver" level of safety may feel right.

For questions like "what should the regulatory bar for launch be", all human drivers seems like an easy answer.

[1] I'm probably guilty to a degree here, on the rare occasions I drive

Re: Cruise is opening driverless cars to the public in San Francisco

#459

Post by ryandrake » Wed, Feb 02, 2022, 5:10 AM UTC

alex_sf wrote: post #434

walrus01 wrote: post #399

Earlier quoted context omitted.

True. I actually replied to the entirely wrong comment, should have been one up... I have a great deal more confidence in truly revolutionary things that have been factually delivered by SpaceX (such as 100+ re-uses of a rocket first stage now), and I question how much of that is really due to Musk at all. Maybe Musk as a figurehead. I wonder if all of the Elon fanboys know that much of what's been accomplished at Sp…

Does it really matter? If there's no Elon, there's no SpaceX. And I say this as someone that doesn't buy into the founder cult.

> If there's no Elon, there's no SpaceX.

Why not?

What mystical gift does this one person have that 7 billion other people don't, that permits him and only him to run the company? This is a really unpopular opinion on a web site that exalts founders, but I don't think it really takes much special skill to run a company. Most (but admittedly not all) CEOs are in their position not because of their know-how, but because 1. They founded the company, and happened to be the one that flipped a coin heads 20 times in a row; or 2. They were born into that Ivy League class that closely gatekeeps CxO and SVP positions for themselves; or 3. Were descendants of one of the above.

Assuming a successful CEO is uniquely skilled is like assuming a lottery winner is uniquely skilled at winning the lottery.

I think many people, if given Elon’s financial war chest and basic knowledge of and an interest in rocketry, could have made SpaceX.

Re: Cruise is opening driverless cars to the public in San Francisco

#460

Post by Pxtl » Wed, Feb 02, 2022, 5:10 AM UTC

honkdaddy wrote: post #313

potatolicious wrote: post #259

Earlier quoted context omitted.

Agreed - I think the "but can it drive in a whiteout blizzard" question is best redirected to: "can a human?" I suspect there are operating conditions the AVs won't solve acceptably - some of those conditions IMO are also conditions where we should not accept a human to solve acceptably. In general I feel we have a lackadaisical culture around driving that encourages/excuses unsafe behavior, and is overoptimistic abo…

I grew up in a part of Ontario where total whiteouts can happen fairly frequently on both major highways leading in and out of my small town. It is in fact possible to drive dozens of km in near or total whiteout conditions simply by the hazard lights of the car ahead of you. You very frequently will see lines of cars km long, all going I don't think vision-based FSV will ever reliably handle winter conditions like t…

The worst is when one goes into a ditch and the car following them follows into the ditch because their main indicator of where to go was the running lights and tire tracks of the car in front.