Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Driverless robotaxi fleet paralyzed for hours in San Francisco

thelastdriverlicenseholder.com

Post reply on HN
Print view View original

331–340 of 514 posts

Re: Driverless robotaxi fleet paralyzed for hours in San Francisco

#331

Post by MauranKilom »

gambiting wrote: post #280
oconnor663 wrote: post #224

Earlier quoted context omitted.

Of course it's a cause for concern. It's a perfectly reasonable concern. It's just about certain that early fleets will have big outages for dumb reasons, just like major websites do. The question will be how frequent those are, how dangerous they are, and how effective the mitigations will be as everything matures. As with many of things related to self-driving cars, the target doesn't have to be "perfection" so muc…

>>the target doesn't have to be "perfection" so much as "better than people on average". You see, I completely disagree with this assumption. They should be perfect or they shouldn't be allowed on the road at all. You wouldn't accept a bread slicer that sometimes erroneously slices off your fingers, as long as it does so less frequently than people slice their own fingers on average, would you?

> You wouldn't accept a bread slicer that sometimes erroneously slices off your fingers, as long as it does so less frequently than people slice their own fingers on average, would you?

And yet I have certainly seen bread slicers that are exactly like that. Possibly even worse than "other hand with knife". Maybe the analogy is also not very helpful though, because (at least for those bread slicers I'm thinking of) the user is still majorly in charge of his own safety, unlike FSD.

Re: Driverless robotaxi fleet paralyzed for hours in San Francisco

#332

Post by henryteeare »

_flux wrote: post #307
carlmr wrote: post #264

Earlier quoted context omitted.

It could be a middle ground kind of thing. The cars have to trust something . For position you have GPS + some interpolation based on your movement sensors + some buildings/streets and maps (also unreliable since construction etc changes a lot, a lot of buildings and intersections look the same, etc). Computers are very far away from understanding where they are from context alone. Hell most humans don't know where t…

Dead reckoning works surprisingly well at least in Teslas, of which the precision can be seen when they move around in underground facilities and finally resurface: the car pops up pretty close to the entry point. In any case, aren't all self-driving vehicles equipped with lidars and work in pre-mapped regions? I would 100% expect they are able to just use those maps with lidar to precisely geolocate themselves, prob…

> aren't all self-driving vehicles equipped with lidars

I believe Tesla has made a lot of noise about not using lidar.[0]

[0]: https://venturebeat.com/2021/07/03/tesla-ai-chief-explains-w...

Re: Driverless robotaxi fleet paralyzed for hours in San Francisco

#333

Post by lifeisstillgood »

I think the best use case for self driving car tech is not self driving but teaching the humans how to drive better "You should slow down approaching that junction two seconds earlier because the last ten times i noticed a car approaching from left that you were not seeing"

Re: Driverless robotaxi fleet paralyzed for hours in San Francisco

#334

Post by fest »

zaarn wrote: post #282
mh- wrote: post #167

Earlier quoted context omitted.

EDIT: the following is incorrect per a reply below. Will leave the original comment below. --- That was going to be my hypothesis. GPS resolution can be reduced in a conflict area. I'll defer to someone with actual knowledge to add more info. :)

Reduction of GPS Resolution is limited to specific areas, it can be targeted. If they are doing it (probably are in Ukraine) then it only affects some areas. You can also bypass that restriction if you have the right military keys (read: Ukraine).

And your systems actually support decryption and entering the said keys.

You might be surprised what proportion of the hardware that operates in Ukraine (from both sides) is using civillian grade GPS receivers (i.e. almost anything with a price tag of less than 1M USD).

Re: Driverless robotaxi fleet paralyzed for hours in San Francisco

#335

Post by BoorishBears »

rob_c wrote: post #325
bjoli wrote: post #316

The promises of self driving cars have always seemed weird to me. Sure, I would love to have a world where all cars are self driving. The security benefits seem large. But decreasing congestion? We already know how to do that, and in many ways it seems like a safer bet than some kind of technogy with large problems and even larger unknowns; public transport. Cheap, well developed, reliable public transport will beat…

> Cheap, well developed, reliable public transport will beat autonomous cars on throughput and resource use any day. Remember, this is happening in a state with massive multiple public health crisis that go un-reported and has maps of human excrement in their inner-cities. The priorities here are completely in the wrong direction I agree.

Are you in a state without an underreported health crisis?

Re: Driverless robotaxi fleet paralyzed for hours in San Francisco

#336

Post by IAmWorried »

SheinhardtWigCo wrote: post #304
bee_rider wrote: post #203

Earlier quoted context omitted.

It couldn't possibly be GPS, right? Networks are unreliable, the cars must be able to reach a safe state if the network drops out. They appear to have stopped in the road. If these cars can't safely navigate their way to the side of the road using local sensors, they are wildly, hilariously unsuited for autonomous operation. > at least a dozen autonomous Chevrolet Bolts from GM Cruise Automation were spotted blocking…

You've nailed the dirty little secret of the AV industry. Everyone's safety model is underpinned by the ridiculous assumption that a human will be (a) reachable with very high reliability over shitty cellular networks; (b) able to correctly interpret complex situations with extremely limited information; (c) able to resolve safety-critical situations by making planner adjustments, without direct control over steering…

Wow. I had no idea that these things had such fundamental limitations. I'd feel a lot less safer using one now, to be honest. If you had to estimate, how long would it be before AVs would be able to do all their calculations locally? Is that something that's even under consideration at the time?

Re: Driverless robotaxi fleet paralyzed for hours in San Francisco

#337

Post by jfoster »

redisman wrote: post #238
unnouinceput wrote: post #153

Earlier quoted context omitted.

1 - they burn in atmosphere 2 - each satellite is under observation by a team 3 - how many satellites do you think are there actually? Not that many, unlike the driverless cars we are setting up nowadays 4 - Earth is actually big. Satellite pouring down all at once and you'll have maximum like 100 casualties - worse case scenario if 1). doesn't happen. Cars OTOH are in busy cities. One battery to explode and can easi…

> One battery to explode and can easily do 100 casualties That’s definitely not true. A 66KWh battery is not a tactical nuke

I think autonomous cars are something that must happen, and the sooner the better, but I am also a bit concerned that they could represent the ultimate weapon if a malicious actor gained control over them.

One battery fire isn't a problem, but if every 2nd car in a city or country crashes into something in a way that starts a fire, that becomes a very big problem extremely quickly.

Re: Driverless robotaxi fleet paralyzed for hours in San Francisco

#338

Post by valdiorn »

CTDOCodebases wrote: post #292
carlmr wrote: post #264

Earlier quoted context omitted.

It could be a middle ground kind of thing. The cars have to trust something . For position you have GPS + some interpolation based on your movement sensors + some buildings/streets and maps (also unreliable since construction etc changes a lot, a lot of buildings and intersections look the same, etc). Computers are very far away from understanding where they are from context alone. Hell most humans don't know where t…

Could it be GPS spoofing that is causing this then?

Malice, incompetence, all that jazz...

Re: Driverless robotaxi fleet paralyzed for hours in San Francisco

#339

Post by gambiting »

Peleus wrote: post #300
gambiting wrote: post #280

Earlier quoted context omitted.

>>the target doesn't have to be "perfection" so much as "better than people on average". You see, I completely disagree with this assumption. They should be perfect or they shouldn't be allowed on the road at all. You wouldn't accept a bread slicer that sometimes erroneously slices off your fingers, as long as it does so less frequently than people slice their own fingers on average, would you?

The bait and switch with this comparison is you're comparing something you have no control over (bread slicer which randomly malfunctions) against something you have full control over (my own cutting abilities). No one thinks their cutting ability is inferior, so of course they don't want to give up control, because they are (obviously) better than the average human. If I posed the question as: Would you prefer to pl…

Ok, maybe a different, well known example - the radiation Theraphy Therac-25 machines which would sometimes(extremely rarely) deliver a lethal dose of radiation instead of the one entered. Undoubtedly those machines saved more lives than they took, and also it's without question that they did so more accurately than any human operated machine could ever do. Yet they are considered one of the biggest failures of medical engineering ever.

I see it the same way - the current "self driving" systems shouldn't be allowed on the road, period, no matter how much safer they are than a statistical driver, unless they can be demonstrated to be completely 100% perfect in all scenarios(other than actual hardware failure - they shouldn't for instance run into a truck that's turning across a highway just because the system chose to ignore it).

Re: Driverless robotaxi fleet paralyzed for hours in San Francisco

#340

Post by astrange »

lelanthran wrote: post #305
astrange wrote: post #277

Earlier quoted context omitted.

Cause the post implied that it's right to be mad at Tesla, but that wouldn't help. Feel free to give the car back to the dealer. Although I'd think this would be obvious on the test drive. Also, being livid is probably bad for your health. Might get high blood pressure.

> Cause the post implied that it's right to be mad at Tesla, Of course it's right to be mad at Tesla - they resolved a property dispute by force. > Also, being livid is probably bad for your health. Might get high blood pressure. I wouldn't be livid in that situation, but I would consider criminal charges against Tesla in my jurisdiction. After all, the dealer is only on the hook for civil claims, Tesla is on the hoo…

Oh, I missed that AP was enabled at the dealer but then disappeared - thought the dealer had just claimed it was there.

That raises a lot of questions about how it could even happen. I'd understand if it was sold as some kind of "user license", but it's on the car, so I assumed it was a permanent extra hardware addition.

Post reply on HN