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

New video of Tesla crash demonstrates the problem of semi-automated driving

theautopian.com

Post reply on HN
Print view View original

701–710 of 714 posts

Re: New video of Tesla crash demonstrates the problem of semi-automated driving

#701

Post by dkjaudyeqooe »

roguecoder wrote: post #140

Video understanding engineer here. Tunnels and underpasses are the worst. They are a pain in the ass, because shadows mess with all the edge detection and motion models and anything else visual. Humans compensate by thinking "I'm in a tunnel: things are weird." But without a reasoning model that can take context into account, the computer is stuck. In the video from behind, you can see the shadow ahead of the car on…

Is this sort of problem an argument for lidar or similar? I'm assuming you're describing a camera only system.

Re: New video of Tesla crash demonstrates the problem of semi-automated driving

#702

Post by watwut »

moduspol wrote: post #698
watwut wrote: post #685

Earlier quoted context omitted.

Delaying self driving cars does not come at the cost of human lives, if those cars are not safer then human driver. If they are less safe with current tech, delaying then saves lives. If they tighten regulation, they are saving lives. Also, mandating companies to put in more testing would force companies to test. Without that, they are economically motivated to make big claims, but to to not invest into reproducible…

The current tech is irrelevant. It is not safer than a human driver, but that's not relevant. If regulation delays the day in the future that it is safer than human drivers and that human drivers are replaced, then it is costing lives. And it would. The only way it wouldn't is if the regulation saves more lives in the meantime than would have been saved by having self-driving cars ubiquitous sooner, but that's tough…

The excess people killed by self driving cars matter too. Self driving is more dangerous until it is not and it is actual task of regulators to delay it until then.

And if regulators won't act, it will never be safe, because it is cheaper to produce unsafe products. It earns more to lie about safety of a product, unless you have functional regulations.

Re: New video of Tesla crash demonstrates the problem of semi-automated driving

#703

Post by Mawr »

Dylan16807 wrote: post #151
Mawr wrote: post #142

Earlier quoted context omitted.

The recommended following distance is "Never get closer than the overall stopping distance." [1], so even if the car in front somehow went from 70mph to 0 instantly, you should be able to stop. [1]: https://assets.highwaysengland.co.uk/stay+safe+stay+back/HE_...

Expecting people to follow that far is like expecting people to read every TOS.

The difference is that not reading the TOS will not kill you.

It's also not just some arbitrary rule, it's a fundamental principle based on the rules of physics.

Re: New video of Tesla crash demonstrates the problem of semi-automated driving

#704

Post by Mawr »

Spooky23 wrote: post #160
Dylan16807 wrote: post #151

Earlier quoted context omitted.

Expecting people to follow that far is like expecting people to read every TOS.

Dude, we’re talking about explaining away a Tesla failure here. Pedantic analysis of DMV regulations is absolutely essential. Clearly any dolt practicing defensive driving would anticipate the self driving car cutting him off and hard braking. Stories like this make people think that full self driving is actually a car the drives itself fully. The car is safer than any human and by critically thinking about this we a…

> Pedantic analysis of DMV regulations is absolutely essential.

If you think following the most basic driving safety rule is 'pedantic', please don't drive.

> Clearly any dolt practicing defensive driving would anticipate the self driving car cutting him off and hard braking.

The Tesla was 100% at fault, nothing can be done when you get cut off like that. But that's not the case for every single car behind them that failed to stop in time.

> Stories like this make people think that full self driving is actually a car the drives itself fully.

In my opinion Tesla has been criminally misleading with their avertising. The level of gaslighting they've been able to get away with is unbelievable.

Re: New video of Tesla crash demonstrates the problem of semi-automated driving

#705

Post by pnt12 »

MengerSponge wrote: post #320
kart23 wrote: post #316

Earlier quoted context omitted.

> think what's interesting here is it's likely the first instance where Tesla FSD has been involved in an accident which affected other drivers. Autopilot has killed multiple motorcyclists, and is suspected in many other cases, totaling 19 fatalities. This isn't the first, our regulatory bodies are just incredibly slow at this. https://arstechnica.com/cars/2022/08/tesla-faces-new-probes-... https://www.theverge.com/2…

But motorcyclists rank just above bicyclists, and just below deer. Nobody actually gets in trouble for hitting them on the road. I would love to see manslaughter charges for more accidents. If I do a whoopsie and stab someone in my home, I'm not going to get off with a "oh my god I'm so sorry! I was tired and it was foggy." People driving should be extended the same courtesy.

We cannot omit the context: in most places, people need to drive to work, to get groceries, and everything. A car is tremendously useful but a 5 second distraction can result in a death. You can't expect people who drive every day to be alert every single second, so there's a leeway when accidents happen.

That's a tradeoff society made.

But I'm all in favor of practicing more careful driving, and banning the shit of these incomplete automated driving mechanism!

Re: New video of Tesla crash demonstrates the problem of semi-automated driving

#706

Post by robbiep »

oblio wrote: post #591
robbiep wrote: post #447

Earlier quoted context omitted.

My Mazda has adaptive cruise control - it has absolutely revolutionised my highway driving. It is intuitive because the heads up display gives you a distance to the vehicle in front; and you can set the distance you want to maintain. It tracks your progress/distance so you can see when it will slow/maintain and it behaves gracefully if someone cuts in

I don't know how the Mazda one works, but for Volvo you actually get 1-pedal driving if you want. Set the speed a bit lower than your expected speed (for example in crowded areas), then accelerate as you wish. Once you take your foot off the gas, the car is set at a lower speed, so it brakes for you if needed. Ta-da! 1 pedal driving :-)

That’s interesting - haven’t seen that

Re: New video of Tesla crash demonstrates the problem of semi-automated driving

#708

Post by osigurdson »

ajross wrote: post #693
osigurdson wrote: post #676

Earlier quoted context omitted.

If it is ever necessary to take over then it is largely pointless. The only value is helping us get to the point where you can chill in the back seat.

Don't you feel the goalposts moving here? You went from "Humans never make mistakes like FSD does" to "FSD doesn't work yet" to "FSD works, but I don't think it has value yet". And you got called out on all of those and shifted. Maybe you could agree to settle on: FSD works pretty well, seems not to be unsafe vs. human drivers, but isn't finished, and is improving rapidly? Might have saved us all some time.

Moving goalposts and having a nuanced view are not the same thing. I agree with your statement above except that human drivers are still obviously better on average. Looking forward to when that changes.

Re: New video of Tesla crash demonstrates the problem of semi-automated driving

#709

Post by dang »

throwawayacc3 wrote: post #153
leaving wrote: post #105

This mirrors my experience of driving in America. People do not adjust their speed or following distance to the conditions. They drive like they have some sort of right to a road free of obstructions and to drive at the speed limit or above, no matter the circumstances. It's as if they expect the speed limit sign to adapt to the circumstances. They have the attitude, "Well if I need to slow down, why does the sign sa…

[flagged]

Nationalistic flamewar is not allowed on HN, regardless of which country you have a problem with.

You've been breaking the site guidelines badly in other contexts too. If you keep this up, we're going to have to ban you.

If you'd please review https://news.ycombinator.com/newsguidelines.html and stick to the rules when posting here, we'd appreciate it.

Re: New video of Tesla crash demonstrates the problem of semi-automated driving

#710

Post by sebastianconcpt »

JoeAltmaier wrote: post #663
sebastianconcpt wrote: post #581

Earlier quoted context omitted.

There is a "pass the test of time" factor involved. If natural selection approves, then you surely can be more confident and even pay less attention because time guaranteed that plenty of diverse cases didn't compromised the survivability of the system and the system needs you less to achieve its success rate as designed. Then it comes innovation. Which necessarily starts with limited cases. And we use to launch busi…

Disagree. We used coal to heat for generations, even though it killed thousands. We drive in traffic but 100 people die each day on the road in the US. It's all about social acceptance, and hardly about safety at all.

Right, 100 in absolute numbers. But that would be what fail rate as % when measuring total success cases of people pleasantly arriving to their destinations uneventfully?
Post reply on HN