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

Tesla loses U.S. designation for some advanced safety features

reuters.com

Post reply on HN
Print view View original

251–260 of 379 posts

Re: Tesla loses U.S. designation for some advanced safety features

#251

Post by ska »

arcticfox wrote: post #211
ffggvv wrote: post #157

Earlier quoted context omitted.

theres actually no proof of the claim that its safer or even as safe as human driving. The only "data" is from tesla themselves. Which will hurtle you 200mph at a wall and then disengage autopilot 100ms before hitting the wall and not count it as an autopilot death.

> Which will hurtle you 200mph at a wall and then disengage autopilot 100ms before hitting the wall and not count it as an autopilot death. This is false: "To ensure our statistics are conservative, we count any crash in which Autopilot was deactivated within 5 seconds before a crash." https://www.tesla.com/VehicleSafetyReport It would be more helpful to focus on actually deceptive issues in how they report data, nam…

You are correct about the 100ms (although 5s is too short, should be something like 30 given our understanding of task switching) but it's worth pointing out that the comparison you quote (and all those like it, not just from Tesla) is disingenuous.

All the self driving companies do this, but it only makes sense to compare # of accidents/mile between like processes. It's an ok proxy for total risk iff accidents were uniformly distributed, but we know that isn't even close to true.

With a system like Teslas', it is preferentially, probably approximately exclusively if normalized by distance, deployed in situations with a lower accident rate - comparing that to total rates is nonsense.

What would be more interesting is Tesla vs. human accident rates on only freeway miles, say, but I haven't seen that anywhere, have you?

Best rough stats I can find is that less than 1/2 of all accidents happen on highways. There are a also a ton of fender-bender, parking lot and/or traffic jam accidents that aren't really applicable either.

Re: Tesla loses U.S. designation for some advanced safety features

#252

Post by tzs »

gauravphoenix wrote: post #58

>removing radar sensors to transition to a camera-based Autopilot system A few weeks back, I had a terrible experience while using auto pilot. I was driving on a highway (in CA) with autopilot engaged. For the most part, there was a concrete median on the highway. Suddenly, a section came with no concrete median and a new left only turn lane gets added. For whatever reasons, autopilot thought it is a great idea to su…

> I haven't engaged autopilot since then. I can't trust auto pilot anymore- it couldn't deal with a dead-simple scenario of a clearly marked lane getting added.

I have a question about Tesla Autopilot.

Looking at the common driver assistance features that are common nowadays even on many entry level models from most car companies, I see adaptive cruise control (often with low speed following), warnings if you start to drift out of your lane (sometimes with minor steering nudges), warnings if you turn on a turn signal when something is in your blind spot, and similar things. They are usually separate features you can individually enable or disable.

I assume Tesla does those things, too, but are they part of Autopilot or are they ? If they are part of Autopilot, do you lose them when you decide not to engage Autopilot or can you turn on and off individual features of Autopilot?

Put another way, if I found myself driving a Tesla, and only wanted it to only use those driver assist features that I'd find if I were driving a low to mid level Honda Civic or Toyota Camry, is that possible?

Re: Tesla loses U.S. designation for some advanced safety features

#253

Post by JoshGlazebrook »

goshx wrote: post #200
t0mas88 wrote: post #78

Earlier quoted context omitted.

They're disabling some functions due to the switch and promising to bring them back with an update. So that's really a fix for something that isn't working anymore. If those features are important to you, it makes sense to wait for the update to be released before you buy. Don't trust a Tesla promise, wait for the feature to be there in reality. It took them a very long time to make the promised automatic wipers work…

"Autosteer will be limited to a maximum speed of 75 mph and a longer minimum following distance. Smart Summon (if equipped) and Emergency Lane Departure Avoidance may be disabled at delivery." Are these really important features that you would postpone a purchase of the car? With the exception of the "Emergency Lane Departure Avoidance", no other car in the industry has these features anyways.

Many states have highway speed limits that _begin_ at 75mph. So yes this would mean autopilot (a huge selling point of these vehicles) is useless for highway driving, which is what autopilot in its current incarnation is meant for.

Re: Tesla loses U.S. designation for some advanced safety features

#254

Post by imoverclocked »

jfrunyon wrote: post #219
ghoward wrote: post #96

Earlier quoted context omitted.

More people need to see your comment because there seems to be a belief that full self driving is just around the corner. However, IMO, it will never happen because it needs a full artificial general intelligence. In your case, the new lane was clearly marked, which just shows how weak the system really is. Imagine if there was snow covering the lane markings. Or sunlight reflecting off of the road confuses the visio…

It doesn't actually need full general AI, but it does need to know and understand all of the roads in its "service area". IMHO the real solution will be building a database of roads (and lanes, speed limits, traffic lights, stop signs, etc). BTW, the average driver gets in a crash every 18 years. The best humans aren't perfect either, and a self-driving car really only needs to be better than the worst humans to be a…

> a self-driving car really only needs to be better than the worst humans to be an improvement

I think this is false because varying abilities of people will use the autopilot, not just the worst. The autopilot needs to be better than average for an overall improvement. I think it needs to be much better than average to be compelling enough for most people to adopt it. Most people consider themselves to be good drivers, even with sufficient evidence to the contrary.

Re: Tesla loses U.S. designation for some advanced safety features

#255

Post by osmarks »

IncRnd wrote: post #147
wombatmobile wrote: post #104

Earlier quoted context omitted.

What's meta disturbing about autopilot is the argument that it could save X lives a year, by cutting the road toll from Y down to Z. For that statement to be a validation for legalising autopilot for open usage, the bar for Z needs to be a lot lower than Y. It needs to be zero or else we would be blessing fatal bugs.

Related to what you wrote, I strongly disagree with the argument that some people make, where it is said X lives were saved. But, what they are really saying is that Y people, who wouldn't otherwise die, were killed by the new technology, yet overall fewer died! What good is any technology if it is known to kill people who would otherwise be alive?

Why does the particular set of people matter instead of just the difference in numbers? If the Y people happened to be a subset of those who would die without the technology, is that better? If you somehow move all the deaths caused by a thing to a different but equally large set of people, is that bad because those people "would otherwise be alive"?

Re: Tesla loses U.S. designation for some advanced safety features

#256

Post by clouddrover »

gmadsen wrote: post #241
clouddrover wrote: post #221

Earlier quoted context omitted.

> Waymo does not have these issues at all. False: https://jalopnik.com/watch-a-waymo-av-get-freaked-out-by-tra...

did you read the article? that is not related to the tesla issue at all. Furthermore, this article just described bad decision making with regards to traffic cones, however was not dangerous. It was aware of drivable surfaces and direction of traffic... because it has a map Do you see how these are different situations? 1. Telsa swerving into an oncoming lane 2. Waymo coming to a controlled stop in front of a traffic…

> Waymo coming to a controlled stop in front of a traffic cone

And then driving off again, and then blocking traffic, and then evading Waymo's roadside assistance, etc.

Do you see how these aren't different situations?

Re: Tesla loses U.S. designation for some advanced safety features

#258

Post by JoshTko »

dangrossman wrote: post #247
JoshTko wrote: post #235

Earlier quoted context omitted.

Interesting. Do you have a source that confirms this?

I searched "Tesla recognizes FSD revenue" and found several immediately. https://www.tesmanian.com/blogs/tesmanian-blog/tesla-autopil... https://loupventures.com/teslas-software-advantage-is-clear-... .

Parent stated that Tesla is using party tricks to recognize a larger than merited portion of FSD revenue. Your links do not make this case.

Re: Tesla loses U.S. designation for some advanced safety features

#259

Post by wpietri »

jfrunyon wrote: post #236
wpietri wrote: post #136

Earlier quoted context omitted.

Yup. Rodney Brooks, famed robot scientist and iRobot founder, says not to expect driverless taxis in cities earlier than 2035: https://rodneybrooks.com/my-dated-predictions/ I don't think we quite need AGI to make something like a commercial urban taxi service work. For any given place, we can substitute a very large amount of training data. And I think we can get there through an approach like Waymo, where the taxi'…

We've already seen several instances of Waymo's approach not working. I agree that ultimately they need a significant amount of local training data. But I believe they would be better off having a human drive a car around to map things out, run their 'AI' on the results, and then manually review/correct lower confidence results. Or, better yet, make a database and have cities provide feeds of changes/permits/closures…

Depending on what "not working" means to you, I'm ok with that. I think the point of the Waymo approach is to not work very frequently so they can flush out the issues.

Re: Tesla loses U.S. designation for some advanced safety features

#260

Post by Syonyk »

jfrunyon wrote: post #236
wpietri wrote: post #136

Earlier quoted context omitted.

Yup. Rodney Brooks, famed robot scientist and iRobot founder, says not to expect driverless taxis in cities earlier than 2035: https://rodneybrooks.com/my-dated-predictions/ I don't think we quite need AGI to make something like a commercial urban taxi service work. For any given place, we can substitute a very large amount of training data. And I think we can get there through an approach like Waymo, where the taxi'…

We've already seen several instances of Waymo's approach not working. I agree that ultimately they need a significant amount of local training data. But I believe they would be better off having a human drive a car around to map things out, run their 'AI' on the results, and then manually review/correct lower confidence results. Or, better yet, make a database and have cities provide feeds of changes/permits/closures…

> Or, better yet, make a database and have cities provide feeds of changes/permits/closures.

"Sorry, I know you crashed, your car's on fire, and you're bleeding out. But I need to get this road closure form submitted before I can help, or a self driving car might ignore the stationary accident and make things worse. Ugh, 5G's not working today... hold on..."

The cars must be able to handle random, unexpected events happening without maps, and without external guidance. A wheel falls off a truck in front of you for metal fatigue reasons and is bounding down the road - that's a thing that happens, and a thing that in no way will be reported before it becomes relevant.

If a self driving car can't handle this sort of thing (or an accident and coming on a flagger) autonomously, it has no business being allowed on the road.

And if it can - then there's no need to have a centralized database of weird things, because it can deal with the range of events autonomously.

Post reply on HN