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

Scenarios in which Tesla FSD Beta 9.0 fails

twitter.com

Post reply on HN
Print view View original

241–250 of 502 posts

Re: Scenarios in which Tesla FSD Beta 9.0 fails

#241

Post by Syonyk »

moralestapia wrote: post #181
Syonyk wrote: post #155

Earlier quoted context omitted.

I'm not sure radar would help - the pillar is a stationary object, and car based radar tends to drop all returns at the same rate as the car is moving forward. Radar isn't terribly precise, so you end up with tons of returns off signs, bridge abutments, etc. Remember, with radar, Teslas will happily autopilot themselves into a semi truck across the road, stopped fire trucks, concrete road barriers, etc. I see no reas…

Here's what happened another time someone argued that "radar can't detect stationary objects". https://news.ycombinator.com/item?id=16722461 Radar does detect stationary objects, come on.

Yes, radar can easily detect stationary objects.

Tesla's implementation of their radar processing drops all returns from anything stationary, given the forward speed of the car. In the context of a Tesla thread, I didn't think that needed to be specifically stated, but will do so in the future.

Re: Scenarios in which Tesla FSD Beta 9.0 fails

#242

Post by maverick-iceman »

moralestapia wrote: post #148
dnautics wrote: post #130

Earlier quoted context omitted.

Actually most of us can do any or all of the following things that Elon did (write a tweet calling a diver a pedo, write a tweet claiming that a stock will go to 420 huh huh, tweet random nonsense about cryptos, criticize the SEC or FAA, etc.) with very little consequence, if any.

Step 1: Become the CEO/major shareholder of a billion+ dollar company. Step 2: Publicly lie about some purported acquisition with the purpose of manipulating the stock of said company. We can talk after you've done that, from jail probably.

I think the SEC gives insiders and C-suite some kind of free pass as long as they don't time their 'overtlyoptimistic' takes on their company with their stock sales or vest.

The SEC is giving Musk free reign because they know that he can't leave the company because the two entities are so intertwined. Musk wealth is effectively just paper wealth.

The SEC doesn't care that investment bankers consider it real enough to give Musk loans for those pledged shares

They also don't care that the cult of personality he managed to create enable him to constantly not produce results and investors still give him money

In the end the SEC cares about real cash leaving the company coffins and into the owners pockets , not paper wealth swelling.

It's really a M.A.D. game between Musk and the SEC at this point...the SEC is willing to bet that it's not remotely possible that Musk lied through his teeth this whole time just to get to #1 in the paper wealth Forbes list only to implode while at the top.

Musk on the other hand doesn't strike me as a cold calculator such as Gates or Brin or Zuck, he is much more impulsive and there are non-zero chances that he did just that.

Re: Scenarios in which Tesla FSD Beta 9.0 fails

#243

Post by throwaway0a5e »

wedn3sday wrote: post #176
throwaway0a5e wrote: post #41

Earlier quoted context omitted.

The fact that a presumably average or not too far from average human can "weave at 90 mph through traffic yesterday, swerving back and forth between lanes inches from other peoples cars" should tell you how wide of a gulf there is between human operators and current FSD tech, which will dodge road debris only if it looks like a car, person, cyclist or traffic cone.

Yes, Im "pearl clutching" because some ahole decided to put hundreds of peoples lives at risk so he (and its always a he) could have a little thrill. There is a wide gulf now between human skill and robo-driver skill, just as 50 years ago it was much faster to call someone then send them a message in the mail. Now we have email, and in the future we will have (actual, real) full self driving.

>Im "pearl clutching" because some ahole decided to put hundreds of peoples lives at risk

The fact that you're saying he put hundreds of lives at risk when a hundred is a score that a terrorist with a semi truck would be proud of and is for all practical purposes unattainable with a light vehicle does seem to point in that direction.

I'm not condoning his behavior but 1) people are jerks, what do you expect and 2) you are getting way to bent out of shape over it.

> Now we have email, and in the future we will have (actual, real) full self driving.

And yet people pick up the phone for the urgent stuff and send snail mail for the super important stuff (though the latter is changing with online forms moving into more and more areas). That says something about the true nature of technological progress. We come up with new better stuff but often old stuff keeps some niches for itself.

Re: Scenarios in which Tesla FSD Beta 9.0 fails

#244

Post by xeromal »

Veserv wrote: post #219
cs702 wrote: post #174

Earlier quoted context omitted.

> ...doesn't matter how well it does in 90% of situations... Based on the full-length videos, I'd say it's more like 99.9% or even 99.99% for FSD 9.

The rate of driving fatalities is ~1 per 60 million miles driven. At an average speed of 60 mph that would be ~1 per million hours. Even if we were to assume there would be only one fatality-inducing error per hour, 99.99% success rate would make FSD 9 100x more dangerous than the average driver. I do not remember the exact numbers, but that would make FSD something like 1000x more dangerous than the classic example…

That's your estimation but we'd need to wait with actual statistics

Re: Scenarios in which Tesla FSD Beta 9.0 fails

#245

Post by jazzyjackson »

heavyset_go wrote: post #104
zyang wrote: post #16

It appears Tesla FSD just ignores things it doesn't understand, which is really dangerous in a production vehicle.

This is what the automated Uber vehicle[1] that struck and killed a pedestrian did, as well. Despite picking her up via sensors and the ML model, it was programmed to ignore them. [1] https://www.nytimes.com/2018/03/19/technology/uber-driverles...

Plus the Volvo’s radar-brake was disabled so they could test the vision system.

Re: Scenarios in which Tesla FSD Beta 9.0 fails

#246

Post by ethbr0 »

belter wrote: post #206
jtvjan wrote: post #144

Earlier quoted context omitted.

>the cyclist bad humor slecht humeur? That's "bad mood".

You are right. Was on mobile. Cannot edit it anymore...:-)

As an English first speaker, I didn't even notice it when reading. It's a somewhat odd phrasing (yet not exceptionally so) but still makes sense.

See second definition: https://www.merriam-webster.com/dictionary/humor

Re: Scenarios in which Tesla FSD Beta 9.0 fails

#247

Post by jazzyjackson »

LeoPanthera wrote: post #116
zyang wrote: post #16

It appears Tesla FSD just ignores things it doesn't understand, which is really dangerous in a production vehicle.

The vehicle is production but this particular software is not, it's a private beta and not available to the general public.

The general public does have the honor of being part of the beta test, in that they play the role of obstacles.

Re: Scenarios in which Tesla FSD Beta 9.0 fails

#248

Post by heavyset_go »

optimiz3 wrote: post #103
croes wrote: post #4

I doubt that autonomous driving will work in the next couple of years without infrastructure adjustments. We built roads for cars to replace horses, we will need to built roads for autonmous cars to replace human driven cars.

> need to built roads for autonmous cars Not any more friendly than they are to humans. AI driving needs to handle the general case. Otherwise a road that has a lapse in autonomous-friendlyness could be catastrophic.

We do a pretty good job of isolating certain freeways. There are freeways where I almost certainly never have to worry about a kid chasing a ball into the road and potentially hitting them, because there are 15ft walls on either side of the road, and it doesn't go through residential areas.

I could see some stretches of highways becoming automation friendly, but it wouldn't be anywhere near even a tenth of them.

Re: Scenarios in which Tesla FSD Beta 9.0 fails

#249

Post by foepys »

colinmhayes wrote: post #210
camjohnson26 wrote: post #190

Earlier quoted context omitted.

And what if it gets stuck in a local maximum and never improves? Then you lost those extra lives for no reason, which consequently is what I believe has happened. At least 4 people are dead in preventable autopilot crashes, and the real number is probably over 10. For the number of Tesla’s on the road that number is way too high. https://tesladeaths.com

4 deaths is worth a 1% increased chance of achieving self driving a month earlier. Way more than 400 people a month die in car accidents.

I somehow get the feeling that you don't actually want to save lives but rather want to experience the future™, whatever it costs.

Re: Scenarios in which Tesla FSD Beta 9.0 fails

#250

Post by rcMgD2BwE72F »

creato wrote: post #60

Some of the issues shown in these videos make me wonder about Tesla's strategy of using non-professional driver (customer) data to train FSD. Things like changing lanes at the last second is a thing that (obnoxious) humans do, and would be a bad example to learn from. There might be a lot of subtle garbage in Tesla's dataset.

Tesla can easily filter these cases out, automatically. They have triggers to catch things they want (e.g sudden lane changes) and campaigns have conditions too (only from prudent drivers, if lane change is forced, etc). FSD does not learn from all drivers all the time.
Post reply on HN