Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register
This site may earn commission on affiliate links.
It is not possible to turn into that driveway with a trailer without turning from the left lane. How could it be outside the training data if I’ve seen it 100 times in my lifetime? I’ve only driven 500k miles!
The issue is probably that NNs require an absurd amount of data to “learn” anything.
Well, I've never seen it before. I don't think anecdotes can determine edge cases or not. For example if someone has a neighbor in their neighborhood that has a trailer that is being used regularly, they will see it many times. But someone living in a neighborhood without may never see one.

I should note even the best AVs have had issues with predicting the path of trailing objects (both Cruise and Waymo have had incidents).
 
Last edited:
  • Like
Reactions: JB47394
I should note even the best AVs have made issues with predicting the path of trailing objects (both Cruise and Waymo have had incidents).
I find it curious that FSD can go around all manner of parked cars, moving cars, delivery trucks, pedestrians, animals, and even blowing trash, but it doesn't seem to have developed a general sense of an obstruction. Perhaps its instinct for moving past the rear of a turning vehicle overrode any notion it had of avoiding obstructions.
 
Fatalities would likely make the press. I haven't seen any reporting in the press or this forum of a fatality. A simple question.
I have not heard of any.

The article says that it was using Autopilot not FSD
It will be good when Tesla just has all vehicles running FSD, then no ambiguity (for newer vehicles).

This does seem imminent - there is probably no way for them to avoid it now. I guess they could stop publishing the mileage data.

Kind of backed themselves into a corner!
 
Last edited:
  • Funny
Reactions: EVNow

Hasn't been updated in a few months, but well researched and sourced.

Thank you very much for sharing that link. As sad as any single fatality is, considering the number of Tesla vehicles on the road and miles driven, I think the outcome is still encouraging. I also note this is AP fatalities data and not FSD and then again not FSD on v12.x. It will be interesting to see how this data trends in future.
 
You probably don't remember, but if the car interpreted you as passing the semi in the next lane, current V11 (highway) behaviour is to slow down when beside the semi. I've seen that commented on multiple times and experienced it as well, and, as usual, this isn't consistent but common behaviour. So it might have been the semi and not delayed reaction to bikes that caused the slow down.
It was a very dramatic slowdown causing me to immediately mash the accelerator. Was more like stopping than slowing. Very obvious freaking out.
 
Well, I've never seen it before. I don't think anecdotes can determine edge cases or not. For example if someone has a neighbor in their neighborhood that has a trailer that is being used regularly, they will see it many times. But someone living in a neighborhood without may never see one.

I should note even the best AVs have had issues with predicting the path of trailing objects (both Cruise and Waymo have had incidents).
I’d like to see a video of how vehicles with trailers turn into driveways where you live.
 
Thank you very much for sharing that link. As sad as any single fatality is, considering the number of Tesla vehicles on the road and miles driven, I think the outcome is still encouraging. I also note this is AP fatalities data and not FSD and then again not FSD on v12.x. It will be interesting to see how this data trends in future.



Not only that, but of the VERY few "confirmed AP on" deaths they list- most....aren't

It lists 2 confirmed AP deaths with this link:
Actual link says cause of crash not determined and does not mention AP at all.

Another is sourced as here:
But again AP isn't mentioned in the story at all, in fact it says police were still determining cause. Further, the death happened AFTER the crash when the Tesla driver, ejected onto the road through the windshield, was run over by another car.

Another is about a 17 year old- speeding, and again NO MENTION OF AP IN USE, hit a pedestrian on a sidewalk.
next one is sourced here:
SHOCKER- no mention of AP in the source.

Next one-
A drunk in a NON tesla went off the road, crashed through a fence, then crashed into a Kia in the opposing lane (and the 2 deaths were in the Kia)... then the drunk in truck ALSO hit a Tesla (no mentioned and nobody killed)

I could keep going but hopefully this gives you an idea what a garbage source that link is.
 
Not only that, but of the VERY few "confirmed AP on" deaths they list- most....aren't

<snip>

I could keep going but hopefully this gives you an idea what a garbage source that link is.

Thanks for doing the legwork on this. Amazing isn't it.

Also of note for me were the number of fatalities involving a Tesla where its most likely not the vehicle that caused the accident. ie Motorcycle hit Tesla etc
 
Stock analyst Toni Sacconaghi was on CNBC this morning (not a TSLA bull). Said he's been driving FSD the last 3 weeks and FSD has made good progress but has "every daily limitations," "still a long ways to go." He thinks:

- full self driving will take 5-10 years,
- it isn't clear that TSLA will be the only company solving it,
- automotive innovation (premium features) are almost always priced away due to global competition and eventually become standard features.

Not saying he isn't going to be right, but why should we care what he thinks?

I guess he got Apple/iphone correct so he must be an expert on this topic as well. ;)

Toni Sacconaghi 2008 -- Others, such as Sanford Bernstein analyst Toni Sacconaghi, say that the discrepancy suggests that more than 600,000 iPhones are stuck in some warehouse gathering dust. The rub? Rabid iPhone demand is pure fantasy.
 
  • Like
Reactions: JB47394
Any reason why you didn't attempt to answer the actual question?
Because I don't know ... not sure anyone outside Tesla knows.

BTW, I tried to figure that out using the mandatory filing of Tesla with NHTSA - but thats for all of AP/EAP/FSD. I've a thread on that somewhere ...

ps : Here it is ...

 
Last edited:
  • Helpful
Reactions: willow_hiller
You were the one who responded that lane splitting was legal in SoCal (but at low speeds.) I was just pointing out that the incident reported was at 80mph.
Lane splitting at 80mph is stupid. That is not traffic flowing at a crawl.

I would not lanesplit if traffic was flowing greater than 20mph. A google search says:
1000029564.jpg
 
Last edited:
I’d like to see a video of how vehicles with trailers turn into driveways where you live.
To clarify, I mean I either don't see them or they don't turn into driveways. The trailers I have seen, the people just park outside next to the sidewalk as the neighborhood I have seen them has plenty of free parking.

I'm just pointing out that the overall instance of trailers at all may be low, much less ones pulling into the driveway. As such, without knowing overall stats as opposed to anecdotal accounts, it's not possible to know if it qualifies as an edge case or not.
 
  • Funny
Reactions: AlanSubie4Life
To clarify, I mean I either don't see them or they don't turn into driveways. The trailers I have seen, the people just park outside next to the sidewalk as the neighborhood I have seen them has plenty of free parking.

I'm just pointing out that the overall instance of trailers at all may be low, much less ones pulling into the driveway. As such, without knowing overall stats as opposed to anecdotal accounts, it's not possible to know if it qualifies as an edge case or not.
It’s also possible you don’t remember seeing it because it’s totally normal. I will keep an eye out. If it’s a true edge case then it’s extraordinarily unlikely I will see it happen again before I die.
Here’s an example of an edge case. Sun shining through construction looked a lot like lane lines causing autosteer to swerve right. I think it’s unlikely I will ever see it again.
 
Stock analyst Toni Sacconaghi was on CNBC this morning (not a TSLA bull). Said he's been driving FSD the last 3 weeks and FSD has made good progress but has "every daily limitations," "still a long ways to go." He thinks:

- full self driving will take 5-10 years,
- it isn't clear that TSLA will be the only company solving it,
- automotive innovation (premium features) are almost always priced away due to global competition and eventually become standard features.



FWIW that's the dude who has had a sell rating on Tesla since mid-2020 when it was $94 a share- a price it's never gotten down to again.

And that rating came not long after he told people the Jaguar iPace was stealing market share from Tesla. A car that sold so terribly that when it's discontinued here shortly it'll have sold in 7 years total on the market less cars that Tesla sells every 2 weeks.


I mean, I don't think we're getting robotaxis this year either- but he's not exactly had the best track record on anything auto-related I'm aware of.