-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
FSD discussion
- Thread starter Buckminster
- Start date
Knightshade
Well-Known Member
...what?
GPS is accurate to within about 16 feet by default, and years ago Tesla patented a method to geolocate significantly more accurate than that-
Tesla patents technology for more accurate GPS positioning
With the advent of self-driving vehicles, GPS accuracy is becoming increasingly important and Tesla believes that it developed a technology that...
electrek.co
So geofencing to a whole state, which is all you'd need since legality is currently state-wide for states that allow robotaxis today is absolutely trivial to do, correctly and accurately
Further- since you would have to input a destination at the start of the trip the car ALSO knows if it's leaving the state because the state is part of the address of the destination.
So none of your objections are.... actual things.
Robo charging for Tesla's dedicated fleet. (Too soon, lol?)
For my area in Chandler Az:
+ Making better lane choices in advance.
+ Smoother turns and setup.
+ Edges over not just for semis but anything encroaching ego lane.
- More ghost braking than normal, mid-day even.
- Random turn signals that come and go. (New)
Ongoing:
- Missing my 2nd stop sign after immediate turn (obscured and shows late).
- Consistently freaks out at Az Ave and Pecos Rd (Warning to take over, but nothing unusual.)
- Thinks a 4 way stop is only 2 way and creeps for no reason.
Looking forward to 11.4
willow_hiller
Well-Known Member
V11.4.1 rolling out to a small group of customers now:
Knightshade
Well-Known Member
The same James who told us on Dave Lees show it made total sense Tesla didn't need LIDAR because vision can do anything lidar can-- but that Tesla using radar made perfect sense because radar is useful in ways vision is not...... and then a couple months later when Tesla removed radar went on the same show to tell us how removing radar will make the system better?
Yes. All of those statements are true. Vision improved to the point that using the radar made the system worse. Tesla are not stupid, when they make decision it's because they believe it will make the system better. Otherwise they would not do the decisions. If you believe that they do, then don't invest in TSLA.
So, sorry to change the subject, but in my 12/16 S P100D the original owner paid for FSD outright so it came with the car when I bought it BUT when I requested the BETA it did not work. Furthermore, my latest software update is 2022/8/11/13. Here's my question; shouldn't this car be able to use the FSD BETA since FSD was sold with the car? I took it to Tesla service and they are saying that the car can not handle FSD!
EVNow
Well-Known Member
He thinks there's 50% chance of Tesla taking liability for when FSD is engaged next year.
Some people tend to extrapolate too much ... If FSD can do a few drives without needing safety related intervention they are happy.
To take liability Tesla needs to be able to do some 10,000 trips without safety errors in all kinds of places, weather and traffic!
Can you do 10000 trips without any safety errors? I know I cannot... Seems like a pretty high bar.. Imo as long as it is safer than most humans it should be okay.
navguy12
Active Member
Would the travelling public in 2023 use a commercial airliner without two humans at the controls?
Safe, software controlled modern commercial airliners still have two error prone humans sitting at the front (and unfortunately, in this pre-AI era, if the aircraft software suddenly doesn’t behave as expected, those two humans most often simply fly the craft to the scene of the crash).
Until the risk (perceived or real) of a fatal (or non fatal but severe) crash is at least on par with today’s commercial aviation world, there is no way Tesla will ever assume legal liability for FSD.
Knightshade
Well-Known Member
I mean, vision objectively makes it worse too--- see the still-lower speed and follow distance limits that have never reached parity with vision+radar over 2 years after they promised it would.
Not to mention they appear to be adding radar BACK to cars (albeit weirdly slowly just like the HW4 rollout)
No, but sometimes they're wrong- something some seem very reluctant to admit. Especially around "what hardware do we need to do something nobody has ever done"
Remember when HW2.0 was sufficient, according to Tesla, for autonomous driving? I do. But it's not.
Remember when HW2.5 was sufficient, according to Tesla, for autonomous driving? I do. But it's not.
Remember when HW3.0 was sufficient, according to Tesla, for autonomous driving? I do. But it's not.
Is HW4 the answer? We don't know yet. Neither does Tesla since they're still not there yet. NOBODY knows how much HW is 'enough' until they achieve it.
FSD Beta is not a paid feature, and nobody is "entitled" to it from owning FSD.
You should have access to all the features included in the actual paid FSD package (which is- everything included in EAP, plus stop light and stop sign recognition). And eventually it'll include city streets once that's moved to the mainline FSD branch instead of the limited beta.
THAT said- your car can't do the beta because it needs updated cameras. Which you are entitled to get, for free... (you might also need a new HW3 computer if you still have the old one- wasn't clear from your post- which also would be free).
Once you get those you'd be able to get the beta, HW wise anyway.
Tesla offering Autopilot Camera Upgrades to owners with FSD capability
Tesla is offering Autopilot camera upgrades to FSD customers with high Safety Scores. The upgrades come just before FSD's wide release.
www.teslarati.com
Link about the upgrades for cameras and how to request them.
Last edited:
EVNow
Well-Known Member
I think a lot of us have done more than 10k trips without errors that result in an accident (i.e. safety errors). Thats like driving for 10 years without accidents (talking about injury / others' property damage).
Moreover - to assume liability - it should make financial sense. An accident caused by automation will result in much higher payouts than if an individual causes accidents. Those payouts need to be balanced with any extra revenue. You do the math. If you search my posts - I've done some estimates before.
And we are not even talking brand damage. Uber automation caused one (fatal) accident - and they finally had to closed down their automation program.
HW3 seems to be performing fairly well on V11 software, V12 is the final version of FSD.
We will only know if HW4 is needed when V12 is mature and stable.
Knightshade
Well-Known Member
Allow me to invite you over here:
AI, Autopilot, & Autonomous/FSD
Discussion about AI, Tesla Bot, Tesla Autopilot (AP), the promise of Full Self Driving (FSD), as well as other Autonomous Vehicles.
teslamotorsclub.com
Where the mods seem to want all FSD discussion to go anyway, and where you'll see most of the people actually using HW3 FSD day to day do not agree with you. (assuming you mean needed for robotaxi--- I think you'd have a lot easier case arguing it's not needed for staying L2 but wide release worldwide)
Performing well? In what universe. I’m starting to think that FSD software is very different dependant on the region. We are definitely not seeing any progress towards a level 4 autonomy system here.
In my experience, the switch to single-stack was a significant positive change, and FSD has improved since then. However, I would disagree with the characterization "that the rate of progress since then has accelerated materially". In my >100 miles of driving today, there were far too many errors that remain uncorrected vs the fewer areas I can say have definitely been improved since single-stack to believe the rate of progress has accelerated. Unquestionably YES, FSD *is* improving every release, just not necessarily at a materially increasing rate. It is still incredible what FSD can do - it's so many orders of magnitude better than any other commercial option available to me, with an ever-growing lead. But as an investor, my own expectation for any 'mass FSD adoption rate' which might impact Tesla's financials materially isn't until much later this decade (depending on how one defines 'mass FSD adoption rate', of course).
Many of today's failures were the 'changing out of the correct lane into the incorrect lane' type - for example, surface street with 2 lanes in each direction and a shared center turn lane, ego is in the left southbound lane with 0.4 miles before the upcoming left turn at a light (where the shared center turn lane becomes a dedicated left turn lane). The right blinker kicks on and the screen shows "Changing lanes to follow route" (not changing lanes to try to speed around traffic or anything)...it thinks it needs to change lanes *right* rather than staying in the left lane and being able to get into the dedicated left turn lane that is about to appear. It did this multiple times today, but the time with 0.4 miles to go before the upcoming left-at-a-light was the most egregious. Also, note that the setting for "Minimal lane changes this trip" does not reduce these errors, as this isn't a lane change to a faster lane, etc...it thinks it *needs* to change lanes to follow the route, even when in these instances if I allow it to change into the wrong lane, it frequently cannot get back left in time to make the turn, resulting in either a recalculation for a new route or a risky dive towards the left turn lane in a manner which leads me to take control again.
Another particularly poorly-handled incident today was heading eastbound on E Lincoln Dr approaching N 24th St, with the navigation planning a right hand turn onto N 24th St southbound. The dedicated right turn lane breaks off before the intersection, and the in-car map shows this; however, ego remains eastbound on E Lincoln Dr until after the gore-zone separator appears, then tries to cross the gore-zone separator until I yanked control back over.
Google Maps
Find local businesses, view maps and get driving directions in Google Maps.www.google.com
Yes, it is true that I haven't received any additional updates since 11.3.6 (v2022.45.15) - all 3 of my cars are on that still to this day! Nice 'strawman' of picking the very latest version as the baseline!
In all seriousness, if the original poster was referring merely to a hypothetical / potential rate of progress increase based on release notes and watching a video online, then my prior post and this can be deleted. My assumption was that their claim was made based on an actual observed rate of progress increase from the first publicly-released V11 builds, and I would refute that based on my experiences. I would also caution that point-releases have been known to be hyped more than the actual build justifies each time, and it is unfortunate IMHO when a release with very good updates is overshadowed by the even-greater-than-possible expectations some set for the release prior to each of us being able to experience it and put it thru a varied range of paces.
@StarFoxisDown! @Singuy
Took V11.4.1 out for a spin this evening in the Model 3 for some first-hand experience with it:
Good - What it does, it does better, and it's learned a new trick:
- Overall driving thru neighborhoods is improved, with smoother turns, better placement within the road when there is no center strip and with various cars parked on the street at places, and very acceptable slowdowns for the large speed humps.
- After making a right onto [removed for privacy reasons] ego now gets into a go-forward lane instead of into the right-hand-turn-into-the-gas-station-and-bank not-a-lane that it used to turn into.
- After making a right onto a many-lane road with the next turn a left in about 0.25 miles, it sometimes turns right into the correct lane and begins to merge left, and sometimes turns right and turns right aiming directly into the 2nd or 3rd lane to get in place for the upcoming left. Very 'human' way to turn / merge when multiple lane changes / crossings are needed, instead of the 'textbook-correct' way to turn / merge.
- Ego handled navigating around the cardboard box that was in its freeway offramp exit lane quite well.
- Northbound on a street with 2 lanes in each direction with a shared center turn lane that becomes a dedicated left turn lane at the light, ~45 mph with a green light, no other traffic, and needing to make a left turn, on 2 of 2 attempts ego goes into the left hand turn lane *after* the opening between the shared center turn lane and the dedicated left-at-the-light turn lane. Arguably, ego never really got into the dedicated left hand turn lane, somewhat going diagonally thru it as it began to make the left turn thru the intersection. Interestingly, going southbound on this same road, needing to make a left turn into a neighborhood street, ego gets into the shared center turn lane at exactly the right point.
- Eastbound on a street with 2 lanes in each direction with a shared center turn lane, where at the light the right lane becomes a dedicated right turn lane and needing to make a right. Ego properly gets into the right lane about a half mile away from the light, but then as the right lane becomes a right turn lane and breaks away from the continue-forward lane, it failed the first time, at the last second attempting to dive towards the left lane until I took control. Second trial run thru the same intersection with the same destination, it wavered back and forth on whether to dive left but ultimately stayed in the right lane.
- Still multiple issues when there are two turn lanes (either 2 left or 2 right, both have issues). Ex: In the right of two left turn lanes onto a freeway onramp, and the nav shows the left turn as next but the blue line shows doing a mid-intersection lane change to the right and going straight. Car hesitated mid-turn but finally made it. Ex: Merging all the way to the rightmost of two right turn lanes just to make the turn, even though the nav shows that after turning right, it'll need to be in the left lane to turn left in 0.25 miles, which traffic there usually doesn't allow. Ex: Offramp which splits into 2 lanes, then has 2 more lanes from the frontage road merge in, then has a left turn lane appear, then has a right turn lane appear and one of the go-straight lanes become a turn lane, ego tends to end up pointed just going straight.
TLDR: If I wanted to make FSDb look great, I know demo routes to use it will do great on. If I wanted to make FSDb look not-so-great, I know demo routes to use it will do not-so-great on. I assume this is also true for those actively promoting their own videos online, so please consider their potential motives before putting too much weight on what you see from online videos, good or bad. Any serious Tesla fan or serious TSLA investor should buy a Tesla, pay for FSD, and experience it for yourself, just as you should be experiencing the product(s) of any company you are a fan of or investor in.
Last edited by a moderator:
If you watch wholemars video closely, his car isn't magical or he picks demo routes purposely to show its good side. The guy just never intervene unless he finds the car stuck or about to hit something. It makes mistakes compared to human driving all the time and do miss turns. He just let it reroute and the car usually recovers. So people claiming that the car require less and less safety related disengagement is just saying it no longer want to smash into things like how unprotected turns on early v10 was really a crap shoot.
You are looking for perfect lane selection and using that as your measurement for improvements. That's just one thing out of a bunch of things....
Last edited by a moderator:
jeewee3000
Active Member
Agreed. As a TSLA investor I bought a model Y, paid for FSD, but since I'm based in Europe I have no clue what FSDb is capable of.
If any Tesla employees are reading this thread, fix this please
Similar threads
