-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Poll: Is FSD a complete crock?
- Thread starter DanCar
- Start date
diplomat33
Average guy who loves autonomous vehicles
I don't think FSD is a crock. But I think it will take longer than 1-2 years before we can sleep in our Teslas while it self-drives.
jebinc
Well-Known Member
"Marketing" at Tesla is pretty good. Names/branding like "Autopilot," "Ludicrous+," "Full Self Driving" (aka, "FSD") , are very misleading. Does anyone know if you can "unbuy" the FSD option - meaning getting Tesla to remove, it via update, and then refunding the $6000 paid? And if so, what the process for doing so is?
Terminator857
Active Member
Would that be the third choice? Half baked, half crock? Or do you prefer 25% crock?
Article says FSD is years away:
Self-Driving Cars Are Still Years Away. That's Probably A Good Thing. | HuffPost
Self-Driving Cars Are Still Years Away. That's Probably A Good Thing. | HuffPost
Was informed of Tesla's definition of feature complete: 25% of daily commutes can be done without driver engagement.
Does that mean the people that voted 75% are right?
Does that mean the people that voted 75% are right?
jebinc
Well-Known Member
If that turns out to be true/accurate, then yes. But no one really knows at this point what real or not.Was informed of Tesla's definition of feature complete: 25% of daily commutes can be done without driver engagement.
Does that mean the people that voted 75% are right?
diplomat33
Average guy who loves autonomous vehicles
jebinc
Well-Known Member
From an engineer at a different driverless car company. I didn't ask how he knew.
diplomat33
Average guy who loves autonomous vehicles
Thanks. In my opinion, I would not consider that person to be a reliable source since it is 2nd hand and comes from a competitor.
I would consider it reliable because it sounds right and there is no incentive to give misinformation. Also the person is reputable. True it is second hand and no date of when the information passed to the next, so Tesla's definition could have changed or someone misinterpreted something.
Oh and I realize I made a mistake. Strike "... driver engagement" and replace with "... driver intervention". Driver must always be engaged. Here it is restated correctly from memory:
25% of daily commutes can be done without driver intervention.
Oh and I realize I made a mistake. Strike "... driver engagement" and replace with "... driver intervention". Driver must always be engaged. Here it is restated correctly from memory:
25% of daily commutes can be done without driver intervention.
Last edited:
Atari2600
Active Member
I assume FSD = Level 5. Well I've seen bad weather, dirty cameras and various road repairs either disable the current system or cause undesirable behavior. So I'm not sure any more. I think HW3 can probably handle the data but I guess I'll have to wait and see how if it does a better job with the camera inputs than 2.5. That said "I want to believe".
diplomat33
Average guy who loves autonomous vehicles
So I interpret that to mean that if you took say 100 daily commutes that people do, 25 of the trips would have no driver interventions at all. Am I interpreting your statement correctly?
If correct, that seems like a very low bar that Tesla is setting themselves.On the other hand, it does kinda jive with Elon predicting "feature complete" by the end of this year. If their criteria is no disengagements for only 25% of commutes that seems more plausible that they might achieve it this year. Also, the last 2 FSD features on the page, traffic light response and "automatic city driving" would seem to fulfill that criteria as well. If the current Nav on AP could also handle traffic lights, stop signs, and intersections reliably, then I could see that covering 25% of people's commutes. So it kinda makes sense with what we know.
Level 5 is the end goal but it will take awhile before FSD actually gets there.
Much longer than you think likely. I wish I could get EAP off my car and get a refund of it's useless features. The old standard AP on my S was better than what I have now. I don't need the car to change lanes if I have to nudge the wheel all the time, that makes it useless and advanced parking lot collision is something I certainly could care less about even if it worked perfectly. Boy was that a $5k money fleecing operation.
Yes, that is how I would interpret it.
Interesting, I have the opposite opinion. I think it is a very high bar. I guess that speaks to our expectations. I have very low expectations and you have high ones.
Changing subjects slightly: Reliably is an interesting word. What I got from speaking to one driverless perception engineer is that the best traffic light recognition is at 99% now. That doesn't seem reliable to me, but if a human is suppose to augment it, then I guess it is passable. Still seems very dangerous to me, because people will tend to tune out and not watch the traffic lights closely.
Daniel in SD
(supervised)
Unless everyone in the industry is lying about their disengagement data there is no way the reliability is 99%. You can't go 10k+ miles between disengagements with 99% traffic light recognition (unless maybe they stop for 1% of green lights and the car eventually figures it out and continues).
I would say doing 99.9% of commutes (on mapped paved roads) 25% of the time would be a good bar for "feature complete". Having a high percentage of commutes that the system can never do is not feature complete IMHO.
Similar threads
- Poll
