-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
FSD rewrite will go out on Oct 20 to limited beta
- Thread starter diplomat33
- Start date
powertoold
Active Member
That 270 degree roundabout by Tesla Owners SV was as hairpiny as turns get. I guess someone with a big battery pack needs to try a hairpin bend on a mountain.
DanCar
Active Member
Vaporware - Wikipedia
When it reaches level 3 or 4 than we can say "full" self driving actually exists. Until then I think we can still call it vaporware.wikipedia said:
It can be level 3 or level 4 tomorrow if Tesla decides to make it such - the only difference is whether it nags for the steering wheel and who pays if there's an accident.
Of course, if Tesla or anyone else decides to make it Level 3/4 too early, the increase in accidents will have serious impacts on the public perception of autonomous cars, and might lead to legislation against it. Not to mention the cost of the accidents themselves.
So clearly Tesla shouldn't want to make the switch until they're confident it'll get into fewer accidents than humans do, which is why they're doing this level 2 FSD period gaining confidence and maturity in the system before tell the drivers they don't need to pay attention.
powertoold
Active Member
No doubt about that, especially for single lane keeping in traffic jams on a highway. All the Tesla vision naysayers are slowly dwindling as more and more FSD beta videos come out. The beta isn't level 5, but it gives us the taste of possibility, considering they haven't even deployed dojo yet.
Semantics, fine.
In that same definition depending on which point you pick, other companies could be defined having vaporware as well. Not only in terms of SAE level expected but also in terms of availability to the general public.
It's more about managing expectations. Elon has not been great about it, has been overly optimistic. Other companies have been years late in different ways.
Last edited:
DanCar
Active Member
Wonder how long before there is pothole detection and avoidance. Many places in the country will need this before Level 3 is practical. Or maybe it is fine to hit a pothole during traffic jam assist.
powertoold
Active Member
Elon said they're labeling potholes, so it'll come eventually, but I'd rather Tesla focus on fixing the dangerous situations and polishing every turning scenario first.
Tesla is in a prime position to accelerate FSD development. It seems they've got their software foundation "right" this time. Their data engine is producing results. They've got a good human driver and NN perception feedback loop. We know this because Karpathy said this FSD beta build was the result of many months (not years) of hard work. It was a subtle flex IMO:
Falkirk
Member
I think the system needs to handle speed bumps first! In the video I saw the car didn't even slow down for it.
Mo City
Active Member
The video from the guy in Miami showed the car slowing for speed bumps.
If America didn’t have such a ‘rugged individualism’ psyche, more money would be spent on fixing roads
LA commute test from Santa Clarita to pasadena and back...longer videos
Tesla Full Self-Driving Commute Test: Santa Clarita to Pasadena and Back [VIDEOS] - TeslaNorth.com
Tesla Full Self-Driving Commute Test: Santa Clarita to Pasadena and Back [VIDEOS] - TeslaNorth.com
Daniel in SD
(supervised)
I really don't think this is true. The capability Tesla has demonstrated so far was achieved a decade ago. The fact that they are achieving it at much lower cost really isn't evidence that they're close to achieving robotaxi capability. Relative to where they were it's definitely a quantum leap. I think most of the skeptics, like myself, have observed that many companies have quickly gotten to the point where the car can do a good approximation of driving for 100 or 1000 or 10,000 miles only to stall out. The fact that every other company has also made overly optimistic predictions should make you less optimistic about Tesla, not more. The entire history of artificial intelligence is full of people making overly optimistic predictions. Maybe this time will be different.
I really like this guy’s vids. He and his wife are calm and he’s polite and takes over when the car might impede other drivers. It looks like he uses AP pretty similarly to how I do, actually. Impressive stuff from the vids.
yup, james taking over to not slow down traffic was a very nice gesture. i wish the camera was showing more of the wheel though
Really? Any sources to demonstrate that, ie real world demos of cars doing what Tesla FSD beta is doing from 2010. Would love to see it.
No, not that I am aware of. But keep in mind that driverless rides in Chandler are in a 50 sq mi area so it is unlikely we would see an hour long commute on freeways. Also, we have less videos from Waymo because Waymo riders are just average people using Waymo to run errands. They are not tech evangelists who spend all day putting their drives on youtube like Tesla fans do.
We do have several complete rides and the longest video we have is the 1hr30mn one that is 4 rides stitches together. These videos do show us that Waymo's FSD can handle city driving very well as they are completely driverless with no problems at all and complete smooth.
Waymo has a disengagement rate in CA last year of 1 per 11,000 miles. So I think a hour long commute video from Waymo would probably be very boring.
My report said that freeways were part of the ODD but that Waymo recently changed it. Waymo will probably put freeways back into the ODD again.
Similar threads
- Question
