Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

FSD Beta Videos (and questions for FSD Beta drivers)

This site may earn commission on affiliate links.
I would almost swear sometimes that FSD Beta can hear and understand compliments, becomes self-conscious, and then screws up. Major mess-ups seem to happen after compliments more often than random chance would dictate.

This is so bad though, I was waiting for a "But the car was trying to be extra safe, so I actually don't mind that"
 
One minute video of worst of FSD from 22 minutes of driving. Kind of click baity in my opinion, so sorry for reposting, but interesting anyways. Is it fun to see FSD failures, while ignoring the many successes?
I honestly think we need more of these videos. There are far too many out there that just gloss over the failures so we only see the "20 miles with zero intervention" kind of BS.
For me, FSD is probably getting close to 70/30 now, where 70% is pretty darned good. Good enough to inspire confidence in its capabilities, however the 30% is just laughably unusable/insane. More specific examples are trying to fit into a 3' wide bike lane, or making a 20mph corner at 40mph, or selecting the ghost island as the travel lane, or coming to a full stop for a flashing yellow warning light on a 65mph divided highway and other such completely insane things that even a student driver wouldn't try to do.
Crazy as it sounds but it still feels like its moving forward and getting better. Maybe 10.10 will be 75/25 brilliant/insane ?
Maybe one day SWMBO will enable FSD 🤣
 
Title: THE BALLSIEST Move I've Ever Seen FSD Beta Pull

Xpz633Y.jpg
 
VERY discouraging and sad to watch and SCREAMS that there is no way we will have Level 3 this year much less L4.
Considering the current disengagement rate is 1 in <10 miles and for human level driving it needs to improve 1000x, why would anyone even think it will happen this year ?

BTW, L3 can be achieved by not taking those left turns. Waymo would never attempt those.
 
Considering the current disengagement rate is 1 in <10 miles and for human level driving it needs to improve 1000x, why would anyone even think it will happen this year ?

BTW, L3 can be achieved by not taking those left turns. Waymo would never attempt those.
““I would be shocked if we do not achieve Full Self-Driving safer than human this year. I would be shocked.””
 
OMG looks like Chuck has had his WORST left turn video ever. He only "died" about 5 or 6 times.:oops: It is a total failure with the sorta exception of the last try. VERY discouraging and sad to watch and SCREAMS that there is no way we will have Level 3 this year much less L4.

I think he should not do the reverse and reset thing. The car thinks you’ve already completed the stop at the stop sign and proceeds.
 
Unfortunate that Chuck had a bad experience there, I watched his Memorial Park test route last night and it seemed solid
I've given up long ago getting worried because of Chuck's penultimate ULT. This intersection is like the final boss in a video game. You can beat it some of the time through repetition, because in video games you have infinite lives. Unfortunately, Tesla drivers are not cats.

There are too many issues at that intersection to use it as a test bed for ULTs. The side street visibility is poor because of obscuring vegetation. The highway has a 55 mph speed limit and there isn't a nearby upstream controlling traffic signal. The median barely fits a vehicle. That last factor is irrelevant because FSD beta presently only stops in the median through its own poor judgement.

But besides all that, it's clear to me without a side-facing camera embedded in each headlight, there just isn't enough lateral visibility with the present camera mix to enter a highway at an unprotected intersection where the vehicles are moving at 25+ m per second.

This extreme intersection type begs for Tesla to implement a crowd-sourced hazard reporting system. Else, FSD will constantly enter situations where it doesn't belong. If a hazard is reported, other drivers can review and comment as to its validity and then Tesla can decide whether to include the hazard, so the navigation will avoid the location. This is creating an anti-waypoint, so to speak. Short of crowd-sourced information, at least let the driver save anti-waypoints so the car will avoid them.
 
I've given up long ago getting worried because of Chuck's penultimate ULT. This intersection is like the final boss in a video game. You can beat it some of the time through repetition, because in video games you have infinite lives. Unfortunately, Tesla drivers are not cats.

There are too many issues at that intersection to use it as a test bed for ULTs. The side street visibility is poor because of obscuring vegetation. The highway has a 55 mph speed limit and there isn't a nearby upstream controlling traffic signal. The median barely fits a vehicle. That last factor is irrelevant because FSD beta presently only stops in the median through its own poor judgement.

But besides all that, it's clear to me without a side-facing camera embedded in each headlight, there just isn't enough lateral visibility with the present camera mix to enter a highway at an unprotected intersection where the vehicles are moving at 25+ m per second.

This extreme intersection type begs for Tesla to implement a crowd-sourced hazard reporting system. Else, FSD will constantly enter situations where it doesn't belong. If a hazard is reported, other drivers can review and comment as to its validity and then Tesla can decide whether to include the hazard, so the navigation will avoid the location. This is creating an anti-waypoint, so to speak. Short of crowd-sourced information, at least let the driver save anti-waypoints so the car will avoid them.
That intersection is indeed very difficult. Where I live that kind of doesn’t exist. It would be either traffic light controlled or speed limit would be lower.
 
The most disappointing part is that Chuck has had much better experiences in that turn on previous builds, so it seems like a regression. It's tough to exactly say if it's a regression or other factors are influencing the performance and previous runs weren't dealing with those factors, but he has previous videos where the car nailed every ULT one after another.

For FSD Beta right now, past performance definitely does not guarantee future results.
 
  • Like
Reactions: n.one.one
I’m out of the beta. Upgrading my M3 to a new one. Got a good trade in price from Telsa on my current one and will pretty much leave it parked until my new car arrives. Not getting fsd on my new M3.

I’ll be looking here for news on how fsd is working. I’ll check it out again if everyone starts saying it really works.

For me.. I don’t need it to drive around stopped cars, but the jerky steering needs to go away. Along with the phantom stopping (at pedestrians crossings with no pedestrians around). And most of the “just push the go pedal” situations need resolved.
 
  • Like
Reactions: Terminator857