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

FSD Beta 10.69

This site may earn commission on affiliate links.
I think Tesla can solve the lane problem easily if they created a cloud server that generated a lane map for a given route, but to them (and Elon), that is a crutch, so they're probably trying to do it the hard way. In the case the lane map isn't downloaded in time, the car will just revert to the current approach.
This right here. Tesla can continue to use the current FSD approach of "generalized" driving intelligence, but if all cars have cameras, why not generate a decent map, record whether a given maneuver worked or not, and next time when someone tries the same maneuver it will have a much higher chance of success. They still use the generalized approach and augment with a recent high resolution map. I understand the "crutch" argument to an extent, and I also understand it's more work, but imo the current vision-only approach has its limits which we are all seeing.
 
That was my point. Nothing is better than FSDb - from a real usability standpoint. For experimentation and fun with friends, it’s fine. I’m more focused on features that are actually usable. FSDb is not currently useable/production ready; it’s an abomination in that regard. @WilliamG

Edit: And sadly, GM SC will likely be more usable than FSD v11 Autopilot for quite some time.
IMO comparing Apple to oranges isn’t useful. Better to compare AP to SuperCruise and UltraCruise to FSD.

MobilEye and Zeekr are supposed to have similar city driving in China next year. Would be interesting to compare.
 
IMO comparing Apple to oranges isn’t useful. Better to compare AP to SuperCruise and UltraCruise to FSD.

MobilEye and Zeekr are supposed to have similar city driving in China next year. Would be interesting to compare.
Fair enough. My only point was FSDb is really not usable "as is". It may be fun, entertaining for the driver, but no one else - passengers, other drivers in your vicinity, etc. At this point, it's still a "Parlor Trick". Note: I left the descriptor, "Cheap" out as that would invalidate the statement! 🤣
 
Fair enough. My only point was FSDb is really not usable "as is". It may be fun, entertaining for the driver, but no one else - passengers, other drivers in your vicinity, etc. At this point, it's still a "Parlor Trick". Note: I left the descriptor, "Cheap" out as that would invalidate the statement! 🤣
To each their own. I use FSDb everyday, and I really miss it when traveling and driving a non-Tesla rental car.
 
Fair enough. My only point was FSDb is really not usable "as is". It may be fun, entertaining for the driver, but no one else - passengers, other drivers in your vicinity, etc. At this point, it's still a "Parlor Trick". Note: I left the descriptor, "Cheap" out as that would invalidate the statement! 🤣
It depends on your expectation.

I was using AP on the roads before FSD. Had to disengage at turns, press the accelerator when there was no lead car at stop signs and lights. More than 90% of that disengagement & intervention is gone with FSD. So, to me it is very useable, and I use it all the time.

But, it is definitely not as reliable("useable") as a crusty old cruise control (though you have to intervene, disengage a lot more), since you know the limitations very well.

I'm used to being at the edge of software for decades. So, feels natural to me ;)
 
That’s not on city roads. So, how is it even comparable?

GM is supposed to come out with UltraCruise next year available on the $200k hand built car. We’ll then be able to compare.
Highly comparable. Every time someone on here mentions using AP on something other than highways there is a chorus of folks saying that AP is only for highways, just like the (still) beta note tells us.
Also Supercruise includes lane change and autopark so is really similar to EAP, not AP which can't change lanes.
Difference is that its already hands free and doesn't have the "beta" tag on it like AP & EAP.
Ultracruise isn't available yet but is very likely to be baked when it is released and isn't likely to have a beta tag either.
Neither Ford or GM are selling their product based on unrealistic promises.
Promises such as -
FSD will drive coast to coast without driver intervention this year, next year, definitely next year, <crickets>
FSD will enable Robotaxis, next year you will be able to enable robotaxi for sure, in the future... <crickets>
FSD is worth 6k, 8k, 12k, 15k based on the value it adds to the car, soon will be 30k. Oh we didn't deliver in the three years you owned the car, oh well give us another $15k you'll get it soon, honest.
Your car already has all the hardware it needs for FSD (to MCU1 car buyers)
 
Highly comparable. Every time someone on here mentions using AP on something other than highways there is a chorus of folks saying that AP is only for highways, just like the (still) beta note tells us.


03A723DE-8311-4378-BE0C-F803EC2DEA7A.jpeg
 
Update:

I broke down and bought the S3XY buttons to get into my glovebox
with fewer than 4 or 3 presses on the display...

It works great - Even works nicely to open the drivers door from inside of the car.

Now 2022.44.25.5 or 10.69.25 has upgraded the display.
The Homelink (Screen) button is no longer 2" x 6" and it
resides nicely on the top row of the display...

So now open the glovebox is a 2 press operation!!!

Blizzard conditions in south Michigan - So not much time to see 10.69.25 in action.

Shawn
 
So I discovered a second driver profile (synced to my Tesla account I guess). I removed the old non-synced one and made sure to update the settings (re-enable no-confirm NoA one more time lol). Now just need to verify if this persists now. If not I guess I’m going back to the offline driver profile again.
Final update on this - the cloud-synced profile kept my NoA and cancel-blinker settings on the next drive. Nice.
 
I haven't posted much recently about FSDb but the Holiday release is downright awful. Today, a trip that had no interventions consistently with 3.1 was a disaster. It missed a right turn altogether despite signalling. It tried to cut off traffic on an ULT. It then forgot to signal to enter a right turn lane and swerved into it high speed before slamming on the brakes to avoid running a red light. Lastly, it swerved from the turn lane to the right lane in order to make a ULT.

And the new UI is cluttered looking. Has anybody minimized the new media card and then realized there is no way to get it back? Is that a feature? I'm afraid that while the cat has been away, the mice have been playing. Elon needs to get back to meddling at Tesla. This is first version that has made me think that Waymo will beat Tesla in the robotaxi race.
 
I played a bit with the new auto-cancel turn signals. They worked as intended, but I only had a couple uses as I ride on FSDb almost all the time where they would be useful.
How does it compare to FSD Beta's turn signals -- at least the turning off part? I would guess Auto Turn Signals might be powered by some FSD Beta learnings, and it might also feed into training FSD Beta as well. Similar to Traffic Light and Stop Sign Control having the driver confirm and override if necessary to send back tricky scenarios of traffic light lane assignments, Tesla could potentially improve turn signal behavior by having the fleet find problems, e.g., the driver manually turned off the signal earlier than it thought it needed to.

From my quick testing today, it seems to understand the difference between a lane change to continue straight (and turn off the signal) vs lane change into a turn lane (and keep the signal on).
 
  • Like
Reactions: FSDtester#1