-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
FSD Beta 10.69
- Thread starter Buckminster
- Start date
mtndrew1
Active Member
I am truly more excited about 10.69.1 being on the 2022.20 branch than the actual FSD update.
More like the place that correlates to "fire"
AlanSubie4Life
Efficiency Obsessed Member
Creeping way too far into intersections (1:15, 11:14, 13:30), stopping in the middle of the road (2:36), jerky steering, bad lane selection (2:36, 14:35), spurious signals (5:18), fastidiously staying in the slowest possible travel lane (~4:00-5:18), phantom braking (7:08), ponderous or failed left turns (7:35, 16:14, 17:20), what’s not to love?
THIS is the glorious FSD future we have been promised.
I mean, let’s be realistic here. Looks like lots of improved features, with tons of work to do still to get driver approval (and then subsequent wife signoff, a much higher bar). But it is not correct to describe this as “good.”
FSD will keep getting better and better but it is not going to be useful until Tesla “turns off” these “features.”
Look at this creep limit (?)! It is perfectly placed. Revolutionary? I have no idea why that is being drawn or what it is, it is so advanced. (Also, why is the the wheel turned? And more to the point, why is it a yoke?)
Let’s be realistic about progress and quality. FSD is coming along, but we need to stay grounded in the reality of what currently exists.
We’ll get there, but not by proclaiming that every release is good.
Last edited:
Above is an example of viewing the video objectively in detail, without Elon colored lenses in your glasses on, and making SURE to spot the failures that the video maker intentionally tries to hide/ignore/minimize.
The whole story/full story. Not the shill version.
AlanSubie4Life
Efficiency Obsessed Member
I would say the video maker was mostly quite forthright about the obvious failures during all (or at least nearly all) of the cited events. I would say a good beta tester providing good feedback.
Daniel in SD
(supervised)
One thing I've noticed about your posts is that you never talk about all the collisions that FSD beta avoids. A human driver could have crashed at least a dozen times during that drive.
AlanSubie4Life
Efficiency Obsessed Member
True. I also have to think about the accidents that it failed to create.
I’ll give this topic some thought before posting in future. This should please everyone watching this thread, I imagine.
Daniel in SD
(supervised)
I think if you grade each individual action the car takes you will find that it achieves greater than 90% success which is an A.
Since the car is doing the driving this means 90%+ of human errors are eliminated, a phenomenal improvement in safety.
jabloomf1230
Minister of Silly Walks
I'll have to try it again, but the first LT was from a State highway to a rural unlined road. The more that I think about this, it makes sense to always have FSD beta phone home.
agreed upon further review, the video shows that FSD is fire
It's drawing both the creep limit / wall as well as the median crossover region. The former is probably fine while the latter is quite wrong as the "median" is just the painted yellow lines that actually curve with the road indicating the region is currently expected to be something like a quadrilateral. Hopefully if it did believe it was a median crossover region, it would have figured out that there isn't enough space to safely wait there.
AlanSubie4Life
Efficiency Obsessed Member
Yes, I thought it might be that, though it obviously made no sense. It was also very brief - but we know that the car will instantly respond to brief changes in the perceived environment (and planning).
It's drawing both the creep limit / wall as well as the median crossover region. The former is probably fine while the latter is quite wrong as the "median" is just the painted yellow lines that actually curve with the road indicating the region is currently expected to be something like a quadrilateral. Hopefully if it did believe it was a median crossover region, it would have figured out that there isn't enough space to safely wait there.
but Elon said..I wonder if Elon is tired of hearing negative comments about known issues in early builds (James Locke) and told the release team to not include them in the initial releases going forward...
"constantly seek criticism"
mtndrew1
Active Member
I believe it’s three fire. You are missing two of the fire.
EVNow
Well-Known Member
We all know you are the glass 10% empty kind of guy
The drive looked ok compared to previous version. Thats all I'm concerned about.
It is beyond stupid to think suddenly FSD will jump from 1 disengagement in 10 miles to 1 in 1,000
ChaoticNeutral
Member
Didn’t watch the video because I prefer the hands on experience in my area, but did they actually figure out round-abouts like Elon said or was that not included in the video?
EVNow
Well-Known Member
I didn't reply to this earlier ... we can't really talk about a release being "good" in absolute terms.
To me release is "good" if it improves upon the earlier release in some meaningful ways without causing too many new issues.
Specifically, we'll know its good if disengagement rate decreases. Unfortunately - Tesla doesn't publish that. We'll have to just be content with crowd-sourcing. I'll be doing my usual recording and collect stats over the next month to come to a conclusion (for me).
AlanSubie4Life
Efficiency Obsessed Member
Yes I agree that is not a good idea! I thought about mentioning that!
I think it is best to focus on specific examples of capabilities or demonstrated lack thereof, without assigning broad non-specific statements about “good” or “bad.”
The capabilities will largely determine the overall disengagement rate, at the current time. They will also determine usability.
Similar threads
- Article
