I'm on this thread just to see if anyone has heard when 10.13 might drop.
-
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 10.13
- Thread starter Todd Burch
- Start date
AlanSubie4Life
Efficiency Obsessed Member
The question is whether they ultimately call failures successes (for example, not going when there is a valid opportunity is a failure, each time it occurs). It’s potentially a key issue which may yet prove me closer to the mark! I sure hope not though. 100% on this turn would be a “quantum leap” in FSD development due to the new capabilities it would unlock, to be capable of competently and consistently executing such a move with the boldness of a competent human.
tmoz
S85D, Y LR, ex Prius PiP
The delay of nailing the UPL might have given the other guys working on problem areas more time to improve their piece of FSD, thus Musk's comments that this is not really a one step upgrade from 10.12. We can only hope and wait and see.
Ah. Just what I thought. It’s hard to get this right but they are holding back release until they do. It doesn’t seem that any aspect of FSD is 100%. Everything has edge cases that the car can’t handle competently.
If they’re so infatuated with getting UPL turns perfect with 10.13 that’s cool, but why can’t we get some 2022.16 updates while they figure it out? I don’t understand with they have to be released together.
Hoping roundabouts get some love; it had been mentioned as a priority for 10.13 as well.
AlanSubie4Life
Efficiency Obsessed Member
I paid closer attention to this last night, and in at least a couple cases, the car succeeds in both being jarring AND slow at accelerating from red lights. Slight kick on the start, followed by ponderously slow acceleration. Another thing to improve in 10.13, though with potential massive capability improvements in 10.13 (required for near 100% success rate on Chuck's turn), they'll need to dial up the jerkiness to maintain safety levels.
everydaychris
Active Member
Based on a comment from Musk in this week's Tesla shareholder meeting, another week or more is needed. So, maybe next weekend at the earliest???
tmoz
S85D, Y LR, ex Prius PiP
Substitute "month" for "week" and there's your guide.
As has been said before, speculating on the release date will neither make it arrive faster or slower. Though the more you think about it the longer it will APPEAR to take.
Pretty much the only reason I check on upcoming releases is to maybe ratchet back my beta feedback, since I suspect Tesla pay less attention as they get a new release finalized.
Pretty much the only reason I check on upcoming releases is to maybe ratchet back my beta feedback, since I suspect Tesla pay less attention as they get a new release finalized.
Also a new screenshot of a long left turn with the update.
TravelFree
Active Member
The problem with FSD beta is that 90% is nowhere good enough. True FSD needs to be 100% perfect if it is a pathway toward Robo-Taxi. If there is one serious accident in 100 million miles of FSD, the regulators will seek to outlaw it's use in Level 3 + so Tesla needs to achieve the perfection of a 4 function calculator to be considered ready for wide spread FSD and Robotaxi. This will be near impossible as an FSD Tesla will need to predict the erratic actions of human drivers, especially those humans who intentionally break the law, speeding, DUI, running red lights and stop signs erratic lane changing and road rage. Hopefully, FSD will one day predict illegal driving and using it's AI avoid them with better than good human driver action in advance.
I continue to have great success with FSD beta but I always shut it down when approaching a condition I know FSD-beta fails often. 10% failure is too often in my book.
I continue to have great success with FSD beta but I always shut it down when approaching a condition I know FSD-beta fails often. 10% failure is too often in my book.
Daniel in SD
(supervised)
Waymo and Cruise vehicles crash all the time and aren’t outlawed. Cruise just had a serious collision a few months ago (though it is being investigated, they haven’t been shut down).
Obviously it does need to be very close to 100% on any given maneuver to exceed human reliability.
yea, and I would like to know the E-definition of "most". Is that 50.1% of the time or 99.9% of the time?
50.1 of the time, of course…
I don’t think there is a lot of evidence to support this assertion.
diplomat33
Average guy who loves autonomous vehicles
You are correct that 90% is not good enough. True FSD does need to be many times better than human drivers. It should be very close to 100% but it does not need to be 100% perfect. Nothing is 100% perfect.
And yes, true FSD does not need to predict erratic behavior of human drivers as best as possible. That is a big challenge. That is why AV companies like Waymo are focusing so much of their machine learning research into behavior prediction. It is important for AVs to anticipate when human drivers run a red light or a stop sign or do something erratic that might cause an accident, as best as possible. But prediction will never be 100% perfect. We want AVs to avoid accidents as much as possible but there will always be some accidents that are unavoidable.
Mobileye's philosophy is that AVs should never cause an accident and should avoid accidents whenever possible but should not be expected to avoid 100% of accidents. I think that is a reasonable approach to AV safety.
I also question your stat that 1 accident in 100M miles would cause regulators to shut FSD down. Not sure where you are getting that stat from. A quick Google search finds that humans get into an accident about 1 per 533,333.33 miles. So 1 accident in 100M miles would be a whopping 200x safer than humans. If FSD only got into an accident every 100M miles, we would immediately deploy them everywhere now because that would be amazing safety.
Similar threads
