FSD Beta 10.69

[Charles Guerin]

FSD Beta 10.69.2.2 definitely wasn't slowing down for the yellow and potentially not even for the red light, so I needed to disengage around here as the visualization was showing bright green traffic lights:
View attachment 856442

As you can probably tell from the car's own shadow, the sun was pretty low, and FSD Beta was already getting confused at unlit yellow school lights but brightened by the sun angle. Here's a view from the back at the traffic light, and even then it's hard to see the next car:
View attachment 856443

Definitely a corner case as the sun setting only lines up with this road only for a few days of the year and happens only for a few minutes each of those days. On the flip side, I've noticed that FSD Beta actually does fine even with the sun directly behind a traffic light as there's multiple traffic lights to look for and doesn't mind "staring at the sun."

I had the same thing happen, same conditions.

Most of my drives have been uneventful but I may have discovered another edge case — anyone are try this? :

When pulling out of a parking lot into a 2-4 lane street, position the car at a right angle to the street and turn on FSD.

Twice now the car has gone in a straight line rather than turning into the street, the first time directly into a bus that was passing. 2nd time it was aiming at the opposite curb.

Both times I disengaged quickly. I think it couldn’t tell I was entering a street and it need to turn either right or left.

 

[Supcom]

FSD completely disengaged going around a 90 degree bend and headed straight for the sidewalk where the orange cones are. This has happened twice. Yesterday I figured I must have grabbed the wheel causing the disengagement. 10.69.2.2 doesn't handle this curve as well as prior versions so I pay extra close attention taking the curve now. Today I specifically had both hands an inch off the wheel and FSD simply disengaged and drove straight towards the sidewalk again. Braked and grabbed the wheel to make it thru the turn safely. Another example of a regression in 10.69.2.2

Note street view doesn't do the curve justice. Much sharper than it appears.

Google Maps

I can't tell from your photo, but I assume that the cones were on the outside of the turn so that when FSD disengaged, it stopped actively steering the car and simply continued straight toward the cones. Did you get the Red Wheel of Death on the display with notice and loud chimes to "Take Over Immediately"?

I don't recall FSD ever simply disengaging without getting a Take Over notice. I did have a scenario on a previous version of FSD that would give me a Take Over notice most times it drove down a particular spot on a residential street. Nothing strange looking to me, but FSD just did not like that spot.

 

[aronth5]

I can't tell from your photo, but I assume that the cones were on the outside of the turn so that when FSD disengaged, it stopped actively steering the car and simply continued straight toward the cones. Did you get the Red Wheel of Death on the display with notice and loud chimes to "Take Over Immediately"?

I don't recall FSD ever simply disengaging without getting a Take Over notice. I did have a scenario on a previous version of FSD that would give me a Take Over notice most times it drove down a particular spot on a residential street. Nothing strange looking to me, but FSD just did not like that spot.

Cones are on the sidewalk not the road. No Red Wheel of Death simply FSD disengaged and went straight to the cones. I was going to post this yesterday but said to myself maybe I held the wheel enough to force the disengagement. That is why today I specifically kept my hands off the steering wheel by a good inch waiting to see what would happen. Very bizarre and dangerous. As Tesla says FSD can do the worst thing at the worst time. This is an example. Fortunately I am always extra careful at this curve.

 

[Dewg]

FSD completely disengaged going around a 90 degree bend and headed straight for the sidewalk where the orange cones are. This has happened twice. Yesterday I figured I must have grabbed the wheel causing the disengagement. 10.69.2.2 doesn't handle this curve as well as prior versions so I pay extra close attention taking the curve now. Today I specifically had both hands an inch off the wheel and FSD simply disengaged and drove straight towards the sidewalk again. Braked and grabbed the wheel to make it thru the turn safely. Another example of a regression in 10.69.2.2

Note street view doesn't do the curve justice. Much sharper than it appears.

Google Maps

What speed were you going? If you wheel down the speed before the curve, does the car handle it better?

 

[aronth5]

What speed were you going? If you wheel down the speed before the curve, does the car handle it better?

Speed limit is 30mph. Prior to .69 it would slow down to 22-24mph as it entered the turn. Doesn't slow down as much now. Tomorrow I'm going to try it again since it handled the turn on .69 the first few times, albeit poorly. It is only the last two drives that it freaked out totally. Other turns/intersections are better with .69. Jekyll and Hyde release for me.

 

[Knightshade]

I've never had FSD just give up without any warning.... (I HAVE had ADAS systems on Toyota/Lexus, Hyundai/Kia vehicles all do that though)

 

[powertoold]

FSD, whether on highway or city, will give up when the perceived environment is distorted beyond bounds set by the AP team. I've had highway NoA give up as it was driving into a curved widened lane that was under construction and painted with makeshift lines.

FSD in city will also give up when the perceived road geometry or planned path changes between too many different variations.

 

[D Good]

There are an infinite number of edge cases with an infinite number of solutions.

Guys. Is it possible???

 

[KArnold]

FSDb even if in rare occasion "silently" disables, that sounds like a big deal. Guess maybe it's more of a question on "how rare an event like this happens". Would love to see a video - maybe someone can ride along and video the screen?

 

[diplomat33]

There are an infinite number of edge cases with an infinite number of solutions.

Guys. Is it possible???

Yes. Remember AVs don't need to be able to handle every single edge case imaginable. They just need to handle the common cases and enough edge cases that they can be safe enough. Statistically, you need 2B miles of driving to prove if your AV is safe enough. That's because 2B miles will cover all the common cases and enough edge cases to prove if your AV is safer than humans. If your fleet can drive 2B miles safer than humans, then statistically you know your AV is good enough since you know it can handle enough edge cases.

 

[aronth5]

I've never had FSD just give up without any warning.... (I HAVE had ADAS systems on Toyota/Lexus, Hyundai/Kia vehicles all do that though)

I hadn't either until yesterday and again today. That is after 11 months using FSD on almost every drive. I have taken this curve hundreds of times before without anything this drastic happening. I specifically avoided posting anything yesterday assuming I did something to force FSD to disengage.

Hopefully Tesla will actually review my email.

 

[GregalseT]

I've never had FSD just give up without any warning.... (I HAVE had ADAS systems on Toyota/Lexus, Hyundai/Kia vehicles all do that though)

It's done this many times to me. Just nopes out - here, you take it. I feel this is very dangerous and IMO is programmed in so Tesla can say FSD/AP was not involved in the ensuing crash.

 

[r1200gs4ok]

I got the new fsd before the .2.2 revision.....the first drive I took was wonderful.....no issues at all except it wont go into the right turn lane (dotted lines) instead, it goes to intersection the then turns right....late on all turn signals.....it was smoother and did not have the jack rabbit starting from stopped position....

then the next release a week later came out......what a pitiful version for me.....herky jerky in turns, out of no where braking, made left hand turn and car started out after red light stop, and just stopped.....car behind me had to slam on brakes.....its awful......hope I am in the minority for this version and it gets fixed soon.....expecting new version within next two weeks

 

[thewishmaster]

Does 10.69.x throw itself at cars at 4-way stops when the other car is making a left turn from your right-hand side for anyone else? So far I’m observing a 100% rate of FSDb starting to accelerate just after the other car starts to move. Ironically it hasn’t done this a single time when the car from the right (or left) is going straight - waits until it clears the intersection fully. Happened again today.

 

[JulienW]

Great video that is not (Chuck) Left Turn or FSD specific (actual about strode medians) but highlights and covers a lot of the technical problems involved. This shows how complex and nuance LHT are.

 

[nvx1977]

It's done this many times to me. Just nopes out - here, you take it. I feel this is very dangerous and IMO is programmed in so Tesla can say FSD/AP was not involved in the ensuing crash.

It's not dangerous if you're already paying attention like you should be. Surprising, maybe. Dangerous, no. And Tesla has made it clear that the instantaneous state of AP is not what is looked at in a collision. I forget the exact interval (maybe someone can confirm), but if AP was active within 15? seconds beforehand, then Tesla considers that AP was active for the collision. Whatever the interval was, it was generous. Let's stop with the conspiracy theories.

 

[SFLDMD]

I just hit over 100 miles on autopilot in the last 30 days, everything else should be good to make me eligible for the update. Is the update continually rolling out? Or was it just a one time release and I will have to wait for the next roll out?

 

[DCGOO]

I just hit over 100 miles on autopilot in the last 30 days, everything else should be good to make me eligible for the update. Is the update continually rolling out? Or was it just a one time release and I will have to wait for the next roll out?

No one here can say definitively where any individual vehicle is in the queue. I was in line with a 100 Safety Score for nearly three months (kicked back to a 99 for the last couple of weeks), plus I had the feature on my 2018 for nearly a year before trading it in. my Model 3 has had FSDb throughout, but it got 10.69.2.2, just a couple of days before my X. I had several hundred autosteer miles, although not too many in recent weeks.

Bottom line, no one really knows for sure. You might get it later today, next month or next quarter. Keep your Safety Score up and don’t give up.

 

[Mardak]

Is the update continually rolling out?

It seems like Tesla might be doing more regular additions with 10.69.2.2 with potentially another batch on Friday after the initial new additions on Monday. If it's like 10.2's continuous additions, it could be daily but seemed to be a manual process with somebody running a query for eligible vehicles then deploying the update, so the exact timing can vary, e.g., 1-2 days after somebody checks scores before update or nothing on weekends, and could stop happening at any time.

To be clear, the current requirements seem to be:

• FSD capability (purchased or subscription)
• Sufficient hardware (FSD computer, RCCB cameras, MCU2+)
• Safety Score 80+
• Autosteer miles 100+
• Software "earlier" than current FSD Beta, e.g., 10.69.2.2 is 2022.20.17 avoids "downgrade" from 2022.24/28.x

 

[KLam]

I just hit over 100 miles on autopilot in the last 30 days, everything else should be good to make me eligible for the update. Is the update continually rolling out? Or was it just a one time release and I will have to wait for the next roll out?

You may want to try the Advanced button on the Software Update if it was not already on.