FSD Beta 10.69

[sleepydoc]

Pretty painful to watch video but this gives a more realistic picture of Beta driving in the city. He lives in the burbs and even states how much better it drives there.

I haven’t watched the video but I can imagine - in places like downtown Chicago you need to be aggressive, move as soon as you have the chance and plan ahead. All of these are significant weaknesses of FSDb.

 

[old pilot]

Are you saying that a 70 year old dude is better off letting Jesus take the wheel, and hopefully figuring out whatever wonky nonsense FSD is doing or is about to do and correct it, than use his 50 years of driving experience to just drive manually?

Debatable.

74 year old here, I don't have a problem with correcting FSD and it is helpful on long drives. I know as I get older I will lose more of my driving abilities so I am participating in this experiment with the hope of some day realizing the advertised potential of full self driving so I can continue to be active and safe.

 

[AlanSubie4Life]

Some classic 10.69.3.1 Beta performances. These were one after another, all within five minutes.

Obviously normally I would just not even bother, or take over in advance (I could see all of these situations coming well in advance).

Keep an eye on that bar! Epic and thrilling!

Curving off ramps, impossible:

Google Maps

Find local businesses, view maps and get driving directions in Google Maps.

maps.app.goo.gl

Unable to stop remotely correctly on downhill grades:

Unable to signal intentions for some reason (is there a plan to fix this?). This resulted in adding 5 minutes to my commute, so convenient:

 

[kabin]

This drive started off poor with needlessly close slow drive next to a parked motorcycle. And then apparently a bad situation was deleted from the video but you can tell it was heading for the tree.

Indecisiveness might be as bad as ever. The driver does his FSDb influencer sell job redirecting cause to distant oncoming cars and FSDb being "extra cautious." Seconds later he states FSDb was doing fine so far in the 3min drive.

I gave up after 3 mins of watching that.

 

[AlanSubie4Life]

This drive started off poor with needlessly close slow drive next to a parked motorcycle. And then apparently a bad situation was deleted from the video but you can tell it was heading for the tree.

Indecisiveness might be as bad as ever. The driver does his FSDb influencer sell job redirecting cause to distant oncoming cars and FSDb being "extra cautious." Seconds later he states FSDb was doing fine so far in the 3min drive.

I gave up after 3 mins of watching that.

Yeah, I would say that for the market and for everyone else here, the bloom is really off of this rose.

We’ll get single stack that doesn’t have to be turned off sometime next year, and then it is time to wrap up HW3 development.

Perfect timing for HW4, when all the faithful will claim that will solve it (I think the best we can hope is that it will make it better, but solving isn’t happening for years).

 

[bhakan]

Nearly everything I mentioned he called out.

I’m just saying it’s not an example of a stellar drive.

We need to be realistic about what it can do. And we need to be realistic about how that compares to a typical driver.

It seems like it should pass a driving test at a minimum.

One of the test elements is probably "do not block the crosswalk." My car was creeping forward and blocking the crosswalk when people wanted to cross as it was a walk signal for them. i had to disengage and reverse my car and of course there was a car behind me. In crowded urban neighborhoods and streets i am not sure it will pass the test.

 

[Goose66]

I would rate it a beginning student driver in my little town.

Source: I have a 15 year old daughter.

It is quite remarkable how quickly my son went from no knowledge of driving to being a very competent and safe driver in the 8 or 9 months of trading in the context of how much progress FSD has made in the last 4 years.

BTW the Model 3 with single pedal operation is an awesome car to train a new driver with.

 

[aronth5]

Yeah, I would say that for the market and for everyone else here, the bloom is really off of this rose.

We’ll get single stack that doesn’t have to be turned off sometime next year, and then it is time to wrap up HW3 development.

Perfect timing for HW4, when all the faithful will claim that will solve it (I think the best we can hope is that it will make it better, but solving isn’t happening for years).

Always the pessimist
To cheer you up, Tesla has gone live with 28 supercharger locations in December to date. (US) Feel better?

 

[JulienW]

Always the pessimist

....and the TRUE optimists have had Robotaxies making money for them for a couple of years now.

 

[Ramphex]

Your reply makes zero sense

 

[Ramphex]

74 year old here, I don't have a problem with correcting FSD and it is helpful on long drives. I know as I get older I will lose more of my driving abilities so I am participating in this experiment with the hope of some day realizing the advertised potential of full self driving so I can continue to be active and safe.

You’re not a typical 74 year old. Speaking from experience working with geriatrics, large majority still have flip phones, are not a fan of “technology,” and I could definitely see them struggling to babysit FSD in its current state.

Now if FSD was what it was promised to be, that’s a different story. I can then definitely see it being helpful to pretty much everyone.

 

[Ramphex]

Always the pessimist
To cheer you up, Tesla has gone live with 28 supercharger locations in December to date. (US) Feel better?

If he’s a pessimist, you’re an apologist. What’s worse?

Debatable.

 

[AlanSubie4Life]

To cheer you up, Tesla has gone live with 28 supercharger locations in December to date. (US) Feel better?

That’s great. I enjoy road trips in my car, 1000 miles per day does not seem difficult. AP is good. Can’t ever have too many Superchargers, to minimize drama. The network is by far the best; no fiddling.

FSD is not good.

 

[AlanSubie4Life]

If he’s a pessimist

I’m not a pessimist. Very bright and cheery. Like to try to predict things correctly.

 

[Ramphex]

That’s great. I enjoy road trips in my car, 1000 miles per day does not seem difficult. AP is good. Can’t ever have too many Superchargers, to minimize drama. The network is by far the best; no fiddling.

FSD is not good.

Couldn’t have said it better.

And that “pessimist” wasn’t a dig at you by any means. There is a huge difference being a realist and being a pessimist, but that dude will probably never understand the difference.

 

[Dewg]

The pessimist says "Things are terrible, they can't possibly get any worse."
The optimist says "Oh yes they can."

 

[Mardak]

Could also be new vehicles in the testing fleet

Has anybody heard of any new additions to FSD Beta 10.69.3.3 anywhere? It seems like especially because it's still in limited release only up to the early access group (which includes the pre-Safety-Score group), it would be quite notable for someone to be newly added. Because we haven't heard much of these new FSD Beta additions from factory vehicle software like 2022.28.300 or what had been too-new-for-FSD-Beta like 2022.40.4.1, maybe it is indeed as you suggest of special "testing fleet" that has non-disclosure agreements with Tesla vs the rest of us.

TeslaFi shows since December 15th, FSD Beta 10.69.3.3 has 18 vehicles updating to it and 12 of those did not previously have FSD Beta. If this self-selection bias to be using a service like TeslaFi is like previous sampling biases, there could have been a few hundred vehicles newly added yet none have said anything publicly?

 

[Ramphex]

Has anybody heard of any new additions to FSD Beta 10.69.3.3 anywhere? It seems like especially because it's still in limited release only up to the early access group (which includes the pre-Safety-Score group), it would be quite notable for someone to be newly added. Because we haven't heard much of these new FSD Beta additions from factory vehicle software like 2022.28.300 or what had been too-new-for-FSD-Beta like 2022.40.4.1, maybe it is indeed as you suggest of special "testing fleet" that has non-disclosure agreements with Tesla vs the rest of us.

TeslaFi shows since December 15th, FSD Beta 10.69.3.3 has 18 vehicles updating to it and 12 of those did not previously have FSD Beta. If this self-selection bias to be using a service like TeslaFi is like previous sampling biases, there could have been a few hundred vehicles newly added yet none have said anything publicly?

It’s possible. Wonder if anyone has ever crunched the numbers, maybe @EVNow, but curious how many Tesla owners are active on this forum vs how many there are total worldwide.

 

[sleepydoc]

It is quite remarkable how quickly my son went from no knowledge of driving to being a very competent and safe driver in the 8 or 9 months of trading in the context of how much progress FSD has made in the last 4 years.

BTW the Model 3 with single pedal operation is an awesome car to train a new driver with.

Yes - humans have a capacity to learn, unlike computers. Part of intelligence is identifying new patterns, generalizing and extrapolating them to new situations, and predicting based on past experiences. Computers do little to none of this which is why I consider the term ‘artificial intelligence’ a misnomer. The vast majority of AI is actually pattern matching, not true recognition.

When comparing one needs to think more of evolution than learning. Tesla is trying to program algorithms that have never been done before so the comparison to a teenager learning to drive isn’t a completely apt one.

 

[Phlier]

The pessimist says "Things are terrible, they can't possibly get any worse."
The optimist says "Oh yes they can."

My wife likes to use the old classic "glass" saying about optimists/pessimists, with her own twist...

Wife, talking to me: "You know, they say the optimist sees the glass as half full. The pessimist sees the glass as half empty. Then there's you, who sees it as that damn glass."