Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register
This site may earn commission on affiliate links.
Even with the prior versions which were not nearly as good, it was very common for people to feel that the software got better after a week or two of use. Give it a little time, get used to it and if you've got any open-mindedness at all you'll probably enjoy it more after a little use. Also don't give up on it to the extent that you won't give the next version, and the one after that, a decent trial and learning curve to see what's happening. In any case thanks for the report and for continuing to work with it as long as you can.

More often than not I hear people bad mouth how horrible prior versions were only after a new version is released. I never understood that.
 
  • Funny
Reactions: DBeckwith
Why don’t these limitations (bugs) apply to simpler turns? Or are you suggesting only right turns, if left is unprotected?
Most UPL's for me have been fine since they are simpler and don't require interacting with crossing traffic in both directions. But when you have a UPL with high speed crossing vehicles, an optional median to get to with limited visibility requiring lots of creeping maybe it's best to go right. Better to take a little bit longer and be safe.
 
Elon said it was a lie.

Elon isn't saying much. Maybe SEC restrictions?

I saw one person suggest the low cost design is still needed for robotaxis so technically it might not be cancelled.

Screenshot 2024-04-05 085417.png
 
  • Like
Reactions: FSDtester#1
Per CNBC's Phil LeBeau - Reuters says sources as well as messages seen by Reuters that Tesla employees were told late Feb the low cost Model 2 would be scrapped. The profit margins wouldn't be enough to warrant it. Instead they would focus on an entry level robotaxis on the same platform (minus steering wheel).

Makes sense given the Mexico giga factory was shuttered a few months back with Austin taking on Model 2 production.

I can hardly wait to make robotaxis profits while sitting on a remote beach. :)
 
Per CNBC's Phil LeBeau - Reuters says sources as well as messages seen by Reuters that Tesla employees were told late Feb the low cost Model 2 would be scrapped. The profit margins wouldn't be enough to warrant it. Instead they would focus on an entry level robotaxis on the same platform (minus steering wheel).

Makes sense given the Mexico giga factory was shuttered a few months back with Austin taking on Model 2 production.

I can hardly wait to make robotaxis profits while sitting on a remote beach. :)
I don't think there's enough (or any) basis to consider this rumor true. Especially with Elon's responses.

He basically approved that both the low cost car and RT will be on the same platform, but it's a lie that the low cost would be cancelled.
 
  • Like
Reactions: STUtoday
More often than not I hear people bad mouth how horrible prior versions were only after a new version is released. I never understood that.
Really? That's basic human nature. If someone gives you discount ice cream and you've never had it before, you'll likely enjoy it. Then someone gives you a better mid-brand and suddenly you realize the previous one wasn't as good as you thought. Then you taste premium ice cream and you're hooked.
 
Piggybacking on this, my car drove itself out of the gated parking lot (once the gate opened) onto the street today. I didn't even know FSD could ever do this. Maybe one day it'll be able to drive from a parking spot to a closed gate, wait for the gate to open, then drive away.
Mine already does this, except for the parking spot start. I've found I cannot start FSD unless I'm already in Drive and my foot is off the brake. But if I pull out of my driveway then engage FSD, it will go right to the gate, wait for it to open (proximity sensor opens the gate on the way out) and go through. On the way back I have the Homelink programmed to open the gate when I get within 20 feet. It will almost stop at 40 feet and creep (very) slowly forward waiting for the gate to open. I have it programmed to open at 20 feet to prevent opening the in gate when I am going out the out gate. Anyway, eventually it creeps forward enough to trigger the gate, and then it waits until the gates completely stop moving before going through. Primary takeaway from me is that it is too slow, but it works.
 
  • Like
Reactions: FSDtester#1
Really? That's basic human nature. If someone gives you discount ice cream and you've never had it before, you'll likely enjoy it. Then someone gives you a better mid-brand and suddenly you realize the previous one wasn't as good as you thought. Then you taste premium ice cream and you're hooked.

True but in this case it's really the other way around. We aren't neophytes. We have been driving for years, refined our craft, know the rules of the road, understand driver etiquette, aware of environmental context, can read signs and nonverbal cues, and expect a certain level of driving comfort. We have certain and specific expectations for FSD beyond that short lived first time excitement of driving in an FSD vehicle.
 
  • Like
Reactions: AlanSubie4Life
I don’t know if it applies to AI or not but my understanding is that the neural net is taught by showing it examples of ‘good’ driving. If that’s the only teaching method then the only way to learn what not to do is by inference. I could be wrong and maybe someone with more knowledge on NN training can chime in.
I don't claim to have more knowledge, but it seems to me that a system like FSD needs to be told what the right thing to do is, regardless of circumstance. For example, it's considered "bad" to drive off a cliff. But the system doesn't need to be told that it's bad to drive off a cliff so much as it's good to stop when seeing a cliff. That is, there is always a correct response to any driving situation.

Is there a specific situation that you think the system needs to be told "this is bad"?
 
True but in this case it's really the other way around. We aren't neophytes. We have been driving for years, refined our craft, know the rules of the road, understand driver etiquette, aware of environmental context, can read signs and nonverbal cues, and expect a certain level of driving comfort. We have certain and specific expectations for FSD beyond that short lived first time excitement of driving in an FSD vehicle.
10.3.1 blew my mind, but it was terrible. It basically couldn't drive, but still wild that it was attempting to. When the next version came out that was better, I was crazy happy and *sugar* all over 10.3.1, but still wasn't good and I knew that the entire time.

V12 is way better than V11 in some many areas that it warrants people bad mouthing the deficiencies, but V12 still isn't close to perfect.

Although with lane selection and random turnlane cutouts on a main road 10.69 has still been our best version.
 
10.3.1 blew my mind, but it was terrible. It basically couldn't drive, but still wild that it was attempting to. When the next version came out that was better, I was crazy happy and *sugar* all over 10.3.1, but still wasn't good and I knew that the entire time.....
You didn't have 10.3? Because 10.3.1 was like heaven on earth compared to the 4 days of a living nightmare 10.3 was. :oops:
 
2024.3.10 finally installed. It took 45 minutes, along with 30 minutes for 2023.44.30.30, plus 150 minutes for downloading the updates.

In the long process tesla pushed 2023.44.30.30 to my car after I inquired as to why I didn't have it since they installed it yesterday. It took two more rounds of downloading that update between 8:15 and 10:20 before it said I could install. Something was obviously wrong because that was on top of the SC downloading and installing it yesterday. I moved the car to a towable location then started the install. When that finished, tesla immediately pushed out 2024.3.10 and 35 minutes later it finished downloading. I triggered the install from home (since the car was in an appropriate place to install.)

I pointed out to tesla that I still don't have the latest NAV. They instructed me to seek out strong wifi for that to download (i.e. "No, we aren't going to push that to you as well as these updates." but at the same time confirming this is too large a file to mobile hotspot update.)

I'll go reboot the car and then start playing with the settings and exploring my new UI since a helluva lot has changed since November. No screaming goat for me, but eventually I'll load up a sound file to use for the walkaway lock confirmation. Something that will penetrate my brain since I don't 'hear' the honk.

I'm in no rush to test out FSDS. Friday afternoon the traffic is too heavy around here to take the car out for a 'drive' and we seldom attempt to leave or return to home because of the volume.
 
The problem of tapping the accelerator, having it speed up, and then drive too fast around a curve is still there. It's dramatic and dangerous. However, I realized that perhaps that problem has always been there, but I never saw it because I never had to loose the car to make it to fast enough??
Meant to write "Goose the car," not "Loose the car." Autocorrect or brain fart and definitely a failure to proofread.
 
  • Like
Reactions: MitchMitch
Another weird behavior I've noticed with 12.3.3 is that it will occasionally get stuck at 5-7mph OVER the set "Maximum" limit (e.g. if the speed limit changes) and won't slow down even if the maximum limit is adjusted with the scroll wheel; it requires a full disengagement and re-engagement in that case to get it to slow down. Seems like a C++ bug more than a neural net error.
I've seen this at some spots where the speed changed from 65 MPH to 55 MPH but not others.

For me it eventually changes to the new limit (after 1 minute?). Time enough to get a speeding ticket. Does yours change eventually?
 
Yeah it's exactly what I predicted. Quote below. I'm not surprised. Not my first rodeo!

The beer is mine (4/7 or 5/7 depending on whether you count driving at 25mph on a ~45mph road a failure), though I maybe will not collect it if Chuck does less than ten attempts (@Daniel in SD and I will have to examine the fine print of this rule in the 10.69 thread). It's a moral beer if nothing else. The best type of beer. 🍺

Do remember that 12.3 and 12.3.x does not incorporate any additional unprotected left training - that all occurred after the 12.3 release (before 12.3.3 but that isn't relevant since it doesn't appear to have been retrained).

NEXT time I will finally lose. I am betting against FSD 12.4 as well. Any perfect performance with 7 or more attempts is ok by me. The march of the 8.6s to 10.0s.
The primary unprotected left I take is to exit our neighborhood onto a feeder road. It’s only one lane, 40 mph so pretty easy. IME, FSD still hesitates, stopping at the sign, then creeping, then stopping again before finally committing and it’s still too slow to start. It doesn’t seem to be as bad as what @AlanSubie4Life describes but is definitely not confident. I’m sure this is magnified with multi lane left turns like chuck does. When negotiating these, you can’t hesitate; you need to find a spot then take it.
 
I got FSD disabled warning three times already (because I was not looking at the road, but on the screen and it didn't like it). I only have two more strikes left.

Question: What happens if I exhaust all my warnings? Will my FSD get disabled permanently and thats it? No recourse after that?
You will only lose it for a week.

Tip: If you are going to look away from the road or at the screen, be sure to constantly jiggle or apply torque to the steering wheel at the same time. Plus, take quick glances at the road, which is a good idea anyway for safety.
 
I would like it to not tap the brakes, then proceed, when the left-hand-turn light goes yellow and my light stays green (I have video of this, which I absolutely will not post). How on earth did the miraculous NN learn to do that completely incorrect behavior?

Maybe it just takes the average of the light colors. 2 out of 3 were green.
That’s what humans do half the time - light turns yellow, they start to brake then decide they can make the light and proceed.