Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Nag reduced with 2017.46

This site may earn commission on affiliate links.
Agreed
Tesla autopilot cant be trusted 100% . Granted it has gotten MUCH better since ap2 first came out. I have to say its been performing wonderfully after the last few updates.

I had a 1,000 mile road trip this weekend. Autopilot 99% of the way. I had only ONE hiccup where the car went a bit out of control. I thought it was going to go into a different lane. It was just a split second before I took control. Not sure if the car was just adjusting and I didn't want to wait around to find out.

Thankfully I keep a hand on the wheel at all times

I still find autopilot to be a safer driving experience compared to manual driving. Car never takes eyes off road it watches cars in front of cars. It can brake in an emergency before you can.
In general what were the big downsides to the original AP2 software? I got my car ~3 months ago and I've been relatively happy with its features because I didn't see what they originally were so awful/ lagging of. Lane keeping could be better when there are turn lanes approaching... and the neat little people/bikers, and the view of cars in other lanes are missing. I've noticed my radar detects cars only when it's gotten much much closer to the stopped car at a light in front of me when I'm driving over about 50.
I'm But what else is really a problem for EAP
 
In general what were the big downsides to the original AP2 software? I got my car ~3 months ago and I've been relatively happy with its features because I didn't see what they originally were so awful/ lagging of. Lane keeping could be better when there are turn lanes approaching... and the neat little people/bikers, and the view of cars in other lanes are missing. I've noticed my radar detects cars only when it's gotten much much closer to the stopped car at a light in front of me when I'm driving over about 50.
I'm But what else is really a problem for EAP

I cant remember the details, but you missed the hurdles.
Car was very jumpy and unreliable. The car would not stay smoothly in a lane and it would sometimes go to different lanes all together. It was also limited to lower speeds back in the beginning when they first released the software.

It really does feel "as smooth as silk" compared to what it originally was. It used to get my passengers sick so I would just take control. It still is not 100%, but it keeps getting better.

Turns out it was a brand new system they wrote in 6 months. The original AP relied on a tweaked version of a system provided by another vendor (i think mobile eye was the name).

As for your radar, I suspect the car actually knows quicker than it displays it on screen. I have had scenarios where the vehicle 2 cars ahead of me brakes quickly and my car braked right away even though the car in front of me did not brake yet. This all happened without the 2nd vehicle being displayed on my headunit. I hope that makes sense lol
 
  • Informative
Reactions: tkdkenny
No hands touching the wheel at all? Are you sure you were inadvertently touching it realized you were present? This is the first report and I find it very skeptical that you went 25mins with 0 nags and no hands on the wheel...
You shush now! Let me enjoy my fantasy of a redbull-can-free steering wheel for a few moments before reality comes crashing back in and this claim about .46 is debunked.
 
Changing lanes does cause the steering wheel to move, yes? I see mine move at least.

Which would reset the nag timer.

Would be nice if we got back to the more relaxed nagging of previous versions. Not a new feature, though. Just parity with the old baseline through tuning.

You wouldn’t consider auto-sensing wipers a new feature, right? Well, some would because they’ve never experienced AP1.
 
Before, you had to tug on the wheel every 1 min 15 seconds, and make sure you did it just hard enough to clear the nag without disengaging. If it’s true that all you have to do now is change the cruise speed, that’s huge, as there’s much less care you have to use with that.

Any chance this applies to follow distance? If so, that’d be even easier and with less of a chance of any change in the car’s behavior.

Just check and yes - changing distance does indeed reset the nag timer as well. Glad Tesla made this change. No need for the car to request a tug of the wheel if you are obviously paying attention by changing speed/lanes/distance. Another example that there is more to the updates than the advertised chill mode etc...
 
.46 model S AP2

I can confirm this morning that changing speed resets nag timer. However, I find it much easier to lightly tug the wheel than to fiddle with the speed knob. I wish they would use some sort of interior camera just to check the driver is still awake and looking more or less ahead once in a while rather than pure physical touches. Maybe the AP2.5 interior camera will do that, but what will happen to us AP2.0 folks?
 
Now this makes sense, sorry @calisnow just get that envelope ready for St. Jude’s....
I had a nice candlelight dinner with my stack of hundred dollar bills last night - I assured them I love them and will always keep them safe with me - and that @BigD0g kindly volunteered his own stack to take their place at St. Jude's next year. FSD features will be so nice in January :)
 
.46 model S AP2

I can confirm this morning that changing speed resets nag timer. However, I find it much easier to lightly tug the wheel than to fiddle with the speed knob. I wish they would use some sort of interior camera just to check the driver is still awake and looking more or less ahead once in a while rather than pure physical touches. Maybe the AP2.5 interior camera will do that, but what will happen to us AP2.0 folks?

Eventually some learn that there are workarounds. I for one feel not an ounce of guilt using a workaround because Elon designed the system to be hands free. It was delivered that way in Oct 2015 and I still believe the nag system is a concession to Tesla's legal and PR teams - a CYA move that's window dressing and nothing more.

Even if there IS data showing that resting a couple fingers or even a hand on the wheel vs keeping hands in your lap or on the arm rests reduces crash rates, WE do not have that data. In addition, Tesla has never once claimed that keeping hands in the wheel during AP reduces accident rates vs completely hands free use.

The ONLY data we have is Tesla's claim that fleet wide airbag deployments in AP equipped cars were reduced by 40 pct after AP went live in Oct 2015.

Every single post on TMC admonishing folks to keep hands on the wheel vs in their laps is founded on 100 percent pure speculation as to what constitutes safe autopilot use and nothing more.
 
This doesn't work for me. If I hold the wheel at the bottom it will nag me.
It is not "just hold" there is a slight resistance. Hard to explain. But as the road turns (generally never on a exact straight line) the wheel will turn with with road. You can feel that in your hand. I tend to help it just a little. If the road does go straight for a very long time I believe you have to still give a little be of movement. But not nearly what is required when you get the actual nag. I think the point is you are pretty much resting you hand or hands on the bottom portion of the wheel in a comfortable manner and it becomes routine. If you don't then you will get the nag where you have to do something special. For me not resting on the wheel means I am just resting them somewhere else so is not a big deal.
 
  • Like
Reactions: kavyboy
On my long drive last night, I played around with different ways of handling it. It seems to me that Autosteer is a little more erratic when my hand is resting on the wheel. From what I could tell, it seemed to be the case that it still takes some input from the wheel even with autosteer on(I’ve noticed this before in that I could nudge it toward one lane or another in a lane split). With my hand on the wheel, it seems to end up being biased toward one direction, resulting in more wavering in the lane and occasionally a motion swinging the car back towards the center of the lane.

Driving for extended periods with and without holding the wheel, it performed significantly better when I let it do it’s thing and just occasionally nudged it.

Of course, it’s possible I’m just crazy or it’s all coincidences, but I tried it enough times to convince myself there’s a noticeable difference, and I find the occasional clearing of the nag becomes more or less automatic.
 
Every single post on TMC admonishing folks to keep hands on the wheel vs in their laps is founded on 100 percent pure speculation as to what constitutes safe autopilot use and nothing more.

While I'd love nothing more than to agree with you 100%, I can't. At least not until AP2 no longer blindly follows high contrast tar lines believing them to be lane markers. Less bad in .44, but was still bad in .42. I don't have .46 yet so can't comment on it.

If you've never experienced a diagonal crack repaired with a pour of tar, consider yourself lucky. Very lucky. It will, at highway speed, veer violently towards something that, without a hand poised to grip the wheel to instantly cause a disengagement, will scare the crap out of you.

Perhaps it's not common in CA without the winter weather and tons of salt, plowing and freeze and thaw cycles, but it is on many roads in PA.
 
Hold the wheel lightly with ONE hand just above the flat portion at the bottom, along the side, just letting the weight of your hand bear down. Don't steer, but don't keep the AP from steering. You will never, ever get a nag. If you need to do something with your hands, then do it and put one back when you can. Switch hands if you need to. Prop your elbow on the console.
I do this as well. I have literally thousands of miles of AP1 in this position without even the first nag that I can remember.
Ironically, when I grip the wheel at 2-and-10 and I am really paying attention, that's when I get the nags, which is the most dangerous time to be distracted.
 
From what I could tell, it seemed to be the case that it still takes some input from the wheel even with autosteer on(I’ve noticed this before in that I could nudge it toward one lane or another in a lane split).
I used to be able to nudge it within my lane quite a bit. It wasn't just a perception, it was noticeable and extremely useful. I like to give semis a bit more room without disengaging, for example. Somewhere along the line, I lost that capability.