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

Things my wife said about Navigate on Autopilot tonight

This site may earn commission on affiliate links.
Someone said I take all this criticism "personally". My experience is much like that of @Skryll and other positives, except maybe that I "drive" the NOA more actively, using the speed max roller as a limiter, and dropping in and out of AP, sometimes several times in a minute in very ambiguous situations. I hardly ever experience any of the problems that are reported. It's not that they cannot happen, but they can be avoided or, worst case, dismissed with a shrug of the shoulders, once you know the car and its current limits.

It's good that people are raising these issues, I hope Tesla is reading. It is a work in progress. But looking at all these rather bitter complaints, a potential buyer would conclude that the whole AP/NOA system is worthless. That e.g. phantom slowing (that's I've seen all of 2 times in the last few months) happens all the time, that the Model 3 automation drives incompetently and dangerously. And that would simply not be an accurate picture.

You’re essentially saying you’re driving the car while dancing around Autopilot’s limitations and failures. You consider that to be normal and expect others to do the same.

It’s not exactly normal though. Most people don’t consider driving using thumbwheels and stalks to be fun, safe or normal. We already have vastly more experience with the car’s regular interface, i. e. the steering wheel and pedals.

I personally get where you’re coming from, I ran the same geeky tests all the time, however, I never considered the AP’s abilities to be either helpful or stress-relieving. Watching that idiotic software and waiting for it to fail wasn’t fun for too long, despite having developed a reasonable understanding of when/where it’s likely to fail.

The point here is that AP needs to be a tool, not a toy. As such, it needs to do whatever it can do much much better than it currently does. I may be in the minority (around here, quite likely, in the real world, likely not) but I believe that fewer features that work very well are preferable to a lot of features that work poorly.
 
Last edited:
Well, the car knows what road it is on

The car also (kinda) knows the speed limit- yet lets you exceed it.

Hell the car knows where the ocean is but it'll let you drive into it anyway.

Level 2 systems rely on the human behind the wheel to use them correctly- so it's 100% absolutely user error when you don't.

Personal responsibility. Take some.
 
It's not an either/or proposition. There are entirely separate teams of engineers with very different skill sets working on Autopilot and infotainment.
...
Building stuff like games and Easter eggs on the side is also a big morale booster, allowing the dev team to achieve some quick wins and lower their crazy-high stress levels while having a little bit of fun.

(Source: 10 years of experience managing software engineers and 10 years of being one)

I’m going to hazard a guess and say that, if you try that resource allocation strategy at your work, it won’t go too well.


Needed (‘cause it was sold already):
Solid, working AP

Not needed:
Games


We already have 256 platforms to kill time on, when we choose to, the car does not need to be one of them.

If the team that does games integration is not qualified to work on AP, then a simple idea would be to hire the correct skillset and let the other guys build real games somewhere else. On the other hand, I think you’re seriously overestimating the skillset needed to work on this AI thingy.

Morale boosting is not a job description. One can boost a team’s morale by throwing some money at them, perhaps the same money that is otherwise wasted on games.

(Source: random internet dude with a a lot more experience doing that software building and management thing)
 
The car also (kinda) knows the speed limit- yet lets you exceed it.

Hell the car knows where the ocean is but it'll let you drive into it anyway.

Level 2 systems rely on the human behind the wheel to use them correctly- so it's 100% absolutely user error when you don't.

Personal responsibility. Take some.

LOL. No, it’s not that easy. If your fancy driving automation can’t handle a HUGE percentage of roads, then you can’t advertise it like Tesla does, add fine print in the manual that it really only works on the simplest traffic conditions on a small subset of our road network and expect that people won’t try it on roads they actually drive on without complaining.

And even on supported roads it performs poorly in many cases as you can see reading this thread.

Haha. It’s hilarious how Tesla sells this as Enhanced Auto Pilot with FULL SELF DRIVING capability but then if someone dares to use it on a road other than the 880 in Fremont people like you change tune that it really is not made for other roads (because it sucks).
 
LOL. No, it’s not that easy. If your fancy driving automation can’t handle a HUGE percentage of roads, then you can’t advertise it like Tesla does


Tesla doesn't advertise. At all. Their advertising budget is pretty famously $0.00. And they certainly don't advertise that the current system works on "all" roads or anything remotely close to it.

But if you bother to read the manual they're very, very, clear on what kinds of roads the system is intended to be used. (and they also have warnings you have to agree to to enable the system at all the first time- and more warnings that show every single time you engage it).

Maybe give those a read between baseless rants?


Haha. It’s hilarious how Tesla sells this as Enhanced Auto Pilot

Tesla doesn't sell EAP, and hasn't in over 6 months.

Nor did they ever advertise that it worked on roads with cross-traffic when they did sell it.

Once again facts and reality don't appear to figure into your complaints.




We already have 256 platforms to kill time on, when we choose to, the car does not need to be one of them.

If the team that does games integration is not qualified to work on AP, then a simple idea would be to hire the correct skillset and let the other guys build real games somewhere else.


Uh... you know nobody is "building" games at Tesla, right?

All the games have been developed (most of the decades ago) by third parties that aren't employed or paid by Tesla.

Tesla is just loading games that already exist onto the car.

Literally the only thing any "programmer" at Tesla had to do was spend 5 minutes adding the launcher to the menu in the car.
 
  • Like
Reactions: Foxtrotter
Tesla doesn't advertise. At all. Their advertising budget is pretty famously $0.00. And they certainly don't advertise that the current system works on "all" roads or anything remotely close to it.

Come on, it's pretty obvious he used "advertise" as in "announce" as opposed to "put up a billboard and run TV ads".

Maybe give those a read between baseless rants?

The Internet would be pretty void and pointless without baseless rants. We need those!
(not that the one you were referring to was one of them, it was an opinion, therefore not senseless)

Tesla doesn't sell EAP, and hasn't in over 6 months.

So, they've been misleading for all this time minus 6 months then?

Uh... you know nobody is "building" games at Tesla, right?

All the games have been developed (most of the decades ago) by third parties that aren't employed or paid by Tesla.

Tesla is just loading games that already exist onto the car.

Which is why I said "the team that does games integration".

Literally the only thing any "programmer" at Tesla had to do was spend 5 minutes adding the launcher to the menu in the car.

What does that team do after those 5 minutes? Sit around and get paid for nothing?
People in the know here keep assuring us that it's a SEPARATE team from the one working on all the other features and that them working on games in no way affects the timeline and quality elsewhere.

^ Pretty good example of reductio ad absurdum, no?
 
Tesla doesn't advertise. At all. Their advertising budget is pretty famously $0.00. And they certainly don't advertise that the current system works on "all" roads or anything remotely close to it.

But if you bother to read the manual they're very, very, clear on what kinds of roads the system is intended to be used. (and they also have warnings you have to agree to to enable the system at all the first time- and more warnings that show every single time you engage it).

Maybe give those a read between baseless rants?




Tesla doesn't sell EAP, and hasn't in over 6 months.

Nor did they ever advertise that it worked on roads with cross-traffic when they did sell it.

Once again facts and reality don't appear to figure into your complaints.







Uh... you know nobody is "building" games at Tesla, right?

All the games have been developed (most of the decades ago) by third parties that aren't employed or paid by Tesla.

Tesla is just loading games that already exist onto the car.

Literally the only thing any "programmer" at Tesla had to do was spend 5 minutes adding the launcher to the menu in the car.

Now that that’s out of the way, what does your wife/husband say about Navigate on Autopilot?
 
Come on, it's pretty obvious he used "advertise" as in "announce" as opposed to "put up a billboard and run TV ads".

I mean- he'd still be wrong in that case... they never ever claimed existing features worked anywhere but the places they say it works.

They have described ADDITIONAL features that they hope to enable some day in the FUTURE, but were pretty clear on that point too.



So, they've been misleading for all this time minus 6 months then?

Also no- since EAP has been pretty clear where it's intended to work and where it's not.

It's right there in the manual. Which apparently nobody reads. It also requires you to acknowledge its limits when you first enable the feature. And then warns you again to a lesser degree every single time you turn it on.

Not sure what anybody found misleading in anything they actually said? Can you be specific?



Which is why I said "the team that does games integration".

Which there's no evidence of existing as its own team.




What does that team do after those 5 minutes? Sit around and get paid for nothing?

What team?

Adding a launch option to a menu takes whoever does all the OTHER UI stuff 5 minutes, then he goes back to his normal work.


People in the know here keep assuring us that it's a SEPARATE team from the one working on all the other features and that them working on games in no way affects the timeline and quality elsewhere.

Not really. They just point out whoever does it isn't the guy working on AI/FSD.

Given how near-zero work is required as I said it's almost certainly a guy already there doing the normal UI stuff, who does this as one more task.

Know how the nav stuff moved from the right side to the left side a while ago? THAT guy does menus. Not AI/FSD.

Games, specifically add them to the launcher, would be roughly 0.01% of his job.


^ Pretty good example of reductio ad absurdum, no?

Nope. Whole thing has been a pretty good example of folks grossly overestimating the amount of "work" anybody at Tesla puts into games though!
 
Last edited:
Now that that’s out of the way, what does your wife/husband say about Navigate on Autopilot?


She loves it- gets annoyed any time we take the non-Tesla vehicle we own anyplace complaining how it requires one to do all the driving work themselves like some sort of peasant.

Like 95%+ of our driving is highway and it makes it tremendously better and less stressful.

EAP is literally the reason we bough the car- otherwise I'd have just kept my Lexus (where even in the then-new model of the same car the nearest equivalent features were absolute crap compared to Teslas)
 
I disagree. If it isn’t safe (!) to use on non divided limited access roads then it should be disabled. It’s not a video game like some on TMC seem to think it is.

I have been driving AP1 since 2016 and AP2.5 since 2018 and it is definitely safer with than without, even when you have to use it as a driver assist feature within the limitations of what it can and cannot do.
 
I disagree. If it isn’t safe (!) to use on non divided limited access roads then it should be disabled. .

It's not safe (or even legal!) to exceed the speed limit- guess the car should auto-limit top speed then, right?

Also it's unsafe to drive without a seat belt- the car should refuse to shift out of park if yours or your passengers isn't buckled, or automatically pull over to the side of the road if anyone unbuckles while driving, right?
 
  • Like
Reactions: Skryll
Browsing this thread, I see lots of suggestions of follow-distance and creep, but has anyone tried lowering regen?

When I get the most jerks is when AP suddenly cuts velocity and regen jumps in. It doesn't seem able to feather the pedal and instead goes into a coast which turns on 100% regen.

I have AP3 (without NoA) and I run with low regen. It doesn't have any effect.
 
  • Informative
Reactions: AlanSubie4Life
Also, the whole "don't use AP when not on interstates" is ridiculous. So we only have cruise control on divided highways? What if I have a 2 hour drive on a secondary road? If Tesla doesn't want us to use TACC on non divided highways, they should run standard cruise control in those areas instead, just like cars without AP.