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

NoA is worthless, Autopilot is essentially unimproved from 2016-2017, and everybody has caught up.

This site may earn commission on affiliate links.
I agree with most of your observations, but the number of speed cameras we have does not quite fit with the idea of decriminalisation.

If you study the Highway Code, you discover that on certain roads, the presence of street lighting dictates a 30 mph limit if not signed to the contrary. If there is a central divide then the rules might be different, and on a Motorway, it's 70mph.

Absolutely agree that driving convention (which has a lot to do with safety) varies from region to region and even within neighbourhoods. I have lost track as to if up hill or down hill traffic has priority when passing parked cars that block the carriageway or if it even makes any difference now. (used to be a rule similar to our Red & Amber traffic light phase that related to allowing for getting your car to stop or go!)

Realistically, AP will not crack these situations while there are still cars with humans at the controls. And, paradoxically, there will always be human drivers at the controls while there are still so many situations that AP can't handle.

you are approaching this from a british mindset. the uk speed cameras are basically slowdown signs. they are painted bright yellow, advertised beforehand, only take a pic from behind. you have to work very hard to get a speeding ticket. in other countries its much easier to get a speeding ticket. and yes indeed there are default limits which apply but its mostly up to the driver to identify those (usually no signs) but if you drive i.e. 80 mph on a motorway noone batters an eye. Try driving 130 in a 110 and you will very quickly find yourself with a 300 pound speeding fine.
 
  • Like
Reactions: mikes_fsd
you are approaching this from a british mindset. the uk speed cameras are basically slowdown signs. they are painted bright yellow, advertised beforehand, only take a pic from behind. you have to work very hard to get a speeding ticket. in other countries its much easier to get a speeding ticket. and yes indeed there are default limits which apply but its mostly up to the driver to identify those (usually no signs) but if you drive i.e. 80 mph on a motorway noone batters an eye. Try driving 130 in a 110 and you will very quickly find yourself with a 300 pound speeding fine.

They are valid points.

I had two cases in mind relating to speed. One on a stretch of road I travel frequently where the limit goes from 30 to 40 to 30 back and forth many times. Not easy for a human driver to remember for sure once you pass the initial limit sign. Limits painted on the road surface are worn and some burned off poorly when changed / relocated.

If geolocated / map based limits were accurate, then there would be no problem, but having multiple inputs for speeds doesn't help as you need only one correct input. The speed camera in this section of road is just as likely to catch local drivers as any. It's just a mind game! AP needs to hit the correct speed as you pass the sign, not 50 m past it.

The second case was on unfamiliar roads with poor lighting, just after a roundabout. Obscured limit signs, no obvious street lighting. Few if any visible clues especially if not familiar with the road.

Again, map data would make this no problem, but not sure what purpose adding in sign data would serve. One accurate value is all you need.

Would you always go with lowest speed limit detected?

I agree that for example roads in France feel more structured and predictable in some respects, and speed cameras less obvious.

That said, I don't think speed control is that big of an issue for AP. Many possible solutions. As long as engineers are looking at all the edge cases.
 
They are valid points.

I had two cases in mind relating to speed. One on a stretch of road I travel frequently where the limit goes from 30 to 40 to 30 back and forth many times. Not easy for a human driver to remember for sure once you pass the initial limit sign. Limits painted on the road surface are worn and some burned off poorly when changed / relocated.

If geolocated / map based limits were accurate, then there would be no problem, but having multiple inputs for speeds doesn't help as you need only one correct input. The speed camera in this section of road is just as likely to catch local drivers as any. It's just a mind game! AP needs to hit the correct speed as you pass the sign, not 50 m past it.

The second case was on unfamiliar roads with poor lighting, just after a roundabout. Obscured limit signs, no obvious street lighting. Few if any visible clues especially if not familiar with the road.

Again, map data would make this no problem, but not sure what purpose adding in sign data would serve. One accurate value is all you need.

Would you always go with lowest speed limit detected?

I agree that for example roads in France feel more structured and predictable in some respects, and speed cameras less obvious.

That said, I don't think speed control is that big of an issue for AP. Many possible solutions. As long as engineers are looking at all the edge cases.

In the UK I just go whatever I feel is best -and limits dont change that often there anyway. I once got pulled over before Dover by an unmarked police vehicles for driving 150km/h in a 110. But tbh I should have seen it coming - at the time I assumed there were no unmarked vehicles in the uk (even though they are advertised in front of dover) and they did follow me for probably 3-5min before they put their bluelights on. The fine wasnt even that big and not that many points. It was roughly the same as driving 10km/h over the limit in Aus.

In Australia you can only drive faster than the limit if there are no cars on the horizon or you are sure none of the adjacent lanes have any unmarked vehicles and you cant see any speed cameras on the side of the road. You cant compare that at all in the UK where you can overtake a policecar at 140 and noone cares.

I agree with the speed limit thing though - AP should reduce the speed limit as soon as it can detect the limit. With regen braking it is already capable as evident by it recognizing the sign maybe 30m in front of you. and yes you have to go with the lowest limit because if the GPS says 80km/h but the AP detects a 30km/h sign its probably because of a building site etc.

I actually think the speed limit detection works really well in Australia. The only issue I have is that if you set it to 10km/h over which it keeps when the limit is i.e. 50km/h where it isnt appropriate really. If you manually reduce it it wont put the speedlimit back up when you leave the reduced speed section.
 
in the UK where you can overtake a policecar at 140 and noone cares

Hmm..... UK as in Ukraine maybe? Not my UK! They care if they are on duty.

The second case was on unfamiliar roads with poor lighting, just after a roundabout.

..... and as it happens, near Dover!

yes you have to go with the lowest limit

I agree. Which is a problem for dealing with signs on sideroads that get read as though they apply to the main road. Can you use the GPS speed for the main road to cancel the irrelevent side road speed limit?
 
Last edited:
  • Funny
Reactions: mikes_fsd
What good is "autopilot" if you have to keep your hands on the wheel? It's pointless. You might as well steer the car yourself.
It's a substantial workload reduction. The car handles maintaining speed,, lane centering,, and not running into the car ahead of you. Eight eyes that don't get drowsy or distracted to help keep you safe. Why would you turn down a helping hand?
 
Again, I must say - more than a year after my original post, AP is still unimproved, and now other carmakers have caught up with Tesla even more.

(I am obviously not counting the ""FSD"" beta that very very very few people have access to - also it's obvious even from the videos that are released that that system is very far from being releasable)

I was watching a Doug DeMuro video - he's not a Tesla fanboy but has made some very fair and realistic assessments of Teslas in the past. He said in his video about the 2021 Mercedes-Benz E class that the driver assist system was every bit as good as Tesla's Autopilot and I think he even said that Tesla no longer had an advantage at all when it comes to this.... so, yeah
 
  • Like
Reactions: Tim42069
Every part of Navigate on Autopilot is worthless

I commute about 20 miles each way every day, on a route that involves all kinds of roads, from small to interstate. Every feature of NoA is not only worthless on all of those roads, it is fairly detrimental to the driving experience

* Speed based lane changes: (Part 1) NoA takes into account literally nothing but the very next cars, and often makes unforgivable navigational errors, such as putting you in a lane that will merge into yours (literally the arrows are there) in just a couple hundred feet, getting you in/out of lanes that don't make sense because you'll have to get out of them in a minute because of which lane you have to be in to reach your destination.

(Part 2) Even without any of those, it still makes bad choices (unless the situation is as obvious as being stuck behind a slow car in an otherwise empty interstate): It hogs the left lane (no idea why they got rid of that feature that made it move back) all the time, unless on "mild" it swerves from lane to lane to lane - and on "mild" it allows you to be stuck behind cars that are FAR too slow.

(Part 2) by itself could MAYBE (really doubtfully) be considered an improvement to regular autopilot, but it comes with (Part 1) and makes everything far worse.

* Navigation based lane changes: On the roads I am on, this is also a detriment. I don't know what input they're basing these decisions on, but it is just wrong so much of the time.

It often has the wrong idea on what lane it should be on eventually. At times, it will not start trying to merge until 0.7-0.8 miles left on a crowded road, making for some intensely anxious moments. At times, it will change lanes 3 miles in advance. I know both of those are results of poor map information, because esp. on 2 lane roads, it will randomly think it needs to get into another lane to follow the route, even though the lane we're on is completely fine. This last thing especially happens 2-3 times per drive, and most of the time I have to manually cancel it.

* Taking exits: "Taking exits" is literally just making 1 lane change and immediately declaring NOA is now off while keeping AP on - 99% of the time you have to immediately take over because (a) you have to make a turn quickly, or (b) it's on a butterfly exit and it's being extremely slow. So now that AP can "take exits" we can keep AP on for 2 seconds longer than we had to before. Yay.

In general, I've realized that NoA just makes me really anxious while driving because I don't know if it will make the right
decision. Often it doesn't, but even when it does I don't feel good about it because I have been anxious about it. I have turned NoA off.

===

Autopilot is essentially unimproved from 2016-2017.

I first got an AP1 Tesla on February 2017. It centered itself perfectly on the lane and adjusted its speed well, and could read speed limit signs well 99% of the time. I know there was the whole debacle with MobilEye and AP had to get a lot worse before it could get better, but the lack of functional, usable difference in Tesla Autopilot between February 2017 and March 2020 is really very little. So little that if you had told me in Feb 2017, I would in no way believe you.

Today what I use Autopilot for is still 99.99% the same as my AP1 Tesla that was built in 2015. As I outlined above NoA takes away more than it brings to the table, so I don't use it. There are a few improvements (blind spot monitoring is much better while still not as good as my old Porsche Panamera from 2012, and the car now stops more reliably for stopped vehicles ahead, which is a very important improvement, albeit it's still uncomfortably late most times) but some stuff that's still inexplicably missing - I think AP2.5 still can't read traffic signs?

I'm not even mentioning gimmicks like smart summon. I love that I have it and I can shock people but I don't use it.

===

Everybody has caught up.

The lane driving assist-related features on even Hyundais are very good now and while they might not be as good as Autopilot yet - remember that years ago when Autopilot was first a thing, none of those companies had anything that even remotely compared. Now, Autopilot is essentially functionally unimproved, but all of those companies now have rivaling technologies that have caught up or are getting really close.

(I'm on a HW2.5 MCU2 Model S.)


Agreed...it seems every update made some weird changes to NOA. The last great update for me was just before the Christmas update. NOA actually took the appropriate exits and off ramps. Nowadays...I have to keep an eye out...like watching out for a new learning driver.

What I still don't get is why NOA would disengage during bad weather... I mean if this automation is supposed to perform better than a human driver? It's so far from coming even close to my driving skills, and many would consider me a granny driver. I take slow lanes and rarely hit 5mph above speed limit. I would think at worst case, it would warn the driver that bad weather is hindering efforts and it will execute appropriate safe procedures to pull over out of harms way...or something like that. I mean what if I suffered a medical condition while driving? NOA and AP would just check out? That would be terrible.
 
  • Disagree
Reactions: mikes_fsd
Again, I must say - more than a year after my original post, AP is still unimproved, and now other carmakers have caught up with Tesla even more.

I agree completely. It's clear that they have been 100% focused on the FSD Beta code branch. Which means we haven't seen any meaningful feature updates since 'traffic light control' (which requires confirmation and is thus borderline useless).

They have the ability to turn on/off features on the FSD Beta code branch (as demonstrated by GreenTheOnly). I do not understand why they don't move all FSD owners to this code base with most of the features disabled. For example, they could disable left/right turns, lane changes to follow the route, and maybe even driving on unmarked roads. At least that way we would get the many neural net improvements AND we'd get traffic light/stop sign control without confirmation.

They could then enable another feature later - say driving on unmarked roads. And then lane changes to follow the route. Then right turns. And finally left turns.

Instead, I get the feeling that they're trying to get 'feature complete' FSD working on the beta branch before the production build gets ANYTHING new. This was understandable when FSD Beta was first released 7 months ago, but unless a wide release is really around the corner they need to push some meaningful improvements to those who paid big bucks for FSD years ago.

EDIT: This is the Twitter thread where Green talks about the feature flags in FSD Beta:
 
Last edited:
I agree completely. It's clear that they have been 100% focused on the FSD Beta code branch. Which means we haven't seen any meaningful feature updates since 'traffic light control' (which requires confirmation and is thus borderline useless).

They have the ability to turn on/off features on the FSD Beta code branch (as demonstrated by GreenTheOnly). I do not understand why they don't move all FSD owners to this code base with most of the features disabled. For example, they could disable left/right turns, lane changes to follow the route, and maybe even driving on unmarked roads. At least that way we would get the many neural net improvements AND we'd get traffic light/stop sign control without confirmation.

They could then enable another feature later - say driving on unmarked roads. And then lane changes to follow the route. Then right turns. And finally left turns.

Instead, I get the feeling that they're trying to get 'feature complete' FSD working on the beta branch before the production build gets ANYTHING new. This was understandable when FSD Beta was first released 7 months ago, but unless a wide release is really around the corner they need to push some meaningful improvements to those who paid big bucks for FSD years ago.

EDIT: This is the Twitter thread where Green talks about the feature flags in FSD Beta:
I think one explanation for why they don't move FSD owners to the Beta branch with the additional features disabled might be that the system is 100% different from non-beta, and can't perform up to par (or, as reliably as non-beta does).