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

So, is that it for AP1? Beta forever?

This site may earn commission on affiliate links.
I've had my AP1 Model S since early 2015 and enjoyed the constant improvements to AP1 until right around the time that AP2 came out, then it pretty much stopped improving. Not that I'm complaining... it's certainly more useful than cruise control in many highway situations, but I sort of expected that lane keeping would continue to improve (e.g. better deal with cars changing lanes up ahead, better position itself in the center of the lane, offset appropriately for semi trucks, properly handle construction barriers, tight curves and stopped traffic) and that it would automatically take highway exits by now.

More than a year has passed since we've received any discernible updates to AP1, due presumably to all the focus on AP2 and FSD. Are we to expect that that's it for AP1 after all the hubbub about fleet learning and constant improvements? It seems to me that if there was any truth to that discussion, we would have a more robust, feature-rich system by now.
 
I don’t think AP1 was ever promised to do much more than what it does now. I doubt it will improve much.

These AP1 features are still not fulfilled:

1) Limited Hands-Free from an On-Ramp to an Off-Ramp,
2) Automatic Highway Interchanges,
3) Summon: You call your car up from its garage and it will meet you at your front door or curb.
 
For some of the missing AP1 features, it may take the new navigation app before Tesla can safely implement the listed 3 features. The current nav app is using data that is at least 2 years old (in the US), and would not be able to reliably control AP1.

AP1 will always be limited due to the number of sensors and lack of serious processing power. Tesla will likely continue to make small improvements to AP1, but after the original goals are achieved, it should become relatively stable - like TACC, and probably not get anything other than bug fixes.
 
at least AP 1.0 was only a 1,500 option when I purchased. 5k for AP 2.0/2.5 is a joke (on us).
Agreed. That price of that singe item is one of the largest factors working against my potential choice of buying another new Tesla (S or 3) when my current lease is up. It would gall me to pay that much. Still not decided, though, until I get to drive a Model 3 and see how good it is, AP or not....
 
Can us owners of AP1 have any reasonable expectation of improvements to it when those improvements at best can only be incremental?

It's far more advantageous for Tesla to focus on AP2 improvements that use SW that they largely dictate. They spent about a year attempting to reach parity with AP1 on SW they control.

I think we have to accept that AP1 will still be better than AP2/AP2.5 in some ways, but we're largely locked into what we have. That's okay by me as the improvements I want have more to with navigation, and other things. Like the recent firmware upgrade that turns on the heated steering while if it was on when the car was used last.

I can't think of any low hanging fruit for AP1. In fact as soon as they announced AP2 I was hoping that Tesla would lock down AP1 as this was during a period we were losing more than we were gaining (as they were cracking down on abusers).

I wouldn't be envious of AP2 owners as they are going to be in the same boat soon. We're so far away from what it's actually going to take that we have no idea what the requirements are going to be.

Where what it takes is?
Can TACC slow down with 99.99 percentage accuracy for a stopped vehicle?
Can AP stay in the lane with 99.99 percentage on roads with really bad lines or no lines?
Does AP work in the rain on a reasonably poor day (not counting ridiculous storms)?
Does Side Monitoring actually work?
Can the rear camera sense the distance to the car behind it to make sure it can respond to a situation similar to how a driver might avoid being rear ended by not stopping too quickly.
Is the car capable of Level 4 or Level 5 autonomous driving?

Instead it's going to a constant "Are we there yet" question.

The simple answer with HW1/HW2/HW2.5 is a no. At the end of the day the sensors in HW2/HW2.5 don't allow for what I have listed. Sure they have more than what HW1 has. Sure that will translate into improvements, but only incremental.

Like I believe AP2 is better at identifying lines, and this is beneficial. But, it's not that's not enough to really motivate me to upgrade. If AP2/AP2.5 is capable of Level 3 driving that's a big enough differential that I would consider it.

I do expect to be somewhat jealous of AP2/AP2.5 owners for the next 6 months. Since I think they're finally going to start getting what they paid for with EAP.

So let them have their fun, and then we can have our fun with AP3.
 
Can us owners of AP1 have any reasonable expectation of improvements to it when those improvements at best can only be incremental?

It's far more advantageous for Tesla to focus on AP2 improvements that use SW that they largely dictate. They spent about a year attempting to reach parity with AP1 on SW they control.

I think we have to accept that AP1 will still be better than AP2/AP2.5 in some ways, but we're largely locked into what we have. That's okay by me as the improvements I want have more to with navigation, and other things. Like the recent firmware upgrade that turns on the heated steering while if it was on when the car was used last.

I can't think of any low hanging fruit for AP1. In fact as soon as they announced AP2 I was hoping that Tesla would lock down AP1 as this was during a period we were losing more than we were gaining (as they were cracking down on abusers).

I wouldn't be envious of AP2 owners as they are going to be in the same boat soon. We're so far away from what it's actually going to take that we have no idea what the requirements are going to be.

Where what it takes is?
Can TACC slow down with 99.99 percentage accuracy for a stopped vehicle?
Can AP stay in the lane with 99.99 percentage on roads with really bad lines or no lines?
Does AP work in the rain on a reasonably poor day (not counting ridiculous storms)?
Does Side Monitoring actually work?
Can the rear camera sense the distance to the car behind it to make sure it can respond to a situation similar to how a driver might avoid being rear ended by not stopping too quickly.
Is the car capable of Level 4 or Level 5 autonomous driving?

Instead it's going to a constant "Are we there yet" question.

The simple answer with HW1/HW2/HW2.5 is a no. At the end of the day the sensors in HW2/HW2.5 don't allow for what I have listed. Sure they have more than what HW1 has. Sure that will translate into improvements, but only incremental.

Like I believe AP2 is better at identifying lines, and this is beneficial. But, it's not that's not enough to really motivate me to upgrade. If AP2/AP2.5 is capable of Level 3 driving that's a big enough differential that I would consider it.

I do expect to be somewhat jealous of AP2/AP2.5 owners for the next 6 months. Since I think they're finally going to start getting what they paid for with EAP.

So let them have their fun, and then we can have our fun with AP3.
I know that having TACC adjust speed according to speed limit changes is very low hanging fruit. Just hold the lever toward you, and it works. I'm satisfied with AP1, but activating this would be nice.

It's slightly possible, I suppose, that some of the other items like stop sign/traffic light reaction may translate without too much effort once they get it worked out on AP2, but I would be extremely surprised.
 
The only advantage at this point I see with AP1 is that it doesn't nag you when driving in stop and go traffic and it is less aggressive when changing lanes. Both of these issues are easily fixed with a firmware upgrade. This has been my observation based on 1800+ miles of AutoPilot freeway driving.
 
I'd really like to see red light and stop sign recognition added. I don't want AP to actually use this info to stop for lights and stop signs, because I doubt it would be accurate enough, but triggering FCW beeps when the driver is about to blow through would be great. FCW doesn't need very high accuracy to be useful, as long as it doesn't have too many false positives.
 
I know that having TACC adjust speed according to speed limit changes is very low hanging fruit. Just hold the lever toward you, and it works. I'm satisfied with AP1, but activating this would be nice.

It's slightly possible, I suppose, that some of the other items like stop sign/traffic light reaction may translate without too much effort once they get it worked out on AP2, but I would be extremely surprised.

The problem with the speed limit changes is it's only as good as the speed limit recognition, and I've never seen it work that well.

There is a perfect test section on I5 around the border near Portland where the speed limit goes from 70 (before Vancouver, WA) to 60, and then to 50. From there it goes to 55, but then go back down because of some curves. Eventually it gets to 60, and then 65 on your way to Eugene.

Most of the time it gets the sign recognition correct, but occasionally it fumbles really bad. If there car responded to the changes it would fumble pretty horribly leaving one embarassed driver.
 
The problem with the speed limit changes is it's only as good as the speed limit recognition, and I've never seen it work that well.

There is a perfect test section on I5 around the border near Portland where the speed limit goes from 70 (before Vancouver, WA) to 60, and then to 50. From there it goes to 55, but then go back down because of some curves. Eventually it gets to 60, and then 65 on your way to Eugene.

Most of the time it gets the sign recognition correct, but occasionally it fumbles really bad. If there car responded to the changes it would fumble pretty horribly leaving one embarassed driver.
Really? With AP1? It is probably over 98% correct for me. Sometimes there are issues that I can't explain, but I'd be glad to run with it. The problems I see are things like, in some areas, there are warning signs that the speed limit is about to decrease -- the car sometimes reads that sign as a speed limit sign. But it seems to only do SOME of those signs, which makes me wonder if it is in the database rather than reading the sign.
W3-5.jpg
 
  • Informative
Reactions: Galve2000
Really? With AP1? It is probably over 98% correct for me. Sometimes there are issues that I can't explain, but I'd be glad to run with it. The problems I see are things like, in some areas, there are warning signs that the speed limit is about to decrease -- the car sometimes reads that sign as a speed limit sign. But it seems to only do SOME of those signs, which makes me wonder if it is in the database rather than reading the sign.
W3-5.jpg

In Washington the detection accuracy is pretty good, but in Oregon it's considerably less than 98%.