How often did you use the upgraded AP functions

[JulienW]

The car backs in when it autoparks. It won't, ever, need rear cross-traffic, because it won't, ever, be backing out of a spot it parked itself in. It'll be pulling out forward (Where it already has multiple cameras and forward radar)

Real world edge situations with millions of cars driving billions of miles they WILL need to back into potential cross traffic 1000's of times a day. There is 100% NO way around it. Can you look at your driving history and say you never have needed to back up into potential cross traffic?

Also pulling out if there is van (or other obstacle) on the side the B-pillar camera can't lean forward to see around/under/over it the way you do. It MUST pull out the front end far enough for the camera to see and this could expose it to traffic.

I still think it may need radar on the 4 corners.

Again, I hope I'm wrong and just playing devil's advocate. I did pay for FSD.

 

[raptor5244]

The car backs in when it autoparks. It won't, ever, need rear cross-traffic, because it won't, ever, be backing out of a spot it parked itself in. It'll be pulling out forward (Where it already has multiple cameras and forward radar)

What does the car do when you are not permitted to back into a space? For example, parking lots with angled parking spaces? I guess it can drive around in reverse searching for spots. Wait, it can't do that because evidently the rear camera does not work like the front cameras.

What about locations where it is against policy to back in such as some college campuses or paid parking locations where security requires the license plate to be visible?

I really hope the rear camera can be programmed in for more features. With 8 cameras including a wide angle rear camera on a system that is trained to "see" it should clearly be capable of identifying objects in the rear camera such people, bicycles, cars, trucks, etc. and it should be able to determine if they are moving. It should be Rear-Cross-Traffic-Alert on steroids. Otherwise, it sounds like we need a rear camera upgrade and program that sucker into the system. Much more useful than Beach Buggy 2.

 

[Knightshade]

What does the car do when you are not permitted to back into a space? For example, parking lots with angled parking spaces?

If the spot is sufficiently angled the traffic would only be coming one way, and you'd see it in the wide backup camera pretty easily.

It's only perpendicular parking where you have much reduced rear visibility to oncoming traffic where it's an issue, and backing in removes that problem (which is part of why that's how the car parks itself)

What about locations where it is against policy to back in such as some college campuses or paid parking locations where security requires the license plate to be visible?

Well, first, the majority of states require a license plate both front and rear- so that's an imaginary issue there.

I'd be curious to see who has such a written policy though? When I was in college here there wasn't such a thing, parking on campus required its own sticker visible on the window, not the plate.

I really hope the rear camera can be programmed in for more features. With 8 cameras including a wide angle rear camera on a system that is trained to "see" it should clearly be capable of identifying objects in the rear camera such people, bicycles, cars, trucks, etc. and it should be able to determine if they are moving. It should be Rear-Cross-Traffic-Alert on steroids.

Except that's physically impossible.

RCTA typically uses rear radars to see far enough to the sides- and the model 3 has no such hardware.

Real world edge situations with millions of cars driving billions of miles they WILL need to back into potential cross traffic 1000's of times a day. There is 100% NO way around it. Can you look at your driving history and say you never have needed to back up into potential cross traffic?

You realize cars were doing that just fine without RCTA for like....100+ years at this point...right?

But as I say, there's a reason the car parks itself backward- it's flat out safer to do so.

If you want to park more dangerously manually, that's your call of course.

 

[Eugr]

With 8 cameras including a wide angle rear camera on a system that is trained to "see" it should clearly be capable of identifying objects in the rear camera such people, bicycles, cars, trucks, etc. and it should be able to determine if they are moving. It should be Rear-Cross-Traffic-Alert on steroids. Otherwise, it sounds like we need a rear camera upgrade and program that sucker into the system. Much more useful than Beach Buggy 2.

The rear camera is not wide enough for a good RCTA, unfortunately. Attached is the example of the view from rear camera from my Mazda CX-9 and from my Model 3. As you can see, Mazda gives you much wider view of the road (and the road curves around the house)... And it also has a real RCTA with radars...

While we are at it, I still don't know why Tesla can't calibrate rear camera guidelines properly. They don't exactly match the tire track and they don't match external dimensions as well. With my Mazda they match external dimensions including the mirrors, so if the lines fit, I can be certain the car will fit as well (that works for front and birds-eye view too).

 

[Eugr]

I have FSD so could be wrong about standard Autopilot but you can just cut the blinker on and it will change lanes in AutoPilot. No need for Navigate on AutoPilot to be on.

Nope, you now need to buy FSD for it. The standard AP has only TACC and Autosteer, but not Lane Change Assist... It is one feature I would love to have, as re-engaging Autosteer after a lane change is annoying. But I don't feel it worth $6K.

I don't see myself using Summon and autopark often. NoA can be useful on long trips, but when I tested it during test drive it was more stressful then just driving myself as it didn't do a great job in SoCal traffic - it hesitated to make a lane change and would miss my exit if I didn't take over.

Granted, that was with version 2019.15.xx, so it is probably already better and should improve in the future, but I'm still not sure it worth extra $6K for me at this point. Plus, with Tesla you never know if there will be another sale at some point, for instance end of Q3? So I think I'll wait.

 

[Kognos]

Nope, you now need to buy FSD for it. The standard AP has only TACC and Autosteer, but not Lane Change Assist... It is one feature I would love to have, as re-engaging Autosteer after a lane change is annoying. But I don't feel it worth $6K.

I don't see myself using Summon and autopark often. NoA can be useful on long trips, but when I tested it during test drive it was more stressful then just driving myself as it didn't do a great job in SoCal traffic - it hesitated to make a lane change and would miss my exit if I didn't take over.

Granted, that was with version 2019.15.xx, so it is probably already better and should improve in the future, but I'm still not sure it worth extra $6K for me at this point. Plus, with Tesla you never know if there will be another sale at some point, for instance end of Q3? So I think I'll wait.

I'll confirm the Lane Change Assist is tied to "FSD" (as it exists today) and not part of "Autopilot" (again, as it exists today).
Autopilot is purely lane keeping, accelerate, braking, all traffic aware.

Otherwise, also agree entirely with the sentiments here - these are exactly my opinions as well. I'd like lane change assist, but it's not worth 6k to change lanes automatically, or to navigate highway ramps. The rest of the features of FSD are not useful for me, or not fully ready.

 

[raptor5244]

The rear camera is not wide enough for a good RCTA, unfortunately. Attached is the example of the view from rear camera from my Mazda CX-9 and from my Model 3. As you can see, Mazda gives you much wider view of the road (and the road curves around the house)... And it also has a real RCTA with radars...

While we are at it, I still don't know why Tesla can't calibrate rear camera guidelines properly. They don't exactly match the tire track and they don't match external dimensions as well. With my Mazda they match external dimensions including the mirrors, so if the lines fit, I can be certain the car will fit as well (that works for front and birds-eye view too).

Good info. This is definitely one area that caught me by surprise. It looks like all the effort went into the front and side cameras but the rear camera lacks a super wide angle and the rear traffic alert capability. RCTA is actually one of the more useful safety features available today and it doesn’t exist on this car with 8 cameras and sensors all over the place. I figured maybe Tesla had another solution by using “vision” based recognition or something but based on the info it looks like the camera make lack proper field of view.

 

[raptor5244]

I'll confirm the Lane Change Assist is tied to "FSD" (as it exists today) and not part of "Autopilot" (again, as it exists today).
Autopilot is purely lane keeping, accelerate, braking, all traffic aware.

Otherwise, also agree entirely with the sentiments here - these are exactly my opinions as well. I'd like lane change assist, but it's not worth 6k to change lanes automatically, or to navigate highway ramps. The rest of the features of FSD are not useful for me, or not fully ready.

Yeah $6k is a lot of scratch for it to change lanes for you. I am in the wait and see camp when it comes to FSD. If Tesla truly does deliver FSD how much is it worth? I think the answer will be different each individual, which is why I think they should offer the features a la carte. For example, this Auto Lane Feature is probably the most desirable feature in the FSD package right now. So, why not make it available for purchase from our Tesla accounts? They could charge a one time fee or even better let us subscribe to the feature and use it as a recurring revenue stream.

Just an example:

- Auto lane change (NoA) $20 a month
- Summon $5 a month
- Autopark $5 a month
- Premium Connectivity $8 a month

 

[JulienW]

....

Just an example:

- Auto lane change (NoA) $20 a month
- Summon $5 a month
- Autopark $5 a month
- Premium Connectivity $8 a month

Bad math for Tesla (and FSD owners). Let's say even if >75% subscribe to at least ½ the features that is only about $250 a year per car. This would be FAR less revenue than Tesla selling the current percent (25%-40%????) of FSD packages at $6K.

*Pulling numbers out of my rear but I believe my premise is correct.

 

[raptor5244]

Bad math for Tesla (and FSD owners). Let's say even if >75% subscribe to at least ½ the features that is only about $250 a year per car. This would be FAR less revenue than Tesla selling the current percent (25%-40%????) of FSD packages at $6K.

*Pulling numbers out of my rear but I believe my premise is correct.

I know, I kind of listed what I would be willing to pay at this point. I was also thinking at lower subscription price you would get a higher take rate and fairly steady revenue stream and even it out.

 

[Trips]

If you did not get in on the deal a few months ago I would wait.

If the features were separate, here is how much I would pay right now.

• Navigate on Autopilot ($250)
  
• Auto Lane Change ($500)
  
• Autopark ($0)
  
• Summon ($1,000)
  
• FSD computer ($0)

Coming later this year:

• Advanced Summon ($1,500 more than Summon)

• Recognize traffic lights and stop signs ($1,000)

• Automatic driving on city streets ($5,000)

 

[Kognos]

I'm not so sure that breaking apart features is really the right way to do this. However, and I somewhat hope Tesla is listening, there is a value in selling something that can actually be used vs what is not. This is precisely why I won't buy FSD in its current incarnation - because most of it is not usable now.

I would seriously consider an adhoc upgrade for NoA or automatic lane changes though, depending on price. I just don't like the process currently for it (disengage AP, change lanes, re-engage AP) and it seems like I should just be able to use a turn signal and have it merge like it does for the FSD variant - after all, all the hardware is there.

 

[Knightshade]

I'm not so sure that breaking apart features is really the right way to do this. However, and I somewhat hope Tesla is listening, there is a value in selling something that can actually be used vs what is not. This is precisely why I won't buy FSD in its current incarnation - because most of it is not usable now.

Eh?

Lane change, auto park, basic summon, and NoA all work. Today.


What's funny is, they use to offer EAP- which came with basic AP plus all those features above, for 5k....and then FSD for another 3k, which added literally no features other than a promise (still not delivered) of future ones.

And people complained EAP was too expensive and they didn't really need/want "all those extra things" just wanted TACC or autosteer for cheaper.


So Tesla listened, and began to sell the new, much cheaper and more basic AP, and moved all the "extra" stuff into FSD,


And now folks are upset about that

 

[Kognos]

Eh?

Lane change, auto park, basic summon, and NoA all work. Today.

When I say that "most" of the FSD features don't work, I put heavy weight into city street driving, which is an awful lot of work.

You're right in that EAP having these features of NoA and lane changing. I just think they could/should add those on for people who choose to want them. I have a sneaky suspicion they will, in the same way that the premium connectivity is also going to be an option. Here's hoping!

 

[Knightshade]

When I say that "most" of the FSD features don't work, I put heavy weight into city street driving, which is an awful lot of work.

But those features explicitly aren't available under the current offering.

They're instead promised to be added for later this year.

All the current features work... (to varying degrees of work, since some folks use and love each, and others less so)

You're right in that EAP having these features of NoA and lane changing. I just think they could/should add those on for people who choose to want them.

They do. If you buy the FSD option.


As I said, they used to offer all that stuff WITH AP for a higher price- and people complained. So they moved them to FSD and dropped the price on AP (eventually making AP standard).

Now folks are complaining about THAT.. (because adding those back to AP would require a higher price, at which point you'd be right back to where things were 6 months ago before the change)

 

[JulienW]

... I just think they could/should add those on for people who choose to want them. I have a sneaky suspicion they will, in the same way that the premium connectivity is also going to be an option. Here's hoping!

I seriously doubt that standard Autopilot will get a SINGLE new feature added. Tesla WANTS you to buy FSD and not really looking for an insignificant amount of money from a small percentage of owners that would just cause disappointment and confusion to the others. I would bet (and Musk has sorta said) that once they get Automatic Driving on City Streets and Red Light/Stop Sign/hopefully Speed Signs too you will see a price INCREASE for FSD (probably $8 and when FSD is finalized at LEAST 10K) since we will be moving well into a game changer that will be in high demand.

 

[Kognos]

I seriously doubt that standard Autopilot will get a SINGLE new feature added. Tesla WANTS you to buy FSD and not really looking for an insignificant amount of money from a small percentage of owners that would just cause disappointment and confusion to the others. I would bet (and Musk has sorta said) that once they get Automatic Driving on City Streets and Red Light/Stop Sign/hopefully Speed Signs too you will see a price INCREASES since we will be moving into a game changer that will be in high demand.

I agree with you but hope you're wrong.

The problem is that Tesla wants you to spend money on something that isn't there yet, and people like me are frustrated with that concept. I agree with Elon's stance that by the time city FSD is "live" it might cost 15k (my own hypothetical number). I'm willing to wait and spend more at that point, because it'll be done/baked/released/vetted/approved. If it sucks, I don't have to buy it. But I believe 6k is unreasonable for what I get right now.

The point is that people don't like buying promises, for the most part.

The counterpoint is that Tesla does have something to sell right now that actually works, and is only software. NoA, lane changes, to be specific. That can be packaged and sold right now. And because people know it really works, or at least knows the limitations of the product, and it only takes software to deploy (assuming HW 2.5 I think) then why not?

I just think that FSD on city streets isn't going to happen soon, without HW3.0 and some serious improvements, so Tesla will look for easy cash revenue, and it's easy to deploy something that's only software. It puts current AP/FSD models back to the AP/EAP/FSD framework. Here's hoping.

 

[raptor5244]

Does anyone actually enjoy driving anymore? Don't get me wrong, the concept of FSD is cool and all but I am not sure I am willing to pay a whole lot for it until we reach a level of autonomy where I am chilling out in the back seat working on my laptop while my car drives me to work.

 

[bwilson4web]

Does anyone actually enjoy driving anymore? Don't get me wrong, the concept of FSD is cool and all but I am not sure I am willing to pay a whole lot for it until we reach a level of autonomy where I am chilling out in the back seat working on my laptop while my car drives me to work.

Worse, talking with my wife.

Bob Wilson

 

[JulienW]

Does anyone actually enjoy driving anymore? Don't get me wrong, the concept of FSD is cool and all but I am not sure I am willing to pay a whole lot for it until we reach a level of autonomy where I am chilling out in the back seat working on my laptop while my car drives me to work.

So you would rather be enjoying driving in THIS than chilling out? Also keep in mind that this enjoyment comes at the COST of the lives of almost 40,000 people a year in the US alone. If everyone switched to FSD today that figure would probably be cut by at least ½ and would be very low in just a few years.