Model 3 with AP 1 instead of problematic AP 2.0

[Drone Flyer]

With all the problems with AP 2 and the soon to be release of the Model 3 with no heads up display, could Tesla be rationalizing the costs and headaches of this car by going with a proven Autopilot and screen display?
Personally, I don't blame them.
AP 2 won't be fully functional anyway by the time the car hits the streets. I think it could be a good move.
HUD is too new for the masses as well.

 

[anonim1979]

From what universe did you came from?
The one where Tesla still works with Mobileye?
Do they make ICE cars too? Or at least hydrogen-cell ones?
/seriously please (I know a day in the office can be boring like hell...)

AP v2 by the time Model 3 will go into sales will be better than AP v1.

 

[dsvick]

No, they can't use AP1, the hardware on the car is all AP2 so the AP1 software won't work with it. The model 3 will have whatever level of AP functionality is available on the S or the X at the time it is released.

No HUD is simply because, as Elon explained, they don't think it's necessary. Whether anyone agrees with them or not, that is their position. Which doesn't mean that they can't add one later.

 

[Reciprocity]

AP2 is not problematic, it is largely user error. People think they should be driving on autopilot on every local road in the country just because it can. Autopilot is not meant for most roads and is unsafe to use it for that. On freeways, with a fair amount of traffic, it is nearly flawless. Full speed on freeways coming in days.

Tesla will never go back, they get so much more value out of the new hardware and sensor data which is required to make fully autonomy work.

 

[calisnow]

With all the problems with AP 2 and the soon to be release of the Model 3 with no heads up display, could Tesla be rationalizing the costs and headaches of this car by going with a proven Autopilot and screen display?
Personally, I don't blame them.
AP 2 won't be fully functional anyway by the time the car hits the streets. I think it could be a good move.
HUD is too new for the masses as well.

 

[lunitiks]

With all the problems with AP 2

Yes ...

and the soon to be release of the Model 3 with no heads up display,

... mhmmm ...

could Tesla be rationalizing the costs and headaches of this car

Rationalizing the costs and headaches of this car?

by going with a proven Autopilot and screen display?

Say what?

Personally, I don't blame them.

Who?

AP 2 won't be fully functional anyway by the time the car hits the streets. I think it could be a good move.

HUD is too new for the masses as well.

 

[Kant.Ing]

It sure will be based on AP2, but with more powerful CPUs. We need 160x to handle all 8 cameras!

 

[M0DEL³]

With all the problems with AP 2 and the soon to be release of the Model 3 with no heads up display, could Tesla be rationalizing the costs and headaches of this car by going with a proven Autopilot and screen display?
Personally, I don't blame them.
AP 2 won't be fully functional anyway by the time the car hits the streets. I think it could be a good move.
HUD is too new for the masses as well.

Sorry, but even if they wanted to (Which I'm sure they do not), without an arrangement with Mobil Eye they do not have a legal path available for doing so. And HUD/NO HUD has nothing to do with AP2 vs AP1, totally unrelated issues.

 

[Big-T]

One thing that has always interested me, is that they made the decision to load up the 35k base car with all AP2 hardware, regardless of whether the consumer wants it (and will pay for it) or not. Isn't that a HUGE profit margin ???? hanging over Tesla's heads? I"m sure they have some thoughts on how many people will go with the base, how many will AP1 equivilant and how many with AP2 but if you're off by a few percent x400,000 that's a huge difference in profit margin.

Added to that as you get to more mainstream buyers or other countries their taste for AP may be different. Seems hard to nail down quarterly profits when you have up to 23% of MSRP tied up on something that is a sunk cost that may never get recouped.

I mean heck they want what a 25% profit margin? And one guy will make a huge margin for flippling the software switch while the other guy will put you into the red?

Probably a discussion for the investment thread, but reading on whether people would get AP got me thinking.

 

[martinicus]

One thing that has always interested me, is that they made the decision to load up the 35k base car with all AP2 hardware, regardless of whether the consumer wants it (and will pay for it) or not. Isn't that a HUGE profit margin ???? hanging over Tesla's heads? I"m sure they have some thoughts on how many people will go with the base, how many will AP1 equivilant and how many with AP2 but if you're off by a few percent x400,000 that's a huge difference in profit margin.

Added to that as you get to more mainstream buyers or other countries their taste for AP may be different. Seems hard to nail down quarterly profits when you have up to 23% of MSRP tied up on something that is a sunk cost that may never get recouped.

I mean heck they want what a 25% profit margin? And one guy will make a huge margin for flippling the software switch while the other guy will put you into the red?

Probably a discussion for the investment thread, but reading on whether people would get AP got me thinking.

No. You're not accounting for 1) The production cost of having various configurations with or without AP 2) The value of having a huge fleet of data collectors feeding into their neural network billions of miles of data, an asset that no other automaker with have anytime soon 3) The relatively tiny cost of AP hardware (compared to the huge cost of software development) 4) A very high percentage chance that AP will be activated over the lifetime of the car

 

[Reciprocity]

One thing that has always interested me, is that they made the decision to load up the 35k base car with all AP2 hardware, regardless of whether the consumer wants it (and will pay for it) or not. Isn't that a HUGE profit margin ????

Mobile Eye just sold to Intel for $15 billion. That is why AP2 hardware will be in every 3 from day one. The data collected passively, even with AP2 not enabled as part of the purchase, is excessively valuable and really a pre-requisite for Autonomous driving to work. That data will enable high definition, constantly updated maps. I think people over estimate the cost of AP2 hardware. The GPUs are fairly expensive, but everything else is very cheap. Tesla is also buying those GPUs in bulk and there might always be some data sharing between nVidia and Tesla for beneficial pricing.

 

[chillaban]

One thing that has always interested me, is that they made the decision to load up the 35k base car with all AP2 hardware, regardless of whether the consumer wants it (and will pay for it) or not. Isn't that a HUGE profit margin ???? hanging over Tesla's heads? I"m sure they have some thoughts on how many people will go with the base, how many will AP1 equivilant and how many with AP2 but if you're off by a few percent x400,000 that's a huge difference in profit margin.

Added to that as you get to more mainstream buyers or other countries their taste for AP may be different. Seems hard to nail down quarterly profits when you have up to 23% of MSRP tied up on something that is a sunk cost that may never get recouped.

I mean heck they want what a 25% profit margin? And one guy will make a huge margin for flippling the software switch while the other guy will put you into the red?

Probably a discussion for the investment thread, but reading on whether people would get AP got me thinking.

8 webcams and approximately an iPad's processing power is not a lot of hardware to put in. The BOM cost probably is less than $500 for those parts at worst, and is great for upsell and essential for fleet-learning. I think it's a far better idea to put the sensors rather than designing an alternate cost cut version of the car chassis without the camera cutouts.

 

[Big-T]

Ok ok I hear you all ty. I have to say it's weird to hear that it's not expensive to put in full hardware for AP2 but in another breath to hear it's to expensive to put in a HUD (or any sort of 2nd screen, binnacle, gauges, etc) into the model 3 because it's 'just' a 35k car.

Don't want to hijack this into another HUD thread, it's just weird to hear that.

 

[182RG]

could Tesla be rationalizing the costs and headaches of this car by going with a proven Autopilot and screen display?

Mobileye divorced Tesla. Not an option.

 

[stopcrazypp]

Ok ok I hear you all ty. I have to say it's weird to hear that it's not expensive to put in full hardware for AP2 but in another breath to hear it's to expensive to put in a HUD (or any sort of 2nd screen, binnacle, gauges, etc) into the model 3 because it's 'just' a 35k car.

Don't want to hijack this into another HUD thread, it's just weird to hear that.

They have AP2 already, they don't have a HUD in any of their cars yet. Maybe if they introduce one to Model S/X it'll get to the 3 (I'm guessing that is how it will happen if Tesla plans a HUD).

However, Elon's stance seems to be extra displays (whether a HUD or second screen) are unnecessary as tech moves toward self driving cars, not the cost of the HUD specifically. For example, to use an extreme analogy, it doesn't cost much to add a CD player to the Model 3, but it's not worth the money to put it in.

 

[Waiting4M3]

Maybe you can contact Tesla and see if they'll sell you one of the prototypes from last year's reveal, those may have MobileEye's HW and AP1 in them.

 

[McHoffa]

and is great for upsell and essential for fleet-learning.

This is what it's all about.
1. The hardware included means that even if the original buyer doesn't want to pay to upgrade, a second or third owner might want to. I'd say a very VERY high percentage of them will have it enabled eventually.
2. Everything single Model S/X and # will be sending data to Tesla to help it learn, giving it HUGE head starts over every other automaker and autonomous system maker.

 

[martinicus]

With the Model 3 the tesla is making an affordable EV, as well as putting on the roads the largest self-driving data-collection fleet in history. You can't ignore that last part.

 

[sg021]

It does seem strange at first to think of AP as a cheap hw option, but as others have said the cost is much more on the software side.

An example is in the US you can add the parking assist package to a base Benz C300 for only $1090. You get 4 cameras and a top down view on your display showing everything around the car. You also get automated parallel parking. This feature is even standard when you move higher up the Benz lineup.

 

[calisnow]

8 webcams and approximately an iPad's processing power is not a lot of hardware to put in. The BOM cost probably is less than $500 for those parts at worst, and is great for upsell and essential for fleet-learning. I think it's a far better idea to put the sensors rather than designing an alternate cost cut version of the car chassis without the camera cutouts.

It's not an iPad - it runs eleven trillion operations per second.