Infinite and 10 years from now is the same thing for a 2018 model 3
I think a better gauge would be 1/2 of the average length of ownership of a new vehicle. That's roughly three years since the average length of ownership of a new car is 71.4 months.
With a 2018 Model 3 (what I have) it's nearly there.
So lets take stock of what I bought versus what was delivered
EAP:
AP -> Meets minimal viability expectations. The biggest drawbacks are lack of smoothness in de-acceleration, and phantom braking. The current phantom braking annoyances seem to be maps related, but Tesla gives no mechanism to the customer to inform them of a map error.
NoA -> does not meet expectations - From a legal perspective it does do what it promises when it works, but it doesn't work well enough for me to be satisfied with my purchase.
Smart Summons -> Total and complete failure
Auto-Lane Change -> Meets expectations, but could be vastly improved if the positions of the cars around the vehicle were better especially with semi's. I have hope that Tesla Vision will fix this.
Cost: $5K
Estimated Delivered Value: $2.5K
Loss: 2.5K
FSD:
HW3 Upgrade -> Mine was delivered with HW2.5, and one of the promises was the HW3 upgrade. This met expectation.
Traffic light/sign Response -> It feels like its a preview release. I do enjoy it, but I find myself having to turn it off when the majority of my driving will be freeway due to too many false positives. My favorite part is the chime on green even though I'm not one of those people with an issue of not going when the light turns green. I just like the game of beating the chime.
FSD Beta -> As a customer this angered me because they selected 1K (or more) "special" customers to give it to, and then made totally false promises to give it to the rest of us.
Cost: $3K
Estimated Delivered value: $1.5K
So in summary I feel like Tesla delivered on value about 50% which isn't terrible when it comes to options on new vehicles. If I ignore the overpromise of FSD that was clearly beyond the price then the biggest issue is the lack of effort to make the current features better. Like we don't even have a proper reporting tool, and Elon is completely delusional about the current state of NoA.
If I look at it from a customer retention problem the solution is very simple. That solution is to simply allow FSD to be transferable to a new Tesla vehicle while also not allowing anyone to buy FSD anymore. So FSD becomes either something you're grandfathered into (in having a car with FSD or buying a used one where the owner opts to give it up) or something you pay a subscription for.
Basically they'd have to give FSD ownership to the paying customer which is the opposite of what they do now. What they currently do is they play silly games with FSD where they delete it from vehicles they auction (or sell to dealers), and by appearances they add it to used vehicles they sell (why so many of them have FSD). They also don't seem to give fair value for it on a trade-in.
Long term the biggest damage Tesla is doing is to themselves as they're not allowing customers to continuously upgrade to something better. Instead they've locked us into an old sensor suite that is so bad that Tesla deleted the radar with a shrug. I have a 2018 with zero upgrade mobility not because I can't money&desire wise, but because Tesla decided to retain legacy HW for considerably longer than my average car ownership. About the only thing FSD related that is better about the new model S is it has IR lights for the interior facing camera.
Heck that $3K I paid for FSD might save me $40K in sticking with the same car for the 10+ years it will take them to achieve any resemblance of self-driving.