Just read
Autopilot Retrofit on Classic P85 | wk057's SkieNET and I have a comment, not to wk057, but responding to wk057's last statement in his post:
wk057 said:
Now, I know the next big question: How much did it cost? Well, not counting my own labor, all together the project cost less than $9,000. That includes costs for parts from Tesla and salvage vehicles, along with getting the windshield installed, the rear bumper cover painted and re-wrapped yellow.
Again, don’t expect to go into Tesla and start demanding they retrofit autopilot. It’s not going to happen. It’s way too much work, and way too involved. Also, if Tesla were to do it they would be replacing basically every wiring harness/fuse box in the car vs. making some modifications like I did, which would be even more work and extremely expensive. I did this as a sort of proof-of-concept, and because I wanted my wife to have autopilot but didn’t want to trade her car in. It shows that it can in fact be done with substantial effort.
I'm not expecting Tesla to bend to my or anybody's demands, but theoretically, the simpler modifications wk057 did vs. the replacing the harnesses could be done in a retrofit; in my opinion, a Tesla-designed retrofit could be made to make a new set of parts that is purely optimized for retrofit, rather than getting a pure as-same parts for the AP native cars, and gain the parts advantages that wk057 designed into his retrofit for his wife's car. And finally, on top of this, Tesla could then go and specify these retrofits to a competitive bed of third parties that are willing to do the labor and facilities for this type of work, and then those third parties (businesses) could set their own prices for this work. So, ballparking making wild guesses, Tesla engineering on this would have some cost, so that would inflate the retrofit-specific parts a bit, so let's say the materials and outside help that wk057 paid $9,000 would actually come to, considering that some of that could be economies of scale for the professional retrofitter and some of that was already pre-inflated by Tesla and some of it would need more engineering priced in, let's say $12,000, but let's assume Tesla is a-hole and ups it to $13,000 just for the hell of it (like is their way). Then, the pro retrofitter has to make profit, has facilities, etc., so, all-told, $20,000. The next year, with volume and negotiating prices down at Tesla and after Tesla has recouped their engineering costs AND they see the value of their older cars being propped up which will cut into new sales but not really because that market segment still gets new stuff all the time and then the whole model line has better value to those people so it's a positive, Tesla is open to making the retrofit part prices less of a profit center, and with capital costs better paid, the professional retrofitters can drop price to around $12,000, eventually in a few more years to about $8,500 or so. That's my wild lack of information estimate. I think it's possible.
Do I think anybody will do it, soon? Not for Model S or X, but for Model 3 in its massive quantities? Maybe, if anything becomes a retrofit benefit ever, but Model 3 will come with a lot of stuff already in it. Probably yes, if something really innovative happens after a few years.
I think by the time these pre-AP cars get a few years older, it won't be worth it to retrofit them any more. I'm already seeing used pre-AP cars on the CPO site for not much more than double ($46K) what I estimate it would be to retrofit them ($20K), and the step up to a new car is the same as my estimated retrofit cost ($20K), so it seems pointless. Sad to say.
If I were Elon, I'd be tempted to make it available to retrofit the entire fleet, just for the point of it. But, I don't have all the labor costs and environmental costs from retrofitting them all staring at me in the face. Plus, that's a sunk cost, and doesn't benefit the goals of Tesla: he'd be using real profit money Tesla's currently getting from people upgrading to newer cars instead for the old car owners for actual materials and labor rather than investing into Model 3 engineering, and that's bad.
So, in effect, wk057 is right.