Here's the thing with those numbers.
First, I've been through this before many times in the debate between Windows and Mac. Numbers ranged from 3% to high-teens for many years and is certainly higher now. But, it didn't really matter that much in the long-run, as the market was plenty big enough for solutions to exist in either camp, and one or the other was better for particular people regardless of market-size. It also never reflected, very well, the numbers of units in actual use at any give time. The numbers were based off metrics that didn't capture all the systems, and also failed to downplay systems serving as cash-registers or systems that broke, or systems that weren't put into use, etc.
Second, as I indicated, there's that actual in-use problem. It doesn't matter much if Android has 81% in some form of distribution figures someone estimats, if they are only 30% in interactivity with the real-world. Where are all those units? People haven't figured out how to connect them to the Internet? They don't visit websites? They don't shop or buy apps? Are they sitting in shoeboxes while person uses their iPhone?
And, ultimately, as I indicated, so long as there are ENOUGH units in any market, market-share is mostly just a braging point. That goes both ways. When I say Android is a small share, it doesn't mean Android is bad or doesn't stand a chance... I'm just correcting the misperception that Android has the majority and thus, Apple must be in trouble again.
Regarding your example, I suppose I agree roughly, in that the more capabilities, the more things people will find uses for them. I'm not opposed to capabilities, but any manufactuer has to cut the list off at some point based on costs, how widely used they feel the feature will be, and most importantly, how well they can integrate and support that feature. All too often, I see features just tacked on for feature and spec sake. That isn't Apple.
And, maybe I'm a bit behind on my 3D these days, but I'm highly doubting your example is possible. First, a 3D camera on a phone isn't going to be able to generate an appropriate 3D model for such a task. Second, a 3D printer might be able to print a cool *model* of that taillight, but then they'd have to have some taillight manufacturer to make a real one out of appropriate materials.