Two years is probably inaccurate, because they likely became ambitious and repositioned the X to be a major upgrade of S technology. Originally they were going to rush out an SUV because they were not confident on MS demand.
That is not saying I believe the model 3 release will go smoothly. I doubt they have the ability to maintain quality and scale production anywhere close to projections. My belief is due to the continuing fit and finish issue with the current cars. Getting current production correct is child's play compared to making 10,000 M3 per week with fairly good quality.
"Run the line faster" is about as likely as full self drive in the near future. I expect troubles before Tesla finally stabilizes with good production of the M3.
Preface: Please don't take this as an attack on your post, it was merely a convenient launching-point for things I've been mulling about for a bit in my mind.
Elon, is an engineer at heart, and that is both a blessing and a curse. The blessing side of it is he sees potential that others wouldn't be able to see. The downside is he might downplay the potential pitfalls. That said, he seems to learn from his mistakes quite well. I think that for Model 3, the learning experiences and pain from Model X left quite an impression on him. If you look at his obsession on the "machine that builds the machine" his current obsession isn't the car itself (which lead to Model X being too complex) but instead on the manufacturing process.
There are 3 major ways I can think of that a full speed production line can't scale well:
1. The car was not designed properly with the assembly process in mind.
2. The assembly line and its equipment were not designed to optimize production for the particular vehicle.
3. Automation is not fully utilized, making humans the weak link. (Often the biggest factor.)
On #1, Model X was and is the pain in Musk's side on what happens when you don't design a vehicle to be automated as much as possible. I suspect this will continue to be a pain point until the Model X receives a major redesign to resolve this. However, he learned his lesson for Model 3. The discussions regarding cutting wiring length in half and other items are easy to understand tidbits that he's throwing out to illustrate this.
On #2, the recent acquisition of
Grohmann Engineering should assist on this. If your assembly process is limited by the certain kind of robots you have, then you're limited to a specific process. However if you can modify and get exactly the kind of robots that you want for your process, you're not limited to a certain function set or sensor set that a particular robot has. That's why Tesla got these guys.
On #3, typically in automotive manufacturing, your humans are handling very detail oriented tasks, such as interior installation, fit and finish operations, etc. Musk wants to automate as much as possible, and by having the M3 be very sparse as a whole on the interior, we minimize human labor. In fact, I suspect that the Model 3 will use the least amount of human labor for assembly compared to any comparably priced car in the US market.
I suspect however, that at least in 2017, Musk's expectations on deliveries will be a bit of an overshoot. I still expect 30-40k M3 deliveries in 2017. Considering the shorts don't expect anything more than 3-5k in 2017, this will be a boost, but they'll try to downplay it as "Tesla didn't hit their 70k mark blah blah." I expect in 2018 for the floodgates to open, and they really will deliver 150-180k M3 deliveries.
Disclosures: Elon fanboy and TSLA long.