Good point and I absolutely agree on their tremendous overall accomplishment of building a new car company. You are right, it is easy to forget that. But I do hope that these niggling build quality issues are reduced with Model 3 production. Otherwise, there will be a whole lot more unhappy buyers, and they mostly won't be the more tolerant early adopters.
They are doing the right things to address the quality. The Model 3 is being built from the ground up to be as easy to mass produce as possible. Elon has said the Roadster was a mess to build, the Model S was easier, but still difficult to build, and the Model X is 98% about the same level of difficulty as the S with 2% that's much harder. The lessons learned from the X were the design engineers and production people weren't in close enough communication and a lot of things were designed that were just too difficult to manufacture.
The two are in close contact through the whole process with the Model 3.
The lesson of Tesla's history and that of any of the other smaller car companies is that making a car that passes all the highway safety tests is difficult, but doable, but to make a car that has consistently high quality is very tough.
Most Americans forget just how badly built most cars were 30-40 years ago. 50 years ago even the Japanese weren't really known for their high quality. Most of the Japanese auto makers were started in the 30s or just after WW II. It took them 20-30 years to figure out the secret sauce for making consistently high quality cars and part of that grew out of their home culture.
In the western world workers and management were fighting like two tom cats in a bag almost constantly. Management wanted things done a certain way and workers did it even if they could see it was stupid because management didn't want to hear about it. Factories were built on the old feudal model in most cases.
The Japanese listened much more to the people on the line and management took their advice. They also empowered any worker to stop the line at any time if they saw a problem.
The Japanese approach was to fix flaws when they were first spotted rather than keep the line moving and fix anything later. American factories could crank out huge numbers, but there were a lot of inefficient finishing lines that would fix flaws after production was done. In World War II keeping production quotas was so important there were airfields dedicated to rebuilding parts of aircraft after production because it was deemed better to not change the design on the line, but make the changes on planes already built.
The late models of the B-17 had a new tail gun assembly that was called the Cheyenne turret named after the rebuilding facility in Cheyenne, WY where the modifications were made. B-17s were built with the old tail gun assembly and a number of other things, then the planes were flown to Wyoming where brand new parts were torn out and replaced. The USAAF did this to thousands of B-17s.
One of Tesla's problems was Elon Musk came from the software world where there is no production like what you see in a factory. New software goes through testing, but once tested, you make as many identical copies as necessary and send them to the consumer or server where the customer will be served. There are no production quality problems, only design and coding quality problems and once fixed, every unit can be updated instantaneously.
Elon didn't really have an intuitive grasp of what happened when you have hundreds or thousands of people who all have to work together to make as identical a copy of the thing as possible, but each unit is going to be slightly different because you can't make the exact same physical object two times in a row.
The Japanese upped the game in precision manufacturing to levels nobody thought possible. Ford and Boeing both learned this in the 80s. Mitsubishi was making some parts for Boeing aircraft and another American contractor was making the same parts. They found the Mitsubishi parts fit perfectly, but the American made parts needed shims to fit right.
Ford learned a similar lesson when they had Mazda make transmissions for about half of Ranger trucks and a Ford plant made the same transmission for the other half. The Mazda transmissions never failed and the failure rate for the Ford transmissions was alarmingly high. Ford took apart transmissions from both plants and found that while every part in both were within tolerance, the Mazda parts were right down the middle in tolerance with variations that required sophisticated measuring devices to see. The Ford made parts were within tolerance, but they ranged from the extreme low to extreme high end.
Nobody in the US thought what the Japanese were doing was economically possible. It could be done if you wanted to spend a fortune per part, but couldn't be done on an assembly line. What the Japanese were doing was not one thing, but a more careful and thought out approach to the whole process. To a large degree American car companies have finally learned the same lessons, but it's taken decades and two bankruptcies to get there.
Tesla is starting from a more advanced point than say GM was circa 1985. The knowledge about how to mass produce high precision parts and assemblies are out there and a lot of schools teach the basics, but there is such a thing as institutional knowledge that isn't written down, but picked up by people on the job from those around them. That's why Toyota had no qualms about showing GM everything they did when they opened NUMMI. Toyota knew that even if the GM manufacturing people took careful notes, they weren't going to pick up Toyota's institutional knowledge until they had actually worked in it for a while.
GM tried to reproduce NUMMI at other plants cycling managers from NUMMI to other GM factories, but progress was so slow GM went bankrupt before it completely took hold. Gm would have been better suited to taking entire teams of workers from NUMMI and moving them to other plants where those workers could teach other workers on the line. The managers would understand what needed to be done but with the institutional hostility between workers and management, progress was glacial.
Tesla is still building that base of institutional knowledge. Some lessons can be brought in from industry people hired from other companies, but there are some lessons Tesla needs to learn on its own because while a Tesla looks like a regular car and does the same function as a regular car, it is fundamentally something different when you get down into the innards. There is overlap with other cars because cars need certain things to meet regulatory requirements, customer expectations, and basic human ergonomics.
But the drive train is a completely different animal from a regular car and Tesla has also incorporated electronics that is unique in the car industry. Because of these differences, the manufacturing techniques are going to be different. The GigaFactory is something that has never existed. It hasn't had to exist until now.
Tesla has gone from 1980s level quality to around year 2000 quality levels in 5 years. That's covering the ground other company's took 20 years to learn in 5. And by all accounts, the pace of learning has accelerated now that Elon is focusing more and more energy into how to build cars over how to design them.
The Model 3 will have some initial hiccups. ALL new car designs have bugs that get worked out in the first year or two, even Toyota and other top brands. (Toyota minimizes these flaws largely by keeping older designs in production longer than most companies.) I expect the number and severity will be a lot less than the Model S or X, but there will be some.
Elon has said that he plans to revolutionize car manufacturing to make it more automated than ever before. I have mixed feelings about this. Having precision robots make everything will probably improve overall quality and will reduce production costs. It will also probably bring more manufacturing back to the US. But it will mean a lot fewer jobs needed and that makes one of the developed world's problems much, much worse. What do you do when half the population is unemployed and unemployable?