Ultimately you didn't. That very comment shows you were able to recognize and correct for your mistake in the end.
Perhaps if you insist you've Dunning-Krugered yourself about Dunning-Kruger, then I may agree that you have in fact Dunning-Krugered yourself about Dunning-Kruger.
It's turtles all the way down.
Look at the graphs in the actual study:
http://psych.colorado.edu/~vanboven/teaching/p7536_heurbias/...
It's more like the range of 10-90% actual skill is simply transformed to 50-80% perceived skill. The absolute ranking of person by skill order is still accurate.
If Elon Musk is actually at about the 75% percentile of skill which maybe he is, if you turn the graphs that DK reported into a function of actual to perceived skill, if you fed 75% into that you would get like maybe 73%. The real vs. perceived skill curves cross at about 70%.
Versus the popular supposed version of DK, being something like "you are wrong and you do not know it." It's easy to describe the DK to sound like it's a profound explanation of that, but if you look at the data they reported, it's a much more timid effect than it sounds like it is.
To put it another way, low skilled people think they are better than most of the other low skilled people, but they themselves do not perceive themselves to be better than someone who does in fact know what they are doing. Low skill people would think they are a bit above average, not a superconfident expert.
I'm pretty sure there were core employees (not Musk) working on every part of the rocket from the beginning. This is documented in the Vance biography, IIRC.
Edit: I changed what I wrote, so as to not claim that "everyone" is a failure, some people worling in visual programming are still having fresh ideas. But they don't have the attitude of being "industry insiders" the ones with fresh ideas are the least arogant and offputting to outsiders. They are the ones who are least likely to put someone like Elon down as being inexperienced.