I figure this will stir up some controversy here. Well worth the 10 minute watch.
My thoughts:
The astronaut simulation was clever, and it makes sense. In a large enough pool for a small enough number of positions (or outliers), the variance in top level talent or ability is going to be very small, and likely smaller than the precision with which we can measure or estimate it. In the astronaut example, a randomly distributed skill score from 0 to 100 could, on average, be expected to produce 180 perfect scores in a pool of 18,000 applicants for 11 slots. If you bumped the skill score precision to 0 to 1000 you could still expect 18 perfect candidates for 11 slots. Even still, there's going to be very little statistical significance between individuals at the top who barely miss perfect scores. So if we throw in some randomly assigned natural variance, or luck, that ends up being what separates out a number of those best of the best of the best individuals in real world progresses. It isn't necessarily individual merit, but a combination of merit and randomly distributed, beneficial noise that pages that tiny handful up to the top. Think of that extra 5% as the random distribution of happy coincidence and how that separates people out.
I also like his advice to ignore the variance of luck, as despite the role of luck, your actions and inactions still play a large part in success (you miss all the shots you don't take). As his own example shows, he got lucky with someone else featuring one of his videos, but his video would not have been featured if he never made it or didn't produce good content. However, producing good content wasn't enough for him to do YouTube full time until he got lucky with the feature. His success has been a combination of both.
My thoughts:
The astronaut simulation was clever, and it makes sense. In a large enough pool for a small enough number of positions (or outliers), the variance in top level talent or ability is going to be very small, and likely smaller than the precision with which we can measure or estimate it. In the astronaut example, a randomly distributed skill score from 0 to 100 could, on average, be expected to produce 180 perfect scores in a pool of 18,000 applicants for 11 slots. If you bumped the skill score precision to 0 to 1000 you could still expect 18 perfect candidates for 11 slots. Even still, there's going to be very little statistical significance between individuals at the top who barely miss perfect scores. So if we throw in some randomly assigned natural variance, or luck, that ends up being what separates out a number of those best of the best of the best individuals in real world progresses. It isn't necessarily individual merit, but a combination of merit and randomly distributed, beneficial noise that pages that tiny handful up to the top. Think of that extra 5% as the random distribution of happy coincidence and how that separates people out.
I also like his advice to ignore the variance of luck, as despite the role of luck, your actions and inactions still play a large part in success (you miss all the shots you don't take). As his own example shows, he got lucky with someone else featuring one of his videos, but his video would not have been featured if he never made it or didn't produce good content. However, producing good content wasn't enough for him to do YouTube full time until he got lucky with the feature. His success has been a combination of both.