Which could also be explained by observing that people born into more fortunate socioeconomic circumstances, with all that entails (different value systems, better health, better connections, better early education, etc.), have both a higher probability of going to college and a higher probability of becoming billionaires. The two outcomes are correlated, but one is not necessarily causal of the other. Completing college could even marginally decrease someone's chances of becoming a billionaire, and you could still see the same correlation.
My reading of the parent post was that the commenter was implying a causation by suggesting the difference between the percentage of college attendees in the general population vs the percentage of college attendees in the population of billionaires was conclusive of something significant ("despite that...").
If readers didn't understand the parent and grandparent posts to imply something more significant than basic math, I doubt the posts would have received so many upvotes.
I'm curious though why you found it interesting. Without further data or analysis, the statements have no substantial meaning unless one conflates correlation with causation. You could just as easily say "people who go to college are more likely to have straight teeth."
I was merely pointing out that the article takes the exceptions (dropouts that became billionaires) as the rule, when in fact the opposite is true: dropouts don't become billionaires; graduates do.
The article says the probability of becoming a billionaire increases if you are a dropout because 20% of billionaires are dropouts. But in fact, because there are many more people who didnt complete college than who did, this is even more wrong than if you didn't take that into consideration(which it seems the article didn't).
Like pretty much every article in any type of popular publication, the statistical analysis in this article is wrong. Not "somewhat off", but "wrong". Like, if I was a stat teacher, I would put "F" at the top of the page. Just like I put "F" in my mind on every CNBC or WSJ headline that says "Market drops because of...".
My comment was very short because if I put in much time debunking improper statistics I'd run out of time fast.
see: http://norvig.com/experiment-design.html
Warning Sign I7: Confusing Correlation with Causation