He's more right than most people are on the Dunning-Kruger effect. What it actually tells us is that self-assessed competence estimates carry low signal: the co

He's more right than most people are on the Dunning-Kruger effect. What it actually tells us is that self-assessed competence estimates carry low signal: the correlation is almost zero. This might be a regular, pointwise regression to the mean, or it might be an artifact of the aggregation (i.e. it could be that some people are great self-estimators but others are way off the mark). What Dunning-Kruger tells us is that, with regard to certain abilities and especially social skills, people are bad guessers of their own competence and that most people think of themselves as slightly above average. I think that he's also somewhat right on money vs. happiness. Self-reported happiness is not the same thing as actual happiness. You might rate yourself a "7" at two times of life, but be behaviorally different. Your "7" when you are poor might be, "commute sucked, but no one yelled at me at work today"; but when you're rich, it might mean "hotel is nice, but staff forgot to fold the toilet paper into a triangle". Self-reported contentment is the same, but stress levels and behavioral measures of happiness are very different. It is true that after $75k per person, other factors have more of an effect on happiness than money itself. You have to account for cost of living and family size, of course. Then there are job quality issues. A writer who reliably makes $100k per year is a smashing success; a professor earning that, at 40, is more than respectable; but someone who's a "Software Engineer II", making that, at age 40 is a failure. I think that success is much more strongly correlated to happiness than income. People with Harvard MBAs making $250k per year are often the most miserable people on earth, while famous artists who make a fraction of that often love their lives. His relating this to type systems is a bit silly. You can't put all "static typing" languages in the same bucket. Java's static typing is a hindrance with little power. Haskell's static typing is extremely effective-- if you know how to use it. Also, quality of the engineers matters more than the language itself: I'd much rather work on a disciplined Clojure or Ruby team than a sloppy Scala team that's still using Java patterns. If it is properly used, a Haskell-like type system can make code quality very high indeed. That said, most businesses aren't willing to budget the time for quality code, and that's a language-independent problem.