Thread by @davidboxenhorn

I think this is the problem that I'm having with @nntaleb's views on intelligence: IQ is well-defined, it's how well you do on an IQ test. But I haven't seen Nassim's definition of intelligence

In The Black Swan, @nntaleb carefully defines Mediocristan and Extremistan. So when he goes to war against Gauss, it's clear that he means that we shouldn't apply the metrics of Mediocristan to Extremistan

So, I know what IQ is, but I don't know what this other thing is, that @nntaleb is going to war against. In earlier tweets, I tried to make sense of Nassim by understanding that IQ is not a good measure of ability to avoid risk. This agrees with my intuition

I also made sense of @nntaleb by pointing out that IQ is a scalar that lacks dimensionality and can hide pitfalls - i.e. lack of some aspect of intelligence - that might be required for a particular task. This also agrees with my intuition

But when @nntaleb published his Medium paper, the example that he gave was that IQ doesn't predict wealth at the high end, but does predict lack of wealth at the low end. This too agrees with my intuition, and I think most people would agree

medium.com/incerto/iq-is-…

I don't think that most people would expect the richest people to have the highest IQs, or the people with the highest IQs to be rich. I think that most people know that something else, besides IQ, is required to become rich, or to have any real-world success

All of which agrees with @nntaleb's quip that IQ is a good predictor of "bureaucratic or academic success". Though, frankly, I'm not sure that it's even a good predictor of that. I think that bureaucratic and academic success depend hugely on social ability

I think that @nntaleb does a good job of characterizing IQ as "a form of intelligence, stripped of 2nd-order effects". This, too, agrees with my intuition, and I suspect most people would agree. But what is this other form of intelligence, the one that captures 2nd-order effects?

The fact that IQ is not a good predictor of many real-world results does not mean that it's a fraud. In fact, it is a good predictor of some real-world results:

1. It is a good predictor of your children's IQ. Parents and children have similar IQ

2. It is a good predictor of who you will marry. Husbands and wives have similar IQ

Who you marry and who your children are, are two of the most important things in most people's lives. And the fact that we marry people of similar IQ shows that IQ is something that human beings can perceive

In fact, in The Black Swan, @nntaleb creates and contrasts two characters who are clearly of high and average IQ: Nero Tulip and Fat Tony. Much of the book is about Nero trying to figure out with his high-IQ what Fat Tony, somehow, already knows

I must say, I identify with Nero Tulip because I have been trying to do much the same thing, all my life

Incidentally, all this was the thinking behind my previous tweet:

https://twitter.com/davidboxenhorn/status/1128540852240961537

Added: I just re-read @nntaleb's Medium paper. He added a lot since the last time I looked at it. I think that everything in this thread is in his paper, so I have no fundamental disagreement with him, *except* that he dismisses the things that IQ *does* measure, and I don't