Hacker News new | ask | show | jobs
by Tycho 2859 days ago
NN Taleb’s books. He has said that books should surprise the reader every single page or else it’s not real writing.
4 comments

I read parts of a couple of Taleb’s books, Black Swan and Antifragile. There are a few original ideas in each and then they are reapplied again and again in different domains.

I often get how the applications go after reading the section titles, thus I think it can still be more concise. That said, the two books have much better signal-to-noise ratio than some high-brow journalistic pieces.

I found the black swan to be 80% fluff. Still a worthwhile reading, though.
Black Swan could have been a pretty good book at 15% of it's current size. I got tired of being told the same thing in at least 5-6 different ways every time and never finished it.
Could not tolerate. I read hundreds of nonfiction books, and this is one of the few I could not get through.
I found it worth the struggle in the end. One of the most influential books in my life, I’d probably put it as #3.
> NN Taleb’s books

Quite the opposite. He expands one single argument ad nauseum into each of his books.

It's exasperating.

which single argument?
Each book has a different one.
And they are?
> Black Swan

Fat tailed probability distributions are fat tailed...! Assuming normal distribution underestimates risky events.

https://en.wikipedia.org/wiki/Fat-tailed_distribution

> and Antifragile

Simple complex systems are too complex to be simple enough ...!

Okay, so this is a bit more technical, but the important thing is that unintelligent systems (= they are simple) that are complex (= have complex dynamics) are bad at adapting to different circumstances.

Also, these systems are products of design trade offs. (We want things to be cheap and get done fast, and also politically okay, and so on.) And thus they are not robust enough, nor resilient enough. They are fragile. (Taleb laments a lot about how antifragiltiy is different from robustness and resilience. And sure, they are because he uses a model in which they are.)

And the book talks about what is needed for anti-fragility, and it turns out that some kind of feedback loop that optimizes for certain problems. Or of course intelligence.

The core point you missed in antifragile is inference in the presence of uncertainty - a conservative barbell like strategy (80-20 rule or 90-10 rule). This is very easy to overlook but this to me was the key takeaway - in structuring a portfolio, picking art projects etc.
Thank you.

I couldn't have explained it better.

> Fooled by randomness.

Selection / survivorship bias

Taleb has often been accused of lacking rigor in his arguments. He does write well, though.