Fragile, robust, antifragile. The fragile breaks under variation, the robust withstands it, the antifragile gains from it. On that distinction Taleb builds a disorderly, repetitive and at times insufferable book, out of which come three or four tools worth a great deal.
What I did with it
The barbell sorted out how I allocate time. Most of it goes to dull, known, paid work; a small slice goes to bets where the most I can lose is that slice and the upside has no ceiling. What disappears is the middle: moderately uncertain projects with large effort and a low ceiling, the ones you accept because saying no feels rude.
Via negativa changed the order of the questions. Before adding anything — a tool, a service, a promise — I look at what can be removed. Depending on a single client, a single platform or a single route to work is fragility, and it is not fixed by working harder.
And skin in the game is how I decide whose advice to weigh: counsel from someone who pays no price for being wrong is worth less, and that includes the salesperson recommending the platform that pays their commission.
What I left out
The tone and a good deal of the content. Taleb spends pages insulting academics and journalists, restates the same idea under different Greek names, and mistakes being right about one thing for being right about everything. The core idea fits in thirty pages and the book runs to six hundred. I also leave out the easy reading that all uncertainty is good: for someone who lives on monthly invoices, most uncertainty is simply risk.