Taleb and Kahneman on Antifragile, at the New York Public Library

I came to this expecting a launch interview and got something rarer: two people who plainly like each other, agreeing on most of what matters, and being exact about the one thing they see differently. Kahneman says on stage that Taleb changed the way he thinks about the world, that the black swan idea is profoundly important and uniquely original, and that he himself pushed Taleb into writing Antifragile by inventing its subtitle before the book existed. On bankers, on skin in the game, on the arrogance of prediction, he says he will give Taleb no argument, and he means it.

The disagreement is one point, and it is a good one. Antifragility asks you to absorb a long series of small losses in exchange for a rare large win. Kahneman's own work, prospect theory, says people feel a small loss far more than an equal gain, so the strategy asks for exactly the pattern we are built to refuse. In his words, we are designed against what you want. He adds, gently, that Taleb can afford to live that way and most people cannot.

My position is that both are right and they are answering different questions. Kahneman is describing how we are wired. Taleb is describing which strategy wins. If the better strategy is the one your wiring fights, then the honest conclusion is not that it is wrong for humans, it is that it costs you the discomfort of working against your own conditioning, every day, for a long time, and most people will not pay that price. I have paid some of it, and I recognise Kahneman's description of the feeling from the inside. I still think Taleb has the better answer about what to do with it, and I suspect Kahneman would not disagree with that either. What he disputes is whether people will.

The opening, on the Lindy effect. Taleb starts by pointing out that Kahneman's ideas are forty years old and his own are new, so by his own rule Kahneman's book is the safer bet. A good-humoured way into a real point: for ideas, age is evidence of durability, and the talk is an early statement of what became the Lindy effect.

The definition, told through an options trader. Fragile things are harmed by volatility, robust things do not care, and there was no word for things that gain from it. Taleb's example is the package: fragile means handle with care, robust means nothing written on it, and antifragile means please mishandle. The tourist is fragile and the adventurer is antifragile, because a late bus ruins one and hands the other an opportunity.

Measuring fragility without predicting anything. The best technical idea in the hour. You cannot predict what will break the coffee cup, but you know it breaks before the table. Fragility is acceleration of harm: if a ten-kilo stone hurts you more than twice as much as a five-kilo one, you are fragile to stones, and a king who crushes his son with pebbles instead of a boulder has understood convexity. Size brings fragility for this reason, which is why an elephant breaks a leg and a mouse does not, and why Taleb prefers many small decentralised mistakes to one large centralised one.

The turkey, and Kahneman's objection. Taleb retells the turkey fed every day until the day before Thanksgiving. Kahneman's response is the pivot of the evening: the turkey has a good life until then, and that is what people want. This is where he lays out prospect theory against antifragility, and where Taleb answers with the transfer: the problem is not that bankers take small losses, it is that they keep the upside while taxpayers eat the downside. That exchange is the seed of skin in the game and the clearest short statement of moral hazard I have heard from him. The psychology underneath Kahneman's side is loss aversion.

Planes and banks. A system that converts every mistake into a safer next attempt against one where each failure raises the odds of the next. Commercial aviation had one fatal crash in the years he was counting because each crash made the next flight safer; a bank failure makes the next one more likely. This is the line I took from the hour, because it turns antifragility from a personality trait into a design choice.

Trial and error as trial with small error. To out-think trial and error, Taleb says, you would need a thousand IQ points. The gains in nearly every field came from people trying things with small downside and large upside, and the point of the barbell is to make the errors survivable. Kahneman, late in the evening, adds the honest counterweight: he is a firm believer in intuition, most of the time it serves us well, and he has spent a career on its amusing errors.

Project size and cost overruns. Taleb claims, from a colleague's data, that hundred-million-pound projects in the UK ran thirty percent worse than ten-million-pound ones. He does not source it on stage, and the entry on the planning fallacy is where the sourced version of that literature lives.