Hello,

I'm a 34 yo programmer/entrepreneur in Romania, with a long time interest in rationality - long before I called it by that name. I think the earliest name I had for it was "wisdom", and a desire to find a consistent, repeatable way to obtain it. Must admit at that time I didn't imagine it was going to be so complicated.

Spent some of my 20s believing I already know everything, and then I made a decision that in retrospect was the best I ever made: never to look at the price when I buy a book, but only at the likelihood of finishing it. Which is something I strongly recommend even (or especially) to cash-starved students. The first one happened to be Nassim Taleb's Black Swan, which was another huge stroke of luck. Not only it exposed me to some pretty revolutionary concepts and destroyed my illusions of omniscience, but he's a frequent name dropper and provided a lot of leads for future reading material. And the rest, as they say, is history.

Introduction aside, I'm a long time lurker and I actually came here with a request for comments. There is an often mentioned thought experiment in the sequences that compares a lot of harm done to a person (like torture) with minimum harm done to a lot of people, like a mote in the eye of a billion people. I've always found it a bit disturbing, but couldn't escape the conclusion that harm is additive and comparable. Except I now think it's not.

I've recently read Anti-fragile and found the concept of "hormesis", i.e. small harm done to a complex system generates an over-compensatory response, resulting in overall improvement. Simple examples: cold showers or lifting weights. So small harm done to a lot of people is possible to overall have net positive effects.

Two holes I see in this argument: some harms like going to the gym create hormesis, while motes in the eye don't. Also, you could just up the harm: use a big enough mote that the overall effect is a net negative, like maybe cause some permanent damage. But both holes are plugged by the fact that complex systems will always find ways to compensate. Small cornea damage gets compensated at processing level, muscle damage turns into new muscle, neuron damage means rerouting etc. There are tipping points and limits, but they're still counter-intuitive. Killing the n-th neuron will put somebody in a wheelchair, but their happiness level still bounces back. There is harm, but it's very non-linear in respect to the original damage. So I can't help but conclude that harm is simply non-additive and non-comparable, at least not easily.