[pirgidb]

That's a story you tell yourself about the internal lives of others, but it is not knowledge. You don't actually know what's inside other people's minds. All we ever catch are glimpses of each other.

It's tempting to make sweeping generalizations about others to explain the ways they confound and frustrate us. But it's essential to hew to the truth and accept that life is ambiguous, people are baffling, and simplistic narratives do more to give us comfort and reinforce our biases than they do explain the world around us.

[ordinaryradical]

I’m not sure this claim has any more validity than his. Calling people unknowable flies in the face of a bunch of modern disciplines. We can see people behave very predictably with Game Theory, for example.

Still, I respect the view that the individual is private unto themselves in a profound way. But I would also say people tend to show you what they believe in how they act. If you pay attention you will begin to notice when someone has put their hopes in power, finances, achievement, or ideology, to give a few popular examples. And it’s not about what people say animates them, or even about what they believe about themselves—it’s how they act.

[pirgidb]

"People respond to incentives and we can model this with game theory" is a model that notably does not assume anything about the content of people's character or mind. And it frequently fails to explain the behavior of individuals, because you may not understand the incentive landscape they're in. You may offer to pay someone to do something, and it sucks the joy out of it for them, or they feel offended by your offer, and they become less likely to do it.

There's light-years of difference between observing an individual person's actions and drawing conclusions about what they believe (this is in line with what I was referring to with "catching glimpses") and generalizing to conclude that everyone is predestined to be either enthusiastic and lively or a miserable wretch. It's far more likely you caught someone on a good or bad day, to pick just one alternative hypothesis. People contain multitudes.

I also pay attention to the difference between what people say and what they do, and I agree it is meaningful and would even says it's a critical part of being around people. But it's easy to get carried away and pretend we know more than we really do, simplistic stories that explains everyone's behavior is very seductive. A model's ability to generalize is inversely proportional to how powerful it's predictions are. Stories like this that take people and bin them into two groups that control their destiny have huge predictive power - and apply to almost nobody.

Whereas to return to game theory, assuming people are black boxes that optimize some unknown and idiosyncratic utility function which we infer based on their revealed preferences can apply to nearly everyone and everything. But it's predictions are very narrow in scope. You can use an A/B test to optimize any change to any UI flow, and it'll tell you which one converts better. But it won't tell you why.