It makes me think of people who have huge impact and success in life, with little obvious explanation. People like Bill Gates, Larry Ellison, Steve Jobs, Barack Obama, Donald Trump, etc. People to whom a lot of success is attributed, but it’s hard to say exactly, specifically, what skill or task they did to get it.
There’s a joke that Steve Jobs “invented the iPhone,” which is funny to people who are familiar with how products like the iPhone are actually created. But on the other hand… Steve Jobs definitely did something that was important to the creation of that product. Maybe it’s enough to say it was a poorly defined problem, which is why it’s also hard to define exactly what he did to solve it.
I also think intelligence itself is a poorly-defined problem, and AI will help us define it. I think this essay leans in that direction by recognizing the distinction between predictive intelligence (which AI is good at), vs a less-easy-to-define mental facility that defies prediction. Or maybe precedes prediction. Like if I want tacos for dinner, I can use my intelligence to navigate the problems necessary to get tacos. But can I reliably predict what I’ll want for dinner? Seems a lot harder.
What people want, vs what they do to get it, are probably a distinction similar to poorly-defined problems and well-defined problems, respectively. If you can figure out what people really want, well, that seems like a huge step toward being successful. But hard to define.