Surrounded by smart people and intellectually curious people (graduate school is probably on the pareto boundary for this criterion)
Learn things about your field that would be impossible to find in books
Learn how to do research (in particular, how to pose and solve problems that take months to years and that no one has solved before)
Learning to do research involves, among other things, learning to reason about what approaches are likely to work and solve interesting problems, which is particularly difficult because you’ll be able to bring P(success) up from 1% to 30% but not all the way up to 100%. So the other side of this is being relentlessly critical of your own ideas, even after you’ve sunk weeks into them, so as to prune away the other 70% of the probability mass after a couple weeks of time investment instead of after a couple months / years of time investment.
I think the main benefits of getting a PhD are:
Surrounded by smart people and intellectually curious people (graduate school is probably on the pareto boundary for this criterion)
Learn things about your field that would be impossible to find in books
Learn how to do research (in particular, how to pose and solve problems that take months to years and that no one has solved before)
Learning to do research involves, among other things, learning to reason about what approaches are likely to work and solve interesting problems, which is particularly difficult because you’ll be able to bring P(success) up from 1% to 30% but not all the way up to 100%. So the other side of this is being relentlessly critical of your own ideas, even after you’ve sunk weeks into them, so as to prune away the other 70% of the probability mass after a couple weeks of time investment instead of after a couple months / years of time investment.