The importance of tutoring, in its more narrow definition as in actively instructing someone, is tied to a phenomenon known as Bloom’s 2-sigma problem, after the educational psychologist Benjamin Bloom who in the 1980s claimed to have found that tutored students
. . . performed two standard deviations better than students who learn via conventional instructional methods—that is, “the average tutored student was above 98% of the students in the control class.”
Simply put, if you tailor your instruction to a single individual, you can make it fit so much better to their minds, so that the average person, if tutored, would become top two in a class of a hundred. The truth is a little bit more complicated than that (and I recommend Nintil’s systematic review of the research if you want to get into the weeds), but the effect is nevertheless real and big. Tutoring is a more reliable method to impart knowledge than lectures. It is also faster.
I wonder what the implications of this are for AI safety, and EA more generally? How beneficial would it be to invest in making some sort of tutoring ecosystem available to people looking to get into the field, or to advance from where they currently stand?
From Childhoods of exceptional people:
I wonder what the implications of this are for AI safety, and EA more generally? How beneficial would it be to invest in making some sort of tutoring ecosystem available to people looking to get into the field, or to advance from where they currently stand?