You are absolutely correct, which is why I keep harping on the Bloom 2 sigma problem. Its an easy way to highlight to readers how significant human 1 on 1 mentoring is to academic outcomes.

That immediately makes it clear to people that we dont have the money or people to afford that outcome, and after a few traversals on that ladder, the issue of education being a huge method to minimize the cost to educate people without overloading on number of teachers and cost of HR becomes apparent.

Mentorship is my new educational holy grail today frankly. Gone is the hope of the MOOC, now we need the Young Lady's Illustrated Primer, a mentor that customizes information that matches a student.

Fundamentally though people learn through play, thats how humans are designed.

Kids pick up 5000 combinations from playing league of legends, and do stupid number crunching on the fly.

However play is to hard to create, because we don't associate fun with a lot of work.

If we ever solve education, it will be solved by using play, even Mentors must understand fun.