Read on Twitter
Some notes on education. I'm absolutely convinced we haven't scratched the surface of what's possible. It'll take a second to explain why, but hopefully will be worthwhile. CC @Austen (you might be interested). Thread 
1/12
1/12
2/12 Over the past six months or so I've been learning the game of Go (aka baduk/weiqi). Like chess Go has an objective ranking (give or take some difference between servers), so you have a very good idea of how well you're doing.
3/12 The game is massively popular in Asia (Korea, China, less so Japan) and orders of magnitude less popular than chess in the west. One consequence of this is that most educational resources are not in english. There are a few hundred english Go books, though. Plenty to learn.
4/12 I've been learning and doing problems for months, without any improvement. I've tried lots of strategy books, tactics books, different problem sets, etc. Rank didn't budge. I've been hearing about Korean Go book series called Level Up! Very hard to get in the US.
5/12 Level Up! was made in Korea and the translation is a labor of love. It was designed by people who actually care about pedagogy in addition to caring about Go. So they did "obvious" things no other Go book I've read has done.
6/12 For example: instead of having a pro-level player guess the difficulty of a problem (hint-- they're so good they can't estimate low-level difficulty for the life of them), the authors actually tested the problem difficulty on real students.
7/12 This is super important because if you keep getting problems that are too easy or too hard, you're getting an awful learning experience. The authors did many other things too-- organization, print size, problem order (problems build on each other), etc.
8/12 It's all "obvious", but I've read dozens of other books and did many many problem sets, and Level Up! absolutely blows them away. Anyway, the proof is in the pudding. Since I started working through Level Up! my rank keeps steadily improving. It's actually quite incredible.
9/12 After this I realized that our lessons plans for basically everything (programming, math, etc.) Basically suck. They're not tested on real students. They're not organized well. They're not designed to be fun (think Slack vs Teams).
10/12 You might think "ok, but it's a board game, math is way harder". Nope. Go is an incredibly sophisticated game. You're thinking through many layers of abstraction, you need working memory, visual recognition, etc.
11/12 If they could make such drastically better learning resources for Go, no field is off limits. Many people suspect that we can make better pedagogical resources. Level Up! is the most incredible proof of this I've seen.
12/12 There is a question of business incentives of course. There is no selection pressure for pedagogically improved materials. So perhaps that's the meta-problem to solve first. When we do, the sky's the limit to what we can achieve.
