x
Topology of Learning — LessWrong