Towards a Formalisation of Returns on Cognitive Reinvestment (Part 1) — LessWrong