GDP per capita in 2050 — LessWrong