Introduction to Reducing Goodhart — AI Alignment Forum