No-Free-Lunch theorems are often irrelevant - History — AI Alignment Forum