Room and Board for People Self-Learning ML or Doing Independent ML Research — LessWrong