[lostmsu]

Low data modes are also progressing quite fast. There is Dreamer and more recent papers based on RL in learned world models.