[ReDeiPirati]

we already know that deep RL is sample inefficient... is meta-RL really useful for something not trivial? This seems rather silly.

[mtrazzi]

The first applications appear really simple because they illustrate cognitive abilities that imitate planning/model-based RL related to neuroscience/psychology.

It doesn't mean that meta-RL won't scale up with more computation (see http://www.incompleteideas.net/IncIdeas/BitterLesson.html).