> We know what it is good at and what it's not. We know what it's good at today. And pretty sure it won't be any worse at it in the future. And 5 years ago sta

> We know what it is good at and what it's not. We know what it's good at today. And pretty sure it won't be any worse at it in the future. And 5 years ago state of the art was basically output of Markov Chain. In 5 years we might be at another place entirely.