You have a poor understanding of natural selection. All you have to do is select for survival those that are apt at looking at and manipulating things on screens, and kill everyone else so they can't reproduce. In a few thousands years, we would be evolved for it.

Also, working hard in the fields is also bad for our health, hence our life expectancies used to be a bit lower; looking and manipulating things would be an improvement from that...we can always hit the gym after our 4 hour workday touching things.

As a species, we really just want to create what will effectively replace us. Given a few more (i.e. 20-40) years of deep learning technology development, we won't even have to look and manipulate anymore. By that time, we hit singularity and hopefully we aren't being exterminated by skynet drones.