1) Software screen switching is not instantaneous

2) Moving your head is lower friction than pressing a button (and more likely a key chord)

3) Humans have very good spatial awareness and multiple screens have this affordance that a button would not.

Glancing from one screen to another is easier and lower mental overhead and you can look for input in your peripheral like text appearing without losing current context. You can also read from one screen while typing in another which is impossible in a non visible workspace.

If your one-and-only screen is touch, you have to lift your arms up away from your buttons to press anything. No matter how much of an impost it is to have to look down a little, I'd say it's still less than having to lift your hand away from the buttons into your eyeline.