> Purely NLP with no screen

Cumbersome and slow with horrible failure recovery. Great if it works, huge pain in the ass if it doesn't. Useless for any visual task.

> head worn augmented reality

Completely useless if what you're doing doesn't involve "augmenting reality" (editing a text document), which probably describes most tasks that the average person is using a computer for.

> contact lenses

Effectively impossible to use for some portion of the population.

> head worn virtual reality

Completely isolates you from your surroundings (most people don't like that) and difficult to use for people who wear glasses. Nevermind that currently they're heavy, expensive, and not particularly portable.

> implanted sensors

That's going to be a very hard sell for the vast majority of people. Also pretty useless for what most people want to do with computers.

The reason these different form factors haven't caught on is because they're pretty shit right now and not even useful to most people.

The standard desktop environment isn't perfect, but it's good and versatile enough for what most people need to do with a computer.

And most computers were entirely shit in the 1950s

yet here we are today

You must’ve missed the point: people invested in desktop computers when they were shitty vacuum tubes that blow up.

That still hasn’t happened for any other user experience or interface.

> it's good and versatile enough for what most people need to do with a computer

Exactly correct! Like I said it’s a limitation of the human society, the capabilities and expectations of regular people are so low and diffuse that there is not enough collective intelligence to manage a complex interface that would measurably improve your abilities.

Said another way, it’s the same as if a baby could never “graduate” from Duplo blocks to Lego because lego blocks are too complicated