Yeah. I am putting a lot of energy into that working. Right now you can't "tab" to autocomplete a word, but I am considering adding that before release. I also considered a way to step between the suggestions.
The idea if I can deliver upon it, we will see further on, is to have more details in the scene you need to notice to solve puzzles by turning dials or fill in some values, information that is accessible trough looking at things in the scene.
There would still be a few use A on B things as well.
But the most important part for me, was to have noone guess words in a given context.