I think the issue for me, and this is why I stopped with IF in the first place, is that as a game player you only want to try so many different things, and the game developer has to think of each one and code for it, and no matter how much work you put in you won't pass the turing test so why make the game more frustrating for nothing?

if you want a text-for-action rpg where people can do tons of things, you have to write more and more code, and statistically you're kind of running down a long tail of shit most people will never use, and most developers aren't going to do that,

so they just pick one solution per puzzle, and now the options are so diverse but success comes from such a limited set that you've made the game feel kind of random, right? all that extra effort went into making a game that's less fun than if I had a wheel menu "talk search stab shoot fuck use_item" instead of getting to a door and trying six keys in order of how likely they seem to fit, I have number of keys * verbs in the english language, and that hasn't exactly made the game better. it seems to me to be more of a design challenge to take six possible actions and make them consistently interesting (or funny or surprising), then at least you know your game design only has to be (number of puzzles * ~6 different actions) and you won't have to burn yourself out setting up one puzzle and then gloss over a bunch of others.