Hmm, your guidelines feel half-baked to me.

We need to keep in mind that bots are a deep paradigm shift. We’ve been collectively coerced into foolishly believing that the best way to…

Hmm, your guidelines feel half-baked to me. Having guidelines is great, but fabricating them just for the sake of having them ends up in deception. I am of the opinion that it is better not to have any guidelines and therefore fly by the seat of one’s pants, than it is to have phony, contrived guidelines.

We need to keep in mind that bots are a deep paradigm shift. We’ve been collectively coerced into foolishly believing that the best way to interact with software is by manipulating the screen. We’re now learning that it was a lame, foolish thing to do, which was unfortunately forced upon us by the severe limitations of the technology.

These limitations are now slowly but surely being threatened, and so we’re learning that screens are only good for consumption, never for interaction.

We’re now moving into the world where we get things done the way we always have — by interacting with others using speech/text. And there is no point in doing that if we’re going to bastardize it and develop some weird clunky way of interaction. So your advice “don’t pretend to be a human” is as meaningless as if you gave an advice to a graphic designer “don’t pretend that this widget on the screen is a button”. It IS a button; it looks like a button, and it feels like a button, and it behaves like a button. Same is with a bot — it reads text, it responds with text, therefore it behaves like a human. Nothing is gained if we rob the experience of that dimension. Only everything gets lost if we charge down that path.