Waterbed Theory (2014)
wiki.c2.comMaybe a new paradigm could be the personalization of the language to each user, where the user controls which areas are exposed in lower or higher levels. The underlying code is stored in a neutral "byte code" and then decompiled on demand based on the user's preferences or job role.
It's totally theoretical and would create so many nightmare scenarios, but very cool to conceptualize.
I could maybe enjoy using such a thing, in some kind of coding based apocalyptic horror game, where AI caused the apocalypse, and you're trying to rebuild the world with technology that is all too complex to understand or classified because no one person is supposed to know it.
So you could have puzzles where you have to go to one site, steal a badge, go to another site and steal a different badge, and they all let you decrypt different views of the master AI you're trying to reprogram to undo whatever evil it did.
You’re basically describing Lisp m-expressions.
Which, incidentally, pretty much all Lispers ever have independently decided are a bad idea.
I don't think this will be out of the realm of possibility let alone practicality with more advanced AI.
prime takeaway here is to never use financing when purchasing a waterbed
One of my mathematics professors used to call this "the law of conserved difficulty" in his partial differential equations class.
This is comment is orthagonally-related, something I read just yesterday: Douglas Adams' short story on "A Puddle's Perspective":
tl;dr:
If a water puddle had perception, then the containing "outside" always seems to be "the perfect fit" for the puddle, no matter how much nor little rainfall.
Until eventually, the last water vapors dissipate, and the puddle is no more.