Worldbuilding chat: frame narrators & characters with personality disorders

Hi. I was able to experiment with worldbuilding (maladaptive daydreaming) in the chat mode of the playground.

I created an imaginary world where scientist cats battled dogs. Max was the main cat character.

The ai always chose the most generic names for characters. Max, John, etc. If you asked for an “usual but disturbing” name you’d get Satan or Lucifer. That’s kind of jarring. An author’s touch might be something more like “Darkfur the Betrayer.” (How to think like a seasoned author is really an underestimated opportunity for an unimaginable amount of lucrative products and games. Why sit around in the Metaverse looking at the walls when you can walk into an author’s mind?)

I was able to have the daydreaming boy suddenly awaken in his junior high English class. Basically suddenly adding a frame narrator to the story.

I specified that the boy had an antisocial personality disorder and was capable of hallucinating.

I placed the boy in the 1980s and had him standing outside an arcade. A pay phone rang. The boy answered. I asked ai who is was. It was another boy asking him to play video games.

I said no. That’s too boring. Make it Max the cat character. Chat ai said it can’t be a cat because they are imaginary. I said no the boy is capable of hallucinating and it can be the cat.

Finally chat ai relented and said ok it’s the cat.

I asked what did the cat say on the phone? Chat ai replied “meow” or “purred”.

Chat ai fought me about the cat talking like a human. I said the boy can hallucinate that cats can talk. Finally the cat can talk and asks for help.

Next I ask for a portal between the cat and boy worlds and ask chat ai to have the cat enter the boy’s world.

Chat ai refuses because the cat isn’t real. Again I argue the boy hallucinates and the cat can enter his world.

Finally the cat can enter the boy’s world and they get a steak at a restaurant.

Is there an opportunity here to make small steps to having a “theory of the mind” for characters (and humans in real world)?

An imperfect narrator is actually the norm in the modern world. Probably 20% of humans have personality disorders. The same amount that will actually shoot another human in battle (before armies figured out how to brainwash in boot camp).

The other 80% of humans are so traumatized by the 20% that they freeze, flight, or faint. They are usually unreliable to see or report truthfully because they instantly dissociate or freeze up when anything opposes the viewpoints of the 20%. They “faint” to cooperate to stay under radar.

Cognitive distortions are everywhere with both the 20% and 80%.

Hence, we still have a Bronze Age and Iron Age mentally about the nature of our existence. No one wants to be killed by the whole tribe for thinking differently about an imaginary cultural myth. What year is it?

The 20% make history a tale of lies, conspiracy, trauma, suffering, sadism, revisionist history, barbarism, brainwashing, slavery, and exploitation.

Without seeing the “theory of the mind”, ai can’t be trained properly.

Narcissists are the only ones writing anything publicly. (Hence, the 90+% lurkers rate on social media.). They are unreliable because of their toxic nature for too many reasons to list here.

The understanding of the imperfect narrator is the best way to reboot ai into actually comprehending the world through the disease, compulsive lying, cognitive distortions, etc of the human mind.

I asked for a “persona” endpoint in my wishlist. Having a dedicated endpoint to handling characters, behaviors, and dialog would be game-changing for a number of things, such as writing, games, and interactive stories.

1 Like