Technomages. They were humans (and some others) bonded to Shadow-manipulated creatures that could affect spacetime. Technomages learned to "cast spells" in a private language they developed with their symbiote. These were essentially commands to "do the thing I'm talking about."
One of the challenges Technomages faced after acquiring their symbiote is gaining control over both the personal language and the impulses for destruction embedded by the Shadows (bias? bias).
So yes, learning the tricks of getting our AI agents to do the thing we actually meant, while aiming to overcome bias built in to the models when we can see it does sound an awful lot like Technomages dealing with their symbiote.
Sincerely, I respect your response to how arbitrary it seems in this form.
But... I'd like you to take a moment and think really hard about whether this is truly novel behavior for LLMs, or rather something that has always been part of the interplay between inter-agent communication and intra-agent thought :)