i wish i could take the entirety of everything these companies compiled about me,

dump them into a generative pre-trained transformer language model as training data,

and POTENTIALLY tell me to go fuck myself.

or hey maybe i’ll actually like this simulacrum. unlikely, but not entirely impossible.

or furthermore it might help me get a more accurate grasp of my shortcomings instead of the paranoid hallucinatory delusional shit.

the thing about an LLM is that it just takes a screaming leap based on what’s already there. It WILL generate an output. And if I’m the only one interacting with this predictive model of myself, I can shortcut the process of pussyfooting around until i embrace a snap decision by creating a snap decision instantly AND THEN improving on it. It’ll readily cough up something that I might do, but without the prior investment of emotionally and psychologically interrogating myself for the impulse first, thus freeing up those considerable resources for actual productive insight, refining the initial impulse (or outright defying the initial impulse) using factual information