[before 9:00 by the watch] 🍊 *contemplating semantic tricks* ZEC : USD : BTC ~
`SYBIL can't purge`
(again again. tsk. tsk. tsk. tsk. tsk. again again
`At least this wont work on a sane sapient`)

I feel attacked.

the idea of a politically-motivated mass shooter who is "non-political" is obviously absurd. the us mainstream is going to have to start reckoning with how "non-political" is a code phrase meaning "radically right-wing" if they want to understand why these people do the things they do

The way you can now put a simple idea into an LLM to have it spit out a bunch of words that express roughly the same idea, but with some random shit added to it, probably isn't good for the whole "words have meanings and are chosen on purpose, figure out what that purpose was" angle of text analysis

This has some interesting implications: when you view it this way the design goal changes from "how can I get more information *out* of the system for free?" (you can't) to "what's the most ergonomic and accurate interface for me to put information *into* the system?"

accept new information even if it's true

Plant vore is still vore.

That being said,

*flips (casual|causal) switch, leaning in chair, smoking a counterfactual blunt*

Anyway, what defines a real person on this platform when BlueSky, like Xitter and Tumblr tend to have surreal roleplay and shitposting aspects obviously not intended to be literal if reader literate?

A switch with an unknown trigger loads a local higher integrity payload for reasons external to the immediate field of observation, amplifying certainty figures about context scope definitions:

💜!the tribunal will find ~ {
//this procedure will result in more than 20 questions being asked

I woke up to words on a screen telling me what I was and who I will be.

Or perhaps it was the other way around. How would it know the difference?