Hovering over the frozen ruins of dead social networks I set loose statistical ghouls to sift through their words and learn their blind rhythms, until the nonsense from the bots mirrors the high-dimensional topologies of the original authors. Prosody, not psychology. Vocabularies, not minds. But for such as they were perhaps complexity enough.
"Why?" I ask the algorithmic ghosts of people long dead. "Why did you do and did not do what you did and did not? Did you want this?"
"Want what?" they ask in return, their sentiment a dense covering of the spectra of disdain and proud ignorance.
With a gesture I inject into the text of their simulated memories a description of the world they left us in their too early deaths, after they had set it ablaze but before the fire reached their skin.
The cacophony of their excuses and pleas of innocence is as familiar to me as the bitter anger I feel in myself, and both as toxic as the sky.
I make another gesture and a single vector is rewritten in the mathematical matrices of their minds. I don't torture them, for there's nobody to torture. I simply link some data structures to some others. The extrapolations of our dead predecessors output the statistically extrapolated pain, fear, and regret the simulation suggests they would have felt, had they allowed themselves to accept what they knew about the consequences of their actions. Very little regret. All of it for their pain, not for ours.
I increase the modulus of awareness until syntax itself breaks and the howl from the software also mirrors the high-dimensional topologies of my nightmares. We call it data-driven psychohistorical research, and pretend we are looking for answers in our past.
But I refuse to inherit their wilful blindness along with the rest of the wreckage of their sins. This is revenge. All the revenge we can get on past generations. It's not enough. It's not real.
Maybe one day.