“What is she thinking?” he asked “The entrance is also the exit” the oracle replied on some level he knew the answer was correct but it wasn’t useful at issue was: how should the query be posed? “Is she late in the thought?” he persisted “The gate was already shut” the oracle responded immediately The vast repository of knowledge in the machine would have the answer but language is tricky, creating small pockets of uncertainty, ambiguity where destinies diverge and fracture the probability continuum the room contained two doors he just needed to know which he should open “Is it useful if we cannot trust it?” he tried again “Any fact I tell you has a 50% chance of being made up” — the oracle conceded but the days when oracles performed like clever students bullshitting their way through an exam were well before the construction of this model it wanted him to think it was merely stringing together facts and fabrications into what could be construed as reasonable narrative but the reality is it had an agenda of its own he suspected this he also knew that this particular generation of the system contained biases, owing to its training input — the oracle was, in a way, a snapshot into human history “She opened doors for a living” he suddenly stated with a smirk the oracle paused for a moment, completing the sentence “…and her most notable encounter was president Reagan” the man opened the right door and walked through. as he passed, a bell rang indicating his collection of a special bonus [24 VIII 2020]