a preprint
aug 29, 2026so there's a paper / essay that I wrote two years ago and then never published or submitted anywhere because there's nothing like the numbing apathy of self-cancellation to ruin a good day. today, I finally finished reading The AI Con (Bender/Hanna), an excellent and comprehensive look at why the current AI industry/hype is both actively harmful and full of crap. What I *didn't* find there, though, was something that really fires a killing nail into the concept of a "Singularity". The pieces are all there, and maybe all I'm doing is saying the same thing in a different rhetoric style, but it still finally hit me that maybe I should just put this out into the world and see if it helps anyone.
with that in mind, here's an ugly PDF export of my scrivener project, slightly cleaned up to reflect my better understanding of GPTs (although I'm pretty sure I still left 2024 in there and probably need to fix a lot still). I guess I would upload this as an arxiv preprint or something, except I have no idea if/how I would get anyone to recommend I be allowed to upload something to the AI category, and no idea if this would ever turn into something actually published. but for any official purposes, this is a preprint that I would love to turn into a published article for someone.
No Singularity - On the Mathematical Impossibility of an Unchecked AGI Explosion
the graphs are a quick hack, there's a diagram missing, but. maybe someone will listen.