4 ms·
If you were around for the heyday of Markov chain email and Usenet spam this whole thing is familiar. Sure AI slop generation is not directly comparable to Mark
by varjag 1y ago
If you were around for the heyday of Markov chain email and Usenet spam this whole thing is familiar. Sure AI slop generation is not directly comparable to Markov process and generated texts are infinitely smoother yet it has similar mental signature. I believe this similarity puts me squarely in the offended 22%.
- handfuloflight 1y ago[flagged]
- varjag 1y agoSee mostly on the subject but just slightly veering off context texts like this is why it's irritating. Arguing something not entirely relevant yet short of complete strawman and debating the tangents that were never brought up while never adding anything of substance.
- handfuloflight 1y ago[flagged]
- varjag 1y agoGrasping nuance and implications of human replies is also something LLMs struggle with.
- handfuloflight 1y agoYou're claiming there's a detectable "mental signature" but dodging the fact that this claim is inherently testable. Either the signature is recognizable enough to put you in that "offended 22%," or it isn't. You can't invoke pattern recognition to justify your irritation, then hide behind "nuance" when the logical implication—that you should be able to spot it blind—gets pointed out. Turns out humans are just as evasive as LLMs when pressed to back up what they actually said.
- varjag 1y agoI am not arguing with bots sweetie. Sorry the little model that almost could, flagging your replies here.
- handfuloflight 1y ago[flagged]
- daemonologist 1y agoNot sure if this comment was written by a language model, or by a human pretending to be a language model.
- varjag 1y agoI'm certain the account operator's parents think he is smart.
- handfuloflight 1y agoYours don't?