No, wait!
I’ve been using the Emacs LLM-backed spellcheck mode to go through my old comics blogs like the one about Epic Comics. Most blog posts are pretty ephemeral so I don’t really care that much, but those comics blogs are sorta kinda meant to be actually readable, so getting rid of obvious spelling mistakes is nice. (Several people have told me they’ve read an entire comics blog of mine! My god! The humanity!)
Going over all those articles has been an experience. It’s just so embarrassing! I know, I said in that first article that I try to be less neurotic about these things, but every time the LLM points out that I’ve incorrectly used “there’s” instead of “there are” I feel mortified. Slightly. Slightly mortified. That’s the ticket.
So here’s My Theory: People use LLMs to generate their texts because that absolves them of responsibility for their texts.
No responsibility, no embarrassment.
I.e., it’s less embarrassing to post an abhorrent text like this than to post something you’ve written yourself. When people point out that it’s unreadable dreck, you don’t have to take it personally.
If you’ve misspelled “embarasing” and somebody points it out, that obviously means that you’re a bad person and a moron and should be ashamed (etc., depending on your level of nerdly neuroticism). But if Claude wrote it, it’s not a reflection on yourself at all.
I mean, this theory only applies to nerds who care about this kind of stuff. Most people obviously don’t, and there are other mechanisms in play for them, but I kinda feel there’s something here.
