Post

continues to be quite wild watching people make lists like this when the consensus remains that humans, even experts, are basically wrong as often as they are right about whether something g is written “with” an LLM. (What the fuck does “written with an LLM” even MEAN?)

To me this makes a lot of sense, even tho it can def veer into ableism stuff. It’s not about accuracy, it’s a proof of work thing. There’s a relationship in communication, and people are renegotiating what it means to invest in that relationship. For some, genAI reduces the work in a detrimental way

work doesn’t have value in and of itself

if someone thinks it's respectful to me, a reader, to generate the text of a post with an llm ... then surely they can have no problem with me consuming their writing in the same way: with an llm

so to be clear, you’re willing to mislabel publicly roughly half the people you put in this list as having generated text with an LLM and of those who actually did, as not having put thought in, in order to catch, what, maybe 30% of people who actually did generate text with an LLM and didn’t think?

The math just isn’t mathing, man. Thats a huge level of cost you’re placing on others for a relatively small benefit to you. And again, we’re talking reputational costs that many folks likely shouldn’t bear, not just “oh, Daniel summarizes my writing with LLMs I’m so sad”

That makes total sense. I always appreciate when people write better than the average of the internet. Takes skill.