"There's literally no way of knowing whether something is written by AI"
if that's the case then why do i keep seeing the same contorted gibberish syntax in every piece of prose published by an LLM user
0 Boosts
0 Likes
@joojmachine @benjamineskola sure. but hiding actually takes effort and knowledge. so I’m already more confident the code is thought through however it’s written.
it’s still a breach of trust though, specially in PRs to projects with a policy. and this will come to light eventually.
@hbons @benjamineskola yeah, agreed
my point being that the main problem with LLMs isn't JUST that it allows people to generated unlimited slop, but that it can AT THE SAME TIME be unenforceable to try and ban them, since it will make the maintainers deploy extra effort to constantly find new ways to identify LLM usage as the sloppers will constantly try to find new ways to hide it
banning LLMs is the right thing to do, but through the course of this whole discussion online the optimist in me keeps trying to find a way to separate the sloppers from people that genuinely want to help, do volunteer work, but are misguided and still see LLMs as "just a tool" and help the latter actually develop the skills to do so, which I believe just a flat out ban/instant rejection of anything that uses them without further discussion misses
@benjamineskola it’s the easiest thing ever. like a bingo card.
@benjamineskola
*opens https://readme.md**
“What it does”
*closes https://readme.md*
@benjamineskola oh shit that’s actually a domain… for AI no less! I’m going to keep it for ironic effect.
@benjamineskola I'm as anti-AI as anyone else and think it has no place in pretty much any kind of work under a capitalist society, but this argument can be answered with one image: