idk I do think it's worth pointing out sometimes that the ways these models mess up are very similar to the ways that humans mess up. It's funny you can almost always look at an obvious failure of an LLM and think of an equivalent way that a human might make the same (or a similar) mistake. It doesn't make the failure any less of a failure, but it is thought-provoking and worthwhile to point it out.
Obviously this particular case is not the failure of the LLM but the failure of the spammer who tried to use it.
But a human can only mess up so many times per second. Even if it wasn't AI, if it was just a pill that allowed them to type unhumanly fast, once they have the power to scale up their incompetence (or predation) they're a new kind of danger.
Sometimes I read comments like this and feel a swell of gratitude that I don't work with braindead novices that make LLM-like mistakes. Are your coworkers actually that bad?
It's certainly useful to draw carefully thought out comparisons between human and AI performance at various tasks.
But this meme is not that. It's literally just a meme that's posted reflexively to any and all posts that unfavourably compare AI to humans, without any thought or analysis added.
Obviously this particular case is not the failure of the LLM but the failure of the spammer who tried to use it.