@mavnn English is not my native language yet I’ve managed to write blog posts in it just fine for decades.
Sure sometimes I wrote broken English but if you have something to say people don’t care.
If you sound boring and artificial—like you certainly will by using LLMs—no one will want to read it and you will never learn how to write.
¯\_(ツ)_/¯
@thomasfuchs It can be more difficult for non-native speakers. I tell that by experience, unfortunately.
@thomasfuchs grok, is this post written by AI?
@[email protected] I agree with a lot of your thoughts on AI, but I'm afraid on this one I'm going to have to say that is a dangerous take that harms the people who were paid to help train the models (generally people in low income countries dealing with enough crap already). If an AI chances to produce information that isn't nonsense, style isn't a good indicator of AI authorship; a good write up of how this causes harm is this one: segunfamisa.com/posts/no-you...
@mavnn There’s more than just writing style to this.
@[email protected] I agree in the sense that if you don't find an written AI book strange there's something wrong - generated writing has nothing much in the way of actual content. But 'LLM, rewrite the points in this article for my precious SEO juice' versus 'I'm going to try and write my first technical blog post and I come from somewhere that isn't the US/UK" is something that needs to be assessed on context not denegrating the people who 'have LLM writing tells.'
Although I suppose some of that comes down to GenAI created (almost certainly meaning free) versus GenAI ripping off (almost impossible to tell from human written because a human supplied all of the meaning that is now being ripped off)
@mavnn English is not my native language yet I’ve managed to write blog posts in it just fine for decades.
Sure sometimes I wrote broken English but if you have something to say people don’t care.
If you sound boring and artificial—like you certainly will by using LLMs—no one will want to read it and you will never learn how to write.
¯\_(ツ)_/¯
@[email protected] I agree with everything you've said except for a caveat on the final point, which is that using LLMs to directly plagerize something interesting will produce interesting (and potentially human sounding) output. Which is just one more reason why I'm not keen on them when it leads to false accusations of plagerism towards people who are writing for themselves. But it sounds like I'm not really managing to explain myself, and I fundamentally agree with where you're coming from so I'll bow out of the conversation now.
@thomasfuchs but here’s the kicker!
