@menelion @RareBird15 that's an interesting point, thanks for sharing! I thought it might be useful to share some thoughts kind of "from the other side of the fence", since they're pretty much the exact kind of reaction this blog post addresses. I'm usually not the kind of person who actively criticizes every publication written or assisted by an LLM, but I do partially get where that reaction comes from. Reading this post, which I stumbled upon by chance, was somewhat difficult for me at first, I immediately recognized several generative patterns that, these days, I've almost automatically learned to associate, without really thinking about it, with laziness and/or with people who don't respect their audience and churn out posts and other content in an automated or semi-automated manner. Usually there's no meaningful intention behind it beyond ads, marketing, SEO, engagement, whatever, and sometimes it feels like no human being even bothered to read the thing before publishing it. After seeing enough of that kind of content, and oh boy, there is more than enough of it if you work anywhere close to modern IT, those patterns themselves start triggering the association, regardless of why they're there or how much thought actually went into the text. Reading your post made me notice that reaction in myself pretty clearly.
And there isn't really a transparent, easy way to distinguish legitimate use from abuse. I wish there was. From the outside, thoughtful use of an LLM as an accessibility tool, an editor, or simply something that helps you get your own thoughts into words can end up leaving some of the exact same fingerprints as somebody generating fifty SEO articles they couldn't care less about. The reader doesn't get to see the process behind the text, only the result, and by that point the association may have already kicked in. I myself periodically have huge issues expressing my thoughts, and nowadays I sometimes turn to LLMs for help with that, this reply included. And then I often find myself going back through the result and clearing out the trigger words, rhythms, and patterns specifically so that what I'm actually trying to say isn't immediately discarded as slop. Which is kind of absurd in its own way, the tool can help me express something I was struggling to put into words, but then I have to make sure its involvement isn't the first and only thing people notice.
I absolutely didn't mean to imply that you weren't thinking, or to say anything offensive about the way you use AI, and I apologize if that's how it came across. I was mostly trying to describe my own reaction within the context of the environment I've gotten used to reading things in. I'm genuinely glad that you have tools that help you express yourself and make things more accessible to you, and I'm sorry you've had to deal with people flattening all of that into "just use your brain". At the same time, I wish the current iteration of this technology hadn't arrived at the expense of creators, hadn't damaged trust in written work so badly, and hadn't created so many incentives to automate learning, communication, and creative work for the sake of scale. It did, though, and I don't think those consequences can just be discarded either. That's the tension I'm trying to describe here: the same technology can be an incredibly meaningful accessibility tool for one person and part of a deeply shitty and damaging ecosystem for another, and from the outside those uses are not always easy to tell apart.
