How AI has changed my approach to writing
1. I am more likely to post "personal opinion"-style, or even emotional, content that's not factual / entirely subjective. Since AI has made it much easier to do research and compile factual summaries of any well-understood area of human knowledge, the value of pure-fact posts has decreased (though of course there's still value in knowing what to ask and curating / clarifying an LLM's findings). However, AIs cannot know or convey your unique perspective on subjective matters or internal experience.
2. I am less likely to edit or polish my writing. I used to spend time ensuring my writing was free of grammatical errors. Now I edit much less, only enough to ensure the thoughts flow coherently and it's easy enough for someone to read. Partly as a signal that the prose was human-written, partly because the effort put into editing is no longer a good signal of caring because editing is effectively free (you can get an LLM to do it).
3. I am more likely to maintain a distinctive style. Another reason I edit less is that I want to maintain a distinctive, personal style. When the first good LLMs came out I'd often use them to rephrase content, asking them to make my prose flow better or be clearer. But these days I want some clumsy constructions or phrases in my writing that are distinctively mine, I want a personal writing "smell" just like AI writing has a smell. As more people use LLMs to write, their styles become more undifferentiated, and so having your own style lets you stand out. Also when it comes to personal, non-factual content, writing in your own distinctive voice helps convey internal thoughts more precisely; some of the meaning is in the precise choice of phrasing.
