I have different types of posts to support different habits of blogging.<p>1. First, I have classic essays that I wrote by hand. I'll ask an LLM for thematic flow or grammar checking, but it's my thoughts. It's more sporatic, but they're things I'm compelled to talk about and think deeply about. Example is on how visual programming is stuck on nodes-and-wires. <a href="https://interjectedfuture.com/visual-programming-is-stuck-on-the-form/" rel="nofollow">https://interjectedfuture.com/visual-programming-is-stuck-on...</a><p>2. Next, I have posts that I write on a schedule. These are what I call lab notes, which I'll post every monday no matter what. These are easy to write because I just recount what I've been doing the past week and what challenges and wins there are. This exercises the muscle of posting something. Example is this past week on type checking <a href="https://interjectedfuture.com/lab-note-60-writing-words-and-type-checking-effects/" rel="nofollow">https://interjectedfuture.com/lab-note-60-writing-words-and-...</a><p>3. I do write TILs, though these don't occur for me nearly as much at the moment, due to the type of work I'm doing. It's like my own stackoverflow, I guess. Example is on the common footgun of useEffect. <a href="https://interjectedfuture.com/today-i-learned/til-message-handler-inside-of-useeffect/" rel="nofollow">https://interjectedfuture.com/today-i-learned/til-message-ha...</a><p>4. Lastly, I have posts where I had conversations with LLMs in depth on a niche topic that I think others would find interesting. If it's a back and forth, at the end of the conversation, I'll have it write a blog post based on what I thought the salient points were. If it's deep research, I'll just post that. I mark it clearly at the top it was LLM generated. The generation can be good enough that it's worth the post for humans to read. But also, in the hopes that the next model might pick it up, as an indication of what's good an interesting. It's basically one single sample eval based on my taste of what's good. Example is how algebraic effects are handled across Koka, Eff, OCaml, and Unison: <a href="https://interjectedfuture.com/algebraic-handler-lookup-in-koka-eff-ocaml-and-unison/" rel="nofollow">https://interjectedfuture.com/algebraic-handler-lookup-in-ko...</a><p>All's to say to people, there's different modalities of blogging. If you pick just one, you might feel pressure to write when you have nothing to write. But if you don't write you don't get in the habit, which is why I do the lab notes. It makes me post even if I don't have anything to say, or the time to say it.