Briefly: no.

Less briefly:

There are two generated images on this site in a blog post I wrote a while back, the only current exceptions to an otherwise uniform policy:

I wrote, sketched, photographed, screenshotted (?), or illustrated everything you’ll find here.

Except for logos I pilfered to make a few post banners. I didn’t make them, I just put them on solid-colored backgrounds. And the fonts. I wish I made the fonts, but the only one I’ve made so far in my life really sucked. Making fonts is hard.

Content comes from my head, and I’ve warmed up to using agents, specifically Claude Code, to review and improve the code behind the scenes.

Of course I had great fun questioning language models and training my own image models and generating stuff from prompts! Of course I think there are great uses for AI that can benefit humanity!

But not here.

The tools I use help me realize ideas that start in my head. They do not provide generated content I can substitute for the inefficient, flawed, painstaking, deeply human work of having something to express and working my way through it.

The closest I get to offending this rule is asking LLMs for word suggestions like I would a thesaurus. I had considered every LLM my regular expression expert until some of them didn’t work and I was worse off than I started.

Perpetually struggling to find avatars for myself that I like, I once trained some models and generated a few I was happy with. I even started using them in a few places until I decided it amounted to theft, and that I skipped the process of making my own art (illustration, photography, whatever). So you won’t find those anymore.

If you even think you’ve found any exceptions to this, you’re cordially invited to call me on it.

If you and I exchange words, you can be assured mine are from old fingers clacking away at keys.

The words and images on this site are from a strange, imperfect, hopefully-always-evolving human being.