ai;dr

For what it's worth, whatever LLMs do extensively, they do because it's a convention in well-established writing styles.

LLMs have a bias towards expertise and confidence due to the proportion of books in their training set. They also lean towards an academic writing style for the same reason.

All this to say, if LLMs write like you were already writing, it means you have very good foundations. It's fine to avoid them out of fear, but you have this Internet stranger's permission to use your em dash pause to think "Oh yeah, I'm the reference for writing style."

In 2020 at the start of covid, I did an experiment I called Project 35 where, for 35 days straight before my 35th birthday, I wrote 3 times per day, for 10 minutes each, I livestreamed it and whatever I wrote I would put directly into a book with no edits. While I didn't invite many people to join the calls (maybe fear, maybe just not wanting to coordinate it all), I found the process to be more raw, more human, and less perfect than 10x edited writing. It also helped me get better at typing in the moment and not rewriting everything, especially for social media, HN, and other places.

Anyway, it's at https://www.jimkleiber.com/p35/ if you wanna check it out, all sessions posted as blog posts, I think there's a link to the ebook (pay-what-you-want) and there may be audio (I recorded myself reading the writing right after each session).

If you check it out, please let me know :-)

I think there’s an uncanny valley effect with writing now.

Yesterday I left a code review comment that someone asked if AI wrote it. The investigation and reasoning were 100% me. I spent over an hour chasing a nuanced timezone/DST edge case, iterating until I was sure the explanation was correct. I did use Codex CLI along the way, but as a power tool, not a ghostwriter.

The comment was good, but it was also “too polished” in a way that felt inorganic. If you know a domain well (code, art, etc.), you start to notice the tells even when the output is high quality.

Now I’m trying to keep my writing conspicuously human, even when a tool can phrase it perfectly. If it doesn’t feel human, it triggers the whole ai;dr reaction.

I actually have worked in this space and it, uh, has not shaken my belief that powerpoint talks are bad.

> Why should I bother to read something someone else couldn't be bothered to write?

This is an easy but not very insightful framing.

I want to read intelligent, thoughtful text that is useful in some way: to me, to society, to humanity. Ceteris paribus, the source of the information does not necessarily matter; it only matters as a matter of association. To put it another way, “human” vs “machine” is not the core driving factor for me.

All other things equal, I would rather read A over B:

A. high quality AI content, even if it is “only” the result of 6 minutes of human question framing and light editing [1]

B. low quality purely human content, even if it was the result of 60 minutes of effort.

There is increasingly less ability to distinguish “human” writing from “AI” writing. Some people fool themselves on their AI-detection prowess.

To be direct: I want meaningful and satisfying lives for humans. If we want to reward humans for writing more, we better reflect on why, and if we still really want that, we better find ways that work. I don’t think “buy local” as a PR campaign will be easily transferred to a “read human” movement.

[1]: Of course AI training data is drawn from humans, so I do not discount the human factor. My point is that quantifying the effort put into it is not simple.