AI-generated thought leadership sucks

And authors and readers both know it

I have no issue with thought leaders using generative AI in their creative process. It's available, it's amazing, everyone is using it in some capacity or another.

However, I'm also familiar with the rhythms and stylistic tics of ChatGPT, Claude, and other tools. And when I see something that obviously came right out of one of them, I just can't take it seriously. Even if the thinking is serious.

ChatGPT uses a stop/start dynamic to make business prose sound portentous and prophetic. Numbering, bullets, short sentences, lots of stop and go, the infamous em dash — which I also use, but more sparingly than ChatGPT does. If I read a few lines presented in this rhythm, I think "Oh this is ChatGPT" and my mind checks out.

Claude is more introspective and sounds more writerly and belletristic. It writes more flowing prose, issues more challenges, makes finer logical distinctions. That's how I think, so Claude prose tends to pull me in more. But I still recognize it, and increasingly when I see it these days, I check out too.

The first thing I wonder when I see gen AI prose is: where did the line of thinking potentially break down? Gen AI uses expert mimicry, based on probabilistic guessing at which letter/word comes next. It can look great, sound great (especially to a short attention span), but still be nonsense.

I personally use gen AI as a research, outlining, and thought-testing engine. I use it to hammer ideas into shape, then I write the prose myself.

Maybe I just like the sound of my own authorial voice too much to let it go. But I don't think it's that. It's that when I use gen AI prose directly, I can feel my line of reasoning being taken from me, and I don't like it.

There are a lot of rounds of AI evolution and adoption left to play out. I don't have a settled opinion on the use of generative AI. But right now, under current conditions, this is how I see it.