joshtronic

Posted in Technology

Mining Conversational AI for Content Ideas

I've mentioned it before, I don't use AI to generate content for this blog. I'm also open about the fact that I very much use AI for a bunch of stuff, including finding typos and misspellings. From my perspective, usage of that nature is no different than having spell check available in your editor.

Initially struggling with what to write about today, I decided to ask my robot friends for some help. Rather than going in and dropping in a generic prompt like:

Give me 5 blog topics ideas

I decided to mine my recent conversations for something that seemed interesting:

Based on our recent conversations, are there any blog-worthy topics?

It spat out some things, mostly just a top-level overview of a few of the topics I had been drilling on this week. Both ChatGPT and Claude said roughly the same topics, as I have been running conversations in parallel to see which one I'm liking more.

Having significantly more context in ChatGPT than Claude, I decided to dig further by asking:

Let's search a bit further back, any super niche topic ideas hiding out in our conversations?

The results at this point were way less generic, and definitely started to get into the deep cuts territory. I didn't pursue it, but we could have probably easily drilled into SEO stuff for the topics and started to put together an editorial calendar.

Of course the robots were quick to want to offer up content, which I happily declined.

What I did find really interesting was a generalization that ChatGPT made about the things I had talked with it about:

A lot of your best hidden topics share a theme:

“The problem isn’t the tool — it’s the assumptions people make about it.”

Definitely a real theme that is worth exploring more than any of the other content ideas it had suggested. Speaking of, none of the topics it suggested were things I want to talk about, hence why I went with talking about the process itself.

The biggest take away for me this week is that I didn't really do much solving of problems. When I don't solve any problems, I find myself at a loss for what to write about.

I've mentioned it on other posts, but I primarily post to have a record of the crap I had to figure out. Not that I don't love my readers, but I do this primarily for myself to have a record that I can look back on quickly and easily if the issue comes up again.