Compounding Content: How To Train ChatGPT To Write New Medium Articles Based On Old Ones

This is how you really start "compounding" your content library.

Hey there, Digital Writers!

The hardest part of writing online is starting.

I don’t mean this in the cliché sort of “You just have to start!” inspiring-Instagram-quote-graphic sort of way. I mean this in a very literal, mathematical way. The hardest part of writing online is starting because on Day 1 you have no historical data. You have no content library. You have nothing “before” to inform what you can or should write “next.”

But once you start writing… you begin to gather data.

And once you start gathering data, you pick up momentum.

I have been writing on Medium since 2018.

Over the past ~6 years I have probably written over 1,000 articles on that platform.

Some were original articles. Some were republished snippets from Ship 30 for 30 blog posts. And some were viral Twitter Threads (aka: proven data points) expanded into longer-form assets. At my height in 2019/2020, I was earning over $100,000 per year writing on Medium—most of which I’d say, by that point, was just “republishing” and “recreating” old material. Not bad!

https://twitter.com/Nicolascole77/status/1409572338870501381

To this day, I still don’t see very many Digital Writers (if any) doing this with their content libraries. Most think that in order to keep building their web of content, they must constantly create net-new material (which, especially over 5+ years, is very mentally taxing). When in reality, the secret to “niche dominance” as a writer is to hammer home “the hits” over and over again.

For example:

There’s a reason why Ryan Holiday continues to pump out “X Stoicism Lessons From Marcus Aurelius” despite the fact he’s written multiple books, thousands of articles, and hundreds of YouTube videos and podcast episodes on the exact same topic.

Because volume one 1 topic is how you become known for a niche you own.

Link to Ryan’s article, if you’d like to read.

The problem is: Repetitive volume is taxing.

It’s not easy to write about the same thing over and over again—especially over a 5 or 10 year time horizon. (That said, we do have a great framework for this, called The Lean Writing Method. But even still, “patience” is the key skill to build.)

Even for someone like me, who has thousands of articles to pull from, it’s mentally taxing to go back through my entire library on Medium, find a topic I can “rewrite,” and then block the time to do it (considering how many other projects I am working on, and how much other writing I do on Twitter, LinkedIn, in the Ship 30 for 30 ecosystem, etc.).

So, while it’s a great strategy… it still takes time, and has to be done manually.

Until now.

How To Get Leverage On Your Content Library

What I am about to share with you can work for any platform, based on any historical content library.

But for the sake of clarity, I am going to keep this explanation specific to Medium.

It recently occurred to me that the manual process I was going through to “rewrite” old Medium articles into net-new content was the exact sort of thing ChatGPT should be able to do quite well. Even if it’s not perfect, it should be able to get me 80% of the way there—and then I can take it across the finish line.

This writing strategy + ChatGPT = massive compounding on your content.

Here’s how it works, and how I trained ChatGPT to write net-new Medium articles based on my historical library.