The Curious Companion: Ep. 6 – An Introduction to Using ChatGPT for Content Creation​

Curious Reader!

Welcome to this week’s ChatGPT Curious companion newsletter.

What you came for is below, and you can CLICK HERE to listen to the episode if you decide you prefer earbuds to eyeballs.

Happy reading!

In this episode I give an introduction to using ChatGPT for content creation with a focus on practical strategies for repurposing your own work. You’ll hear about reverse prompting, using Projects to save instructions, and how to keep your taste and voice at the center of what you create.

Quick Updates on ChatGPT

• All the models are back and available for use, so much for model unification.
  • GPT-5 models: Auto, Instant, Thinking mini, Thinking
  • Legacy models: 4o, 4.1, o3, o4-mini
    • ​Here’s a resource that explains the legacy models
• There were two recent ChatGPT outages
  • The one on August 20th was possibly tied to roll-out of the new $5 tier in India, which makes me wonder if (aka when) they’ll roll that out in the US to push more people into paying.
  • I’ll likely do an episode about this in the future, but OpenAI has to make more money somehow, and dependence is part of the equation. Only about 2% of ChatGPT’s users are actually paying (which is normal for freemium SaaS at global scale) and even with billions in revenue, OpenAI is burning through an estimated $1–2 billion a year (loss) just to keep the lights (and GPUs) on. If they make the free tier worse, people will pay.
    • I’m curious: Are you a paid or free user? Hit reply and lemme know.

Fun fact, I only knew about the outage because I was at the store and was trying to use ChatGPT to compare elderberry syrup labels when I got the OpenAI equivalent of the wheel of death.

I decided to try an experiment and put the same query into Google to see what it would give me. Gemini, their AI summary tool, was able to understand my question, but the response wasn’t nearly as in-depth or specific as that from ChatGPT (I ran the query in ChatGPT once I got home to compare).

Two takeaways:

• You can use natural language to search on Google. ChatGPT (and other LLMs’) ability to understand everyday language (aka anyone can immediately use it) has been a significant driver in the rapid growth of LLM usage.
• People want the answer to their question, not just a list of links. This is a big reason that I think LLMs will disrupt search as it currently exists.

This does NOT mean ChatGPT will “kill Google.” Everything evolves, everyone wants money, and they’ll all do what they have to do to stay relevant. It’s just something that is worth noting and noticing.

Everything Is Content (Even With ChatGPT)

Shoutout to Stephanie Hine for inspiring this episode with her question box submission.

When I talk about social media, I often say “everything is content.” For today’s episode I want to flip that idea slightly: Everything you create with ChatGPT is content.

That means you shouldn’t think of using ChatGPT for social media as something radically different. You’ve already been using it for emails, lesson plans, recipes, workout programs, business ideas, and sales pages. Social media is not an exception. Same idea, different packaging.

When It’s a Good Idea

Using ChatGPT for content creation can be a good idea if:

• You’re leaning into the strengths of the model.
• You’re generating content from content you’ve already created.
• You’re editing, adding your flavor, and making sure it sounds like you.
• You trust your own taste more than anything else.
  ​
  ​

If you’re just using it to generate a shit-ton of AI slop that you’re going to spam the interwebs with, then no, it’s not a good idea.

How To Start

The Components: Picking a format, repurposing the content, and using reverse prompting

Pick a format: Carousels, reels, quote cards, whatever flips your pancake (feel free to pick multiple formats). You can even chat with ChatGPT about which format makes the most sense.

Repurpose the content: Look to repurpose one type of content (that you’ve already created) into another. This way you have a solid input to give ChatGPT, and it can generate as opposed to just making something out of thin air.

• Blog → Carousel
• Podcast → Reel
• YouTube transcript → Carousel
• Presentation → Quote cards
• Transcript → Newsletter (that’s how my podcast outline becomes the Curious Companion newsletter)

Use reverse prompting: Create the input material and the desired final output material → paste both into ChatGPT → then ask ChatGPT to generate a reusable prompt for how to generate that output when given a similar input.

Example: Write the blog post (input material) and the copy that will go on the carousel slides (desired final output material), then paste both into ChatGPT and ask:

“I’m going to paste X and Y. Please analyze both and generate a prompt that I can use to consistently and reliably get you to create Y when I input X. Then explain what makes it work and how I can adjust it if needed.”

Voila! You’ve got a repeatable system for creating content. You’ll still need to check it, tweak it, and make sure it sounds like you.

Of note, you can absolutely have ChatGPT help you create that desired final output material that you will use to create the reverse prompt. Just make sure you edit it so that it’s exactly what you want.

Projects: A Favorite Feature

If you’re wondering how to repeat this process without pasting that same prompt in every time, this is where one of my favorite features comes in: Projects.

You’ll need the paid version of ChatGPT (Plus works fine). Think of Projects like folders, a place to separate and organize the different things you’re working on. One of the best parts about Projects is that you can save instructions for each of them!

So that reusable prompt you created earlier? You can store it in the ‘Instructions’ section of a Project, add to it if/as needed, and set it up so whenever you paste something in the input field, ChatGPT automatically applies those instructions.

Projects are a really cool feature but honestly, and I may do a dedicated episode on them later. For now, just know they’re a way to make your system organized, repeatable,and easier to use.

How I Used ChatGPT Recently

Each episode I include a section where I briefly discuss how I used ChatGPT that day/week.

This one’s a throwback, but it ties into the elderberry syrup story from earlier. When Lex and I were in Gstaad, Switzerland last month, most of the grocery store labels were in German. At one point, I grabbed a package of meat for a charcuterie board, took a picture, and asked ChatGPT: “What does this say?”

And it gave me an answer.

Here’s why that worked so well: Yes, LLMs are probabilistic models. They generate answers based on patterns in the data they were trained on. The reason you can trust them for translations? Well, for things that are well-established and well-represented in training, like general conversation or common words in a major language, the models will almost always produces the facts correctly. What changes is the style of the response or how much detail the model gives.

Worth noting, translation and language learning were actually part of the GPT-5 demo rollout, so it’s a use case that OpenAI is really leaning into.

That’s it for today’s episode. Always grateful for you.

Questions, comments, concerns, additions, subtractions, requests? Hit reply or head to the website (​chatgptcurious.com​) and use that contact form. I’d love to hear from you.

Catch you next Thursday.

Maestro out.

AI Disclaimer: In the spirit of transparency (if only we could get that from these tech companies), this email was generated with a very solid alley-oop from ChatGPT. I write super detailed outlines for every podcast episode (proof here), and then use ChatGPT to turn those into succinct, readable recaps that I lightly edit to produce these Curious Companions. Could I “write” it all by hand? Sure. Do I want to? Absolutely not. So instead, I let the robot do the work, so I can focus on the stuff that I actually enjoy doing and you get the content delivered to your digital doorstep, no AirPods required. High fives all around.

Did someone forward you this email?
​Click here to join the ChatGPT Curious newsletter family.​

Stay curious.
​