How to Begin

Or my First Writing Post

While I’m not a New Year’s resolution person, I am making use of the day to begin writing here. As part of my transition from primarily writing academic pieces to something that can communicate to a different audience, I’m going to experiment with using this Substack platform to share various thoughts over time.

I’m not sure what it will become, but I intend to have fun with the process and make sure to learn along the way. I do have an initial focus, to elaborate on some of the foundations of various projects I’m working on. I expect I might expand this later, but for now, it’s necessary to just begin.

AI and the Writing Process

While I have not used AI in this post, I want to be explicit about how I’m using it in the series to come. Having had bits of time on my hands over the last couple years, I’ve been able to explore how several people are approaching this. My main take away is that it can be a good tool in many ways, requires work on both ends, (preparing prompts and editing drafts), and that being reasonably transparent about this is the responsible thing to do.

So far I would characterize my writing process like this. I begin with a need of some kind. At the moment, it is to have a bit more outflow, to share things I’ve learned with those who might find it useful. As well, a more specific focus is to elaborate on how I’m approaching the design of learning activities focused on skill building for leaders.

Next, I’ll take the topic and draft a prompt for Gemini (my main LLM). The prompt will include some context I think might be useful and then a specific task. My favorite part of the prompt is to ask Gemini to ask me clarifying questions as the first step in the task. This helps me take the messy implicitness of how I described things and communicate it better. An added bonus I have discovered by doing this is that it often asks me questions I haven’t really thought about. They might be how the topic looks from a lens I don’t usually consider, or provide some information relevant to the topic I wasn’t aware of.

The result is generally a much improved output from Gemini. However, I love to play further, so now I have taken to asking Gemini how to set up a good system instruction for Gemini in the Google AI Studio. This itself requires a substantial back and forth to help me understand what will be useful as ongoing context for it and how to shape the role and foundational approaches it will take. While I can be more conversational in Gemini inside my browser, I can set tighter constraints in the AI Studio version.

Another bonus is the long context it has, so I played around by feeding it articles relevant to what I wanted to work on, as well as several articles and book chapters of my own. This enabled me to construct a role for it to use in drafting that could to at least some degree reflect how I might approach writing about the topic.

Iterations

Then it was back to the strategic conversational Gemini in the browser, to explore what might be a good set of topics for a series to explore this project. More back and forth, ending up with an initial set of seven themes for essays that later expanded to nine. Then finally, a first draft of the first theme from AI Studio Gemini. I was initially impressed. A first read through gave me the feeling that it had managed to make a reasonable draft that did three key things. One, it felt enough like my voice that I didn’t need to throw it away and start over. Second, it stretched me, as I had instructed it to write in a non-academic style. While I felt a bit of discomfort at some of the phrasing, I also want to experiment with how I communicate ideas and know that just doing what I’ve always done the way I’ve always done it won’t make the cut.

The third thing it did was stimulate me to dig in and edit. This wasn’t just to correct it, (although this was needed at times), it was that I got glimpses of connections I hadn’t seen before and inspiration to make those connections explicit. While my initial hope that the hours of set up would generate ready to go text was deflated, the next two hours of editing (which I always find much easier than the initial drafting) was fun for me. I felt much better about the outcome now.

Of course that wasn’t the end of it, as I then fed my edited version back to Gemini with the task of analyzing the differences to update its understanding of how I wanted it to write, as well as make suggestions for improvements. This led to the second essay, a similar round of editing and feeding it back into the system and so on. Along the way I realized a need for the two additional pieces that were not in the original outline and now have a complete set of essays to launch.

My final round was to feed the entire set of essays to Gemini and have it check for consistent use of metaphors, gaps in places like transitions and other areas to enhance the cohesiveness of the series. As of posting this on New Year’s day 2026, the job of using the recommendations Gemini provided to make a final pass through the texts is still awaiting me.

The Importance of Having Fun

Well, after putting this off for a long time, I now see the text flowed and I actually had fun. I am aware that this was easy work as descriptive narratives are generally easier (at least for me) than figuring out how to communicate complex abstract theoretical concepts in a useful way.

This feeling of it being fun is important. “If it isn’t fun, it isn’t worth doing” was a phrase I remember some of us using a few years back when working on what could easily be a not so fun project. This mantra served as a kind of compass, a check point that enabled us to keep from losing energy by taking things too seriously. That perspective kept things in perspective and helped us avoid getting bogged down in minutiae that could suck the life out of us.

So I’m aiming to keep this project of posting here fun. Where the fun is located might not always be apparent to you as a reader, but I hope it leaks between the lines and makes the coming essays enjoyable.