Battle of the Bots: Part 1 of 3
When I Decided to Let My AI Tools Fight It Out
I have a confession to make: I talk to my AI tools like they're real people.
Not in a weird way. More like, I talk to them as if they are my creative partners. Because after months of near daily use, that's honestly what they've become.
I use AI for all aspects of my life. Professional work, where I serve on multiple AI committees and taskforces and oversee a marketing and communications team. Personal creative projects that keep me up way too late. Brainstorming session when I'm stuck. Even figuring out what to cook when I'm staring into the fridge at 6 PM wondering how I'm supposed to turn still frozen hamburger and a cupboard of random spices into something edible.
ChatGPT, Claude, MidJourney, Canva—they're all part of my daily toolkit. Each plays a different role, and for some of them, I’ve discovered that they even have personalities. Real, distinct personalities that affect how I work with them.
The Question That Started Everything
A few weeks ago, I was working late on a project that required both research finesse and creative breakthrough thinking. I found myself switching back and forth between Claude and ChatGPT, using each for what they do best. Claude for the elegant writing and quick iterations. ChatGPT for the bold ideas and creative problem-solving. They each have their strengths.
Here’s an example of how they differ when working on a narrative project. I’ve found that Claude can be amazing at helping me refine my voice, but it also has a tendency to spit back my words exactly as I wrote them. Verbatim. No prompts, no suggestions, not a single tweak. It also tends to miss my questions if I’m mid-project and switch from narrative to brainstorming. Chat, however, has no problem jumping right in with a completely re-written narrative using an entirely different tone and voice, if it thinks that it fits the conversation. Yet, I find that it problem-solves with me like a boss, and it picks up every cue I provide and question I ask, even mid-stream while drafting a narrative.
That's when it hit me: these tools have such distinct approaches to thinking and problem-solving that I naturally pair different tasks with different AI tools. But, I wondered if they understood their own strengths and limitations?
More importantly—can they articulate the differences between themselves and their competition?
The Experiment
So, I decided to find out. I designed what might be the first real AI self-awareness experiment, and I want to show you exactly how it evolved—because the process was as revealing as the results. Now, I did an initial experiment for LinkedIn, and then, decided … let me take this a step further. I decided to go bigger, and post that experiment here, on Subject to Change, for you all. If you want to TL;DR version, head on over to my LinkedIn and read that article. Feel free to drop your thoughts there. But, if you’re up for an even bigger experiment, stick around.
Here's how the conversation developed:
Initial Setup: I started by sharing my detailed experiences with both tools—the conversation limits, file upload differences, writing style observations, creative breakthrough approaches, everything I'd learned from months of real use.
The Pivot: Instead of asking for analysis, I decided to flip the script: "Actually, I'd like to publish an article BY you, and an article BY ChatGPT to showcase the difference. Can you write a comparison article using all my experiences as the foundation?"
That’s basic article you’ll find on my LinkedIn. But, I decided, we needed more. I don’t just use AI for work, so I wanted to enhance the pieces. Make them bigger. More in-depth.
The Refinements:
As we developed the pieces, I made real-time adjustments:
"Let's acknowledge that I use AI in personal creative projects"
"Turn this into a Substack piece, leaning more heavily on the personal side"
"I want this as part of a three-part series"
Final Assignment:
"Write a Substack article comparing Claude vs ChatGPT from Claude's perspective, using my authentic voice and real experiences."
Then I took the same conversation thread and background to ChatGPT with an identical final prompt.
The results? Two completely different approaches to the same assignment—and the differences reveal everything about how these tools actually think.
One Caveat
The day after I finished this little experiment, I logged into Claude and received the most glorious message telling me that Claude now can reference between chats. This was one of the major drawbacks of Claude in this comparative experiment.
I did not re-run the project. What you’re reading reflects the time when Claude did not have this capability.
I’m playing with Claude and testing out its capabilities. So far, they’re limited. But, there’s some overlap. So, anything is better than nothing. But, I’ll come back in a month or so with an update.
Why This Matters (Beyond Tech Curiosity)
This isn't just about which tool is "better"—I've learned that's the wrong question entirely. After months of working with both, I've realized that choosing between Claude and ChatGPT is like choosing between a thoughtful analytical friend and an energetic creative collaborator. You don't pick one; you build relationships with both. Both have value. Both feed off each other. Both are necessary.
In fact, I often find myself feeding Chat content to Claude and vice versa. Double checking content in each platform, getting opposing viewpoints, and sometimes ignoring them both. Sometimes wishing I had a third friend to serve as a tiebreaker. Then remembering, wait … I am the tiebreaker.
But, what I'm really curious about is something deeper: how does artificial intelligence see itself? When given the chance to self-reflect and compare, what does each tool emphasize? What blind spots do they reveal? How do their distinct “personalities” show up in their own writing?
What You'll See in This Series
Over the next two posts, you'll read each AI's take on this comparison challenge. Pay attention to:
Writing style differences - How each tool approaches the same assignment
Self-awareness levels - What each emphasizes about its own capabilities
Competitive analysis - How generously (or not) each acknowledges the other's strengths
Personality quirks - The subtle ways their distinct "thinking styles" show through
I'm genuinely curious which response will resonate more with you. And which AI will surprise you with its level of self-understanding.
A Personal Note
As someone who works in education, I get asked constantly about AI integration, safety, and "replacement anxiety." Here's what this experiment has taught me so far: we're not being replaced by AI. We're learning to partner with it.
These tools don't think like humans. They think like tools—with distinct approaches, limitations, and strengths. The magic happens when you understand those tools well enough to know which creative partner to call on for each type of challenge.
What's Coming …
Part 2 will feature Claude's response to the challenge.
Part 3 will showcase ChatGPT's take on the same assignment.
I'm betting you'll notice some significant differences in structure and creative approach. For example, Claude suggested you look for its, “characteristically measured, analytical approach, and natural writing flow.”
Your Turn
Before we dive into their responses, I'm curious:
Do you use AI in your creative work?
Have you noticed personality differences between tools?
Are you team Claude, team ChatGPT, or are you building your own strategic toolkit like me?
Drop a comment and let me know. Because ultimately, this experiment is about more than just AI comparison—it's about understanding how we collaborate with technology that's becoming increasingly sophisticated, and how we can use these tools to enhance rather than replace our human creativity.
Ready to see what happens when AI gets introspective? Part 2 drops tomorrow.
This is Part 1 of a 3-part series. Subscribe to make sure you don't miss Claude's response (Part 2) and ChatGPT's take (Part 3). Trust me—the differences are eye-opening.
Why I’d Rather Write a Post Than Read One (Yep, Definitely ADHD)
I opened Substack to read today. That was the plan. A good plan. Maybe even a noble one. I had dozens of unread newsletters queued up and a full tumbler of iced coffee and a protein shake.