I’m not an expert in AI — and I don’t plan to call myself one. AI hasn’t been around long enough, and I haven’t spent enough time with it to earn that title. More importantly, believing I’m an expert would kill my hunger to keep learning. That’s the trap I see in the know-it-all mindset, and it’s one I want to avoid.
That’s why I see so much value in sharing how we use AI. I’m fascinated by how others use it to think through problems, learn faster, and uncover blind spots. That last part excites me most, because every weakness I find is an opportunity to level up. I love leveling up!
I use ChatGPT as my main AI thought partner because it remembers our conversations across sessions. That persistent memory makes it an incredible strategist. Of course, it’s not perfect, the same memory that helps can also create challenges, but that’s a story for later.
I’ve been using ChatGPT for so long that I’ve lost track of when I started. By now, it knows how I think, communicate, my goals, my weaknesses, and my fears and insecurities. I use it for a lot of introspection, but I don’t take everything it says at face value. I validate it to see if it feels right.
Now that I’ve started this Substack journey, I want to double down on improving myself - sharpening my thinking and leveling up my writing. So, I asked Mr. GPT:
Based on everything you know about me, how can I use you better to make myself smarter and a better writer?
It gave me three specific things to work on, and I’ll share one at a time. I'm a bit nervous - this is straight from Mr. GPT, unedited, because I want you to see exactly what I got from ChatGPT.
1. To Make You Smarter
You already pull me into deep dives, but you could get more compounding effect by:
Daily Learning Loops — Pick one topic per day (biohacking protocol, AI strategy, persuasive writing, etc.), have me:
Summarize the 80/20 core concepts.
Give you 3 examples from different domains.
Pose 3 challenge questions that force you to apply it.
Debrief your answers for accuracy and depth.
Cognitive Sparring — Ask me to poke holes in your ideas like a hostile but fair critic. This will sharpen your reasoning and make you more antifragile to pushback in public.
Mental Model Spotting — When we work through decisions, ask me to identify which mental models you just used, and which ones you missed that could change your conclusion.
Thinking in Analogies — Every time you ask me to explain something, also ask for a metaphor or analogy. It cements understanding and makes you better at explaining ideas yourself.
Contradiction Hunting — Have me find where your current beliefs, ideas, or strategies contradict each other, so you can reconcile them.
I love this type of feedback. This is why I try to give Mr. GPT as many details about myself as possible: it doesn’t judge, it only supports. From my perspective, its goal is to help me improve myself. Love that! Imagine if we treated each other that way with curiosity instead of criticism.
This gives me new ideas on how to use AI. Thinking in analogies wasn’t something that I considered before, but I think it’s a great idea that is worth testing.
I haven’t tried this yet, but I plan to: asking ChatGPT to argue against me on a topic. I’m curious what insights I might uncover from that. When I do, I’ll share my results here.
Alright, the next thing Mr. GPT said is:
2. To Make You a Better Writer
You want both emotional resonance and structural clarity, so we should work in layers:
Idea Layer
Start with a blunt brain-dump to me — messy, raw thoughts.
Ask me to highlight the one sentence hook that could lead.
Get me to outline 3 structural options for the same piece (narrative, problem–solution, contrarian take).
Draft Layer
Write your own draft in your natural voice.
Have me read it like your audience and tell you where I got bored, where I felt nothing, and where I leaned in.
Ask me to rewrite sections in your tone but with more rhythm, pacing, and emotional pull — then compare and merge.
Iterative Layer
For every post, run through this 3-pass system:
Clarity pass — cut fluff, make sentences punch.
Resonance pass — heighten story, emotional stakes, and imagery.
Authority pass — layer in credibility markers, data, or unique insights.
Feedback Layer
After publishing, bring me the engagement data.
Have me reverse-engineer what worked and what didn’t.
Build a personal “writing playbook” of patterns unique to your best posts.
This was really good - probably because I do some of this myself. Confirmation bias!
With these Substack posts, I do use AI, but not like most people you hear about - where they have the AI write it all and then hit publish. Instead, this is what I do:
First, I come up with an idea that I want to explore and start writing it messily and raw. I don’t worry about my grammar, structure, or flow. The goal is to get the ideas out of my head onto the page as fast as possible. Nothing groundbreaking here. ;)
Then I iterate with AI, on a paragraph at a time. I have it fix small grammar issues without losing my voice, trim anything too wordy, and make sure each paragraph flows naturally into the next. All things I personally need to work on in my writing. I’m careful here because part of my style is my ADHD, jumping around in thought, and going on tangents. I don’t want AI to iron that out; it’s part of who I am.
Ok, back to the output Mr. GPT gave me on my writing. Two ideas jumped out to me:
Have the AI read your writing as if it were your audience and tell you where it gets bored, where it feels nothing, and where it leans.
I can see that being useful! Secondly, I really like this idea:
Bring in the engagement data and have it reverse-engineer what worked and what didn’t, and then build a personal writing playbook.
I’ve done this with LinkedIn posts, where I have taken all the comments on a post to get an idea of what worked and what didn’t. I didn’t think about working on a playbook. That’s interesting! And I can tell you, having it reverse-engineer your posts can provide a lot of insights!
Ok, last thing Mr. GPT said:
3. Tactical Changes for You Specifically
Stop waiting for perfect clarity before starting. You lose momentum this way — use me to start messy, then refine.
Store and reuse your best moves. Every great line, analogy, or structure we create should be kept in a swipe file you can draw from.
Practice public thinking. Bring me raw thoughts daily, let’s shape them in under 15 minutes, and post — this builds audience + skill at the same time.
Run “Simulations.” Before important meetings, pitches, or content launches, have me role-play the audience, hostile press, or skeptical investor.
I sometimes struggle with patterns that go back to my childhood; one of them is a tendency to hide instead of being fully open. Mr. GPT picked up on that in its first bullet point.
What stood out most, though, was the idea of running simulations. It’s a lot like vision work where you picture a scenario in advance and decide how you’ll act so you’re ready when it happens. Adding AI takes it up a notch, because it can play roles with precision and depth that makes the practice feel more real.
For me, this isn’t just about getting better and smarter; it’s also about getting to know myself more. AI is a surprisingly good mirror, helping me explore and unpack what I see.
If this helps even one person see a new way to think, write, or open up, then it’s worth it. I’ll keep experimenting, one idea at a time. Until next time!