I’m 99% in goal with my private project: www.thewheelofclarity.com. This is something I’ve dreamed of for a long time—writing an e-book based on the teachings of a mentor. The original material was difficult to work with; it felt too spiritual and wasn't grounded enough for most people to actually use. I had a vision of making it simple and relatable, but getting there has been a long journey of trial and error. The Battle with "AI-Speak" I’ve tried to create this book several times. If you saw the first drafts and compared them to where I am today, you wouldn't believe it was the same book. The first many versions were done in ChatGPT, but I almost gave up several times. No matter how much we tried to improve it, the AI couldn't stay on track across all the chapters. I eventually switched to Gemini, and that’s where the version I have today finally took shape. But then I hit a wall with the "AI mindset." When I started looking at marketing, the AI kept insisting the target audience should be "high achievers" and "fast-paced CEOs." But that’s not what this book is. It’s for everyone—from stay-at-home parents to office workers to CEOs. It’s about understanding why certain patterns repeat in your life and how to change them. I ended up going through the book chapter by chapter to make it more human and accessible. Even after I thought I was done and created an audiobook version using ElevenLabs, I could still hear too much "AI" in it. It used words and phrasing that real people just don't use. I spent many nights listening to every single section, manually stripping out the "AI-isms" until it felt like my own. Building the Site with Claude Code The audiobook idea actually came from a colleague. I used ElevenLabs to find the right voice, and while you can still tell it’s AI, I’m proud of the voice I created and I’m sticking with it. For the website, I saw a video by Jake about folder structures in Claude Code. It sparked my curiosity, and even though I didn't fully understand the structure yet, I just jumped in. We built the site www.thewheelofclarity.com slavishly from top to bottom. We focused on one section at a time—brand voice, colors, and layout—until I was satisfied. Once the foundation was there, I added the technical side, like heatmaps and tracking tags, to help me optimize it later.

I wanted proof that my prompts improved from four months ago. The results turned into this post. Around early January I added these instructions to my Claude.ai user preferences: If required information is missing, ask clarifying questions before answering. Before giving the final answer: list assumptions, identify missing data, state confidence level. If appropriate, advise on how to write a prompt more efficiently in the future. Then I had Claude pull my chat history from before and after, and look for patterns. I figured I'd see changes in what I was asking. The actual change was in how I structured conversations around the asking, in three phases. Phase 1: one-line prompts (early January) Real prompt from January 8: "How do I set up a eSIM on a Windows laptop?" I was asking the way you'd ask a search engine. Claude wrote a generic eSIM tutorial. I bounced because it didn't match my situation, and never came back. That was my default. One sentence prompts. No context, no constraint, no goal. Phase 2: Claude starts showing its work (mid-January) This is where the instructions started doing actual work. The "list assumptions" line forced Claude to write down what it was filling in for me. When a response opened with "Assuming this is a Windows endpoint with standard user permissions and no recent OS reimage," I could correct the wrong guesses before they corrupted the rest of the answer. About half the time, at least one was wrong. "Identify missing data" produced a list of the questions Claude wanted to ask but was about to silently guess at. Reading that list every response taught me what to include upfront. Every "missing data" bullet was a future prompt fix. "State confidence" forced Claude to mark which parts of the answer were solid and which to stress-test. "High confidence that one of the first three checks will identify the cause" is useful in a way that a confident-sounding wall of text just isn't. The prompt-efficiency line pulled the other three together into a habit. After enough rounds of "next time include the OS version and whether the machine is domain-managed," I stopped needing to be told.