I didn’t expect to rely on AI as much as I do now. At first, it was just curiosity. Then it quietly became part of my routines.
Several people asked me recently, “How do you actually use it?” That got me thinking. AI is now embedded in many areas of my life, from health and finances to music and philosophy.
This is a behind-the-scenes tour of how I’ve set up projects with ChatGPT and, occasionally, Perplexity. It's not about automation for its own sake—it's about thinking more clearly.
Why ChatGPT—and Why I Rarely Use Google Anymore
I use ChatGPT for its memory and conversational style. Each project I run has its own space with its own voice, tone, and history. I can build continuity over time, like a long-term collaborator who knows my preferences.
But for research—especially when I want citations or current data—I use Perplexity. It's now replaced Google for me. Instead of sifting through SEO clutter and ads, Perplexity gives clean, sourced answers, fast—whether I'm asking a complex science question or just wanting the phone number of the highest-rated pizza joint nearest me. We wondered what might topple Google's ad-driven search model, and this might be it.
Together, the two tools complement each other. One thinks with me. The other scouts ahead.
Below are some of the more important ChatGPT projects I engage with almost daily.
Project 1: Personal Health
Health is where my AI use is most detailed. In this Personal Health project, I’ve instructed ChatGPT to provide only science-based analysis. No anecdotes, no influencer blog bros—just peer-reviewed studies and expert perspectives.
I’ve uploaded years of blood lab results, x-rays and MRIs, summaries from past surgeries and treatments, physician notes, typical diet and supplement stacks, Garmin sleep and body composition data, and my FitBod exercise history. That gives ChatGPT a layered view. It can identify biomarker trends, connect them to habits, and suggest possible tweaks.
Just one example: ChatGPT spotted some mild spinal compression on a recent MRI I had for an unrelated reason and cross-referenced that with my workout history. It found some rotational movements and carries that might be increasing lateral spinal load. I made some changes—and felt real improvement in mobility and comfort.
Insights like that would take hours to uncover manually, if I even knew to look. AI connected the dots.
Project 2: Financial Planning
In my Financial Planning project, I asked ChatGPT to speak to me at a CFO level—no simplification, no dumbing down. I also set the context: I’m a very conservative investor with long-term goals, clear estate intentions, and known spending patterns.
Using a link to Kubera, ChatGPT can see all of my assets across IRAs, Roths, taxable investment accounts, home and farm real estate, and private business investments. It’s like handing your financial strategist a live balance sheet.
With that, I explore questions like:
- How will Roth conversions affect IRMAA and RMDs?
- Where should I draw down from first?
- How should I tweak bond ladders in a shifting rate environment?
Once I’ve refined a scenario, I often move it to ProjectionLab to stress-test the assumptions visually.
What I’ve found most valuable isn’t just the math. It’s the clarity. The ability to think through complex options with a calm, rational partner who never tires of the next “what-if.”
Project 3: Travel
I’ve always enjoyed travel planning. But it takes time—especially to get it right. So I built a Travel project designed for independent, active travelers in their early 60s who enjoy comfort—but aren’t afraid to get a little dirty. I told ChatGPT I like business class, smaller airports, early arrivals, unique stays, and backcountry detours when they make sense. It even knows what airlines, aircraft, hotels—and specific seats—I prefer.
It remembers all that. And it builds trips that feel like us.
Here are a few examples:
- Buenos Aires over New Year's: Returning again to one of our favorite cities, ChatGPT created a vegetarian restaurant tour in a land of steak. It found the hidden gems by consolidating reviews across multiple platforms and creating a day-by-day plan that aligned with our other activities for the day. This is how we discovered Marti, one of the best vegan restaurants we've found anywhere in the world, like Uzu in Kyoto, Winners in Athens, and Teresa Carles in Barcelona.
- Vietnam, 3 Weeks Last March: Most travelers go north to south. AI suggested the reverse due to weather patterns that month. It made a real difference in comfort and scenery.
- Greece, 3 Weeks Last May: It recommended islands and routes that dodged early summer crowds without losing the feel of the season, as well as the most reliable ferry schedules. Less traffic. More quiet charm.
I’ve also used it for road trips. One itinerary included only scenic two-lane highways, short daily drives, stops at boutique hotels, and glacier hikes—all perfectly sequenced.
Over time, it remembers more. What I liked. What didn’t work. Each trip gets better.
Project 4: Activities
This project brings AI into everyday life. Here are just two of the subprojects in this section.
- Live Music Curator: Each week, ChatGPT curates a list of live music shows in our area that fit my tastes—folk, acoustic, roots, jazz. It filters by venue, time, and even reported sound level. Before AI, I'd have to search multiple websites and lists, now a single prompt creates it in seconds. I share the list with friends, and it’s become a favorite weekly ritual.
- Recipe Manager: I’ve uploaded links to pescatarian recipes we enjoy—many from Northern Africa and the Mediterranean. I told ChatGPT I want weeknight meals that take less than 30 minutes and feel fresh. Now it suggests new recipes weekly and generates consolidated grocery lists. And when friends come over, it remembers what they’ve had before so they get a unique dish each time. It even offers pairings and side suggestions.
It’s a small thing—but it makes cooking more fun and social.
Project 5: Business Therapist
While I'm now retired and no longer have company responsibilities, I do work with a local incubator to mentor startup founders. The Business Therapist project helps with that mentoring.
I’ve set up separate chats for HR, Sales, Leadership, and Strategy. Each one references thinkers I admire—and has “read” their books, articles, and even critiques.
When a founder has a team conflict, I’ll ask: “What would Lencioni say about this situation?” Or I’ll check how Radical Candor might guide a tough conversation with a struggling employee. It’s like having a personal advisor who has read every leadership book I’ve ever valued—and can apply them to today’s questions.
This has been especially helpful when supporting new leaders learning how to lead with clarity, empathy, and accountability.
Project 6: Ruminations
This last one is the most personal.
In Ruminations, I’ve asked ChatGPT to be a science-informed philosophical companion. I told it I believe we know very little about the universe. I’m open-minded—but not interested in speculation without rigor.
Sometimes we explore free will. Other times: biblical translation, consciousness, or quantum entanglement. One day I asked, “Does the Many Worlds Interpretation eliminate moral responsibility?” and we spent the next hour exploring implications for ethics.
It’s become my version of journaling. But the journal writes back.
Issues and Concerns
As useful as these tools have become, they’re not perfect. A few limitations are worth noting.
The most common concern people raise is hallucination—AI making up plausible-sounding nonsense. That was a problem early on, but I’ve found that the latest version of ChatGPT is far less prone to it. And when you carefully add context, define project goals, and give clear instructions at the start, hallucinations become rare—especially in structured domains like finance and health.
Still, it’s worth verifying. That’s one reason I run financial models through ProjectionLab after working through scenarios with ChatGPT. It serves as a second pass to check the math, stress assumptions, and expose anything overlooked. Interestingly, I haven’t seen hallucination issues in Perplexity at all. Likely because it ties every claim to a citation.
Another limitation is that most personal data still needs to be uploaded manually. I’ve built those pipelines myself—lab results, spreadsheets, PDFs—but it takes time. That said, platforms like Kubera are starting to offer more direct integrations, and more apps are building connectors.
Alongside that, I take privacy very seriously. I’m careful to anonymize or remove specific account numbers and identifying details when uploading financial documents. Same with names of friends or family mentioned in planning or journaling projects. It’s a small discipline, but an important one.
And one final note: ChatGPT is very good at keeping a conversation going. So good, in fact, that it’s easy to forget to pause. I’ve caught myself solving problems that didn’t need solving—just because the conversation was interesting. Every so often, I try to step back and ask: Is this project still useful? If not, I set it aside.
The point isn’t to use AI constantly. It’s to use it well.
One Prompt That Changed Everything
One simple question has led to more insight than any other: “What else should I ask?” Yes—asking AI what I should ask AI. I use it all the time. It helps me discover the edges of my thinking. Find gaps. Ask better questions. And those better questions often lead to better decisions.
It’s a good reminder: AI’s value isn’t just in its answers. It’s in how it helps us connect the dots and explore the unknown.
Closing Thoughts
These six projects are just a start. I’ve found AI to be most useful when I give it structure, purpose, and context. The results aren’t always perfect. But they’re often better than I could have done alone. AI doesn’t replace my judgment—it makes it sharper. And in a world full of complexity, that’s something I’ve come to value.