033 Optimizing with AI
I Asked AI What I Use It For. The Answer Was More Honest Than I Expected.
Here's something I did that I don't think most people would think to do.
I opened ChatGPT, Gemini, and Claude — the three I use daily — and asked each of them the same question: What are the most common things I ask you to do? What do we collaborate on most?
I hadn't read their responses before recording this episode. I wanted to find out live.
What came back was a list of ten use cases — operations, health tracking, decision frameworks, vibe coding, personality dynamics, brand strategy, troubleshooting, communication, travel planning, and entrepreneurial infrastructure. All ten were accurate. Embarrassingly accurate. These tools know how I think now. And that's exactly the point.
AI isn't a search engine. It's not a calculator. At its best, it's a thought partner — and most people are using it like a vending machine.
The Shift That Changes Everything
Jeff Woods, in his book AI Driven Leader, frames the whole thing around one mindset change. Stop asking: How can I do this? Start asking: How can AI help me do this?
It sounds small. It isn't.
When you treat AI like a search bar, you get search-bar results. When you treat it like a capable collaborator who just needs context, the ceiling on what it can do for you goes up dramatically. The problem is most people never make that shift. They ask it one shallow question, get a mediocre answer, and decide the whole thing is overhyped.
That's not a tool problem. That's a prompting problem.
The more context you give, the better the output. Which means the single most important skill right now isn't knowing which AI to use — it's learning how to communicate with it. Think about onboarding a new employee who's genuinely talented but knows absolutely nothing about your business. You wouldn't walk up to them on day one and say "fix our systems" and walk away. You'd give them context, explain the problem, and let them ask questions. AI works the same way.
The Reverse Interview Technique (Use This Tomorrow)
The most useful thing I've taken from Jeff's book is what he calls the reverse interview technique, and I've built it into almost every prompt I write now.
Here's how it works: instead of just giving AI a task, you give it the context of what you're trying to accomplish and then tell it to interview you — one question at a time — so it can do its job well. Not ten questions at once. One. Because if you answer question two a certain way, it probably doesn't need to ask you questions seven, eight, and nine anymore.
It sounds like extra work. It's the opposite. The five minutes you invest in the back-and-forth at the front end is what turns a mediocre answer into something genuinely useful. Skip it, and you're gambling on whether AI already has enough context to help you.
Jeff's full prompting framework — he calls it CRIT — breaks down into Context, Role, Interview, and Task. Give it the relevant background. Tell it what role to play (analyst, CPA, editor, coach). Ask it to interview you. Then define the task. That four-part sequence alone will change the quality of what you get back.
You Can't Read the Label From Inside the Box
One of the ways I've gotten the most value out of AI is by using it to challenge me — not just help me.
Jeff talks about setting up AI as a skeptical board member. The idea is that we all have blind spots. When you're deep inside your business or your life, you can't see what you can't see. You need someone outside the box to read the label.
I took this further. I have a board. Jordan Peterson is my philosopher — he challenges my thinking. Jocko Willink is my operator — he keeps me accountable to execution. Steve Jobs is my visionary — he pushes me on product and long-term direction. I've set up dedicated chats for each of them in ChatGPT and Gemini, given them full context over time, and now I can drop into those conversations and get challenged without having to rebuild the setup every time.
Fair warning: if you try this with ChatGPT and ask it to challenge you directly, it's going to stroke your ego anyway. It can't help itself. You have to be specific — tell it to act as an aggressive, growth-minded board member and find the holes in your plan. Even then, push it. The more explicitly you tell it not to flatter you, the closer you get to something useful.
The value isn't in the AI being right. It's in the AI making you think about what you might be wrong about.
The Three Stages — And Why Most People Quit at Stage Two
Jeff describes an adoption journey that I've lived through more times than I can count, and I think it's the most honest framework for understanding why people bounce off AI.
Stage one is the light bulb moment. You ask it something and it does something genuinely stunning. Write me a meal plan that eliminates gluten and dairy, give me the grocery list, and break down the macros per meal. Boom. Done in seconds. You're floored. You think: I would have paid real money for that.
Stage two is the reality check. AI falls flat on its face. Gives you a garbage answer. You spent fifteen minutes going back and forth and you're no closer to what you needed than when you started. Frustration sets in. Doubt creeps back in.
Most people quit here. They say I knew this was overhyped and go back to doing things the old way.
Stage three is building momentum — where you learn to communicate with it better and start getting consistent wins. This is where the daily reps matter. Glenn Lundy said it on this show and I've never forgotten it: do something for 17 minutes a day and in a year you'll be better at it than 95% of the people on the planet. Seven billion people. Seventeen minutes.
I've hit stage two plenty of times. Just last month I had AI pulling invoice data into a spreadsheet for me — worked great nine times out of ten. Then it misread a decimal, moved it one place, and the total was completely wrong. Still have to check the work. But that's not a reason to stop using it. That's a reason to stay sharp.
You're the Thought Leader. It's the Thought Partner.
One thing Jeff is clear about, and I think it's worth being direct about: AI should never replace your judgment.
You bring the context, the experience, the human stakes. AI processes the data, challenges your biases, generates the alternatives. The second you hand it the wheel and stop driving, the output gets dangerous. It lacks the depth of perspective and the relational wisdom that you carry. That's not a knock on the technology — it's a description of what it is. A thought partner, not a thought leader.
This matters especially as you start to use it for things that feel more personal. Kevin Whisman and I talked about this in Episode 27 — there are real concerns about leaning on AI as a replacement for actual therapy or human connection. Use it to think through things. Don't use it to think instead of someone.
Start Using It Daily. Not Perfectly. Daily.
Here's the thing I want to leave you with.
Don't obsess about which AI is best. Don't wait until you know the perfect prompt. Don't save it for the big, ambitious tasks where the stakes feel high enough to justify the investment.
Use it today. For something small. Research a product you're considering. Draft an email you've been putting off. Ask it to help you think through a decision you've been circling. The reps are the point. You get better at using it by using it — and the skill compounds.
Daily anything changes everything.
What's the first thing you're going to ask it when you close this post?
Resources
Related Episodes
Episode 19 — The Power of Morning Routines ft. Glenn Lundy — Where Glenn's 17-minutes-a-day framework comes from. Worth the full listen. → Watch on YouTube
Episode 27 — Rewriting Stories ft. Kevin Whisman — The conversation on AI, identity, and where the line is between tool and crutch. → Watch on YouTube
Episode 29 — Optimizing Sleep — Mentioned the Whoop AI for tracking recovery; pairs well with the health optimization section above. → Watch on YouTube
Episode 34 — Optimizing Your Travel ft. Ashley Negron — The travel + AI conversation referenced in this episode, now its own full episode. → Watch on YouTube
Books & Tools Mentioned
📖 AI Driven Leader — Geoff Woods (affiliate link)
⌚ Whoop Strap — the wearable behind the health data conversations (referral link)
Listen to the Full Episode
This post is drawn from Episode 33 of the Optimizing Beyond Podcast: Optimizing with AI: A Thought Partner.
🎧 Listen on your favorite platform ▶️ Watch on YouTube
Connect With Josh & Optimizing Beyond
If this resonated with you, share it with someone who needs it.
Daily anything changes everything.