5 Mindsets for Thriving With AI 🚀

When I used ChatGPT for the first time, I was awestruck. The future was here … and it spoke back. But even with my immediate optimism that this technology could solve some of the world’s biggest problems, I felt waves of anxiety: Would AI dull my critical thinking? Replace me? Take over the world?

As I’ve explored AI for both personal and professional use, I’ve found myself building my own mental models to help ground my thinking¹. These aren’t rigid rules. They’re evolving mindsets and operating principles that help me make sense of how AI fits into work and life.

Here are five mindsets I’ve adopted that have helped me work with AI more effectively:

AI as a Copilot - Visual representation of working alongside AI as a collaborative partner

1. AI as a Copilot

In the early days of using ChatGPT, I had real concerns that my critical thinking skills were going to atrophy. It was so easy to treat AI like an answer vending machine: input a prompt, get a solution. That convenience comes with a major risk.

Research shows over-reliance on AI can lead to cognitive decline and skill loss, especially for developing students and children².

Over time, I realized that AI is at its best when treated like a creative thought partner. It can help you challenge assumptions, brainstorm ideas, and expand your perspective—but only if you engage with it deliberately. Satya Nadella, CEO of Microsoft, coined the phrase:

AI as a Co-Pilot, Not Autopilot.³

I constantly ask myself: Am I using AI to refine my ideas and deepen my thinking, or am I just outsourcing the hard work of problem-solving?

In Action ⚡:

I’ve always been drawn to philosophy and often wondered what it would be like to have Socrates as a mentor. So I created a simple GPT with this custom persona: Socrates as my personal tutor. Its only instructions? Ask me thoughtful, probing questions until I find my answer. It may sound quirky, but this setup has become one of my favorite ways to refine ideas and help me think more critically.

Prompts for ChatGPT:

- What might I be missing here?
- Challenge my assumptions.
- Use the Socratic method to help me work through this problem.
- Create a personalized quiz for me on this topic to assess knowledge gaps.

Don't Lose Your Voice - Importance of maintaining authenticity in the age of AI

2. Don’t Lose Your Voice

It’s easy to fall into the habit of letting AI write everything: emails, strategy proposals, even blog posts like this one. I definitely gave in to that temptation early on.

But I quickly noticed something. The drafts were polished, sure, but also bland. The more I relied on AI for writing, the more my own voice got buried. And as I scrolled through my news feeds, I saw more and more content that felt soulless and eerily similar.

That’s why your unique voice, with all its quirks, imperfections, and personal stories, has never mattered more. Authenticity is the new superpower in the age of AI. It’s what helps your message stand out. While AI can tell great stories, it can’t replace the power of human vulnerability and connection.

In Action ⚡:

My personal rule is that I always start with a blank screen. It’s the only way I’ve found to make sure my original ideas come through. Sure, that first draft is often messy, scattered, and overly verbose. But that’s part of the process. AI excels at tightening my writing and expanding on my ideas. But only after I’ve gotten my raw thoughts and personal stories down.

Prompts for ChatGPT:

- Tighten this draft to distill it down to the core idea while maintaining my voice.
- Review this draft from the perspective of someone with little context. What needs clarification?
- How can I make my messaging more compelling while staying authentic?

AI as a Scalable Intern - Leveraging AI for productivity while maintaining oversight

3. AI as A Scalable Intern

One popular analogy I keep coming back to is something I first heard from Block’s CTO, Dhanji Prassana: treat AI like a very productive intern. It clicked immediately and has shaped how I approach integrating AI into my work.

There are two key parts to this mindset:

Interns are eager and can provide serious value, but they often lack context and experience. You wouldn’t let an intern’s work go straight to a client without review, but that doesn’t mean their contribution isn’t worth it. Same with AI: you get serious leverage, but you still need to validate that the output is accurate and relevant.
Unlike a human intern, AI is infinitely scalable. That upends some long-held assumptions many of us may have, like:
- You don’t need to hire anyone to launch something meaningful.
- You don’t have to be a manager to influence strategy and drive deep impact.
- You don’t need to hire experts when you have the world’s greatest thinker and personalized consultant in your pocket.
- You just need clear prompts, the right context, and a little bit of vision.

In Action ⚡:

Multiple AI agents in parallel: I run several AI agents simultaneously depending upon my objectives for the day; one might be optimizing ETL processes, another validating data tables, another summarizing Slack threads, and another brainstorming strategic proposals with me.
Always maintain oversight: None of these agents operate unsupervised. I set up validation loops, require sign-off, and never let AI outputs go directly into production.
PM work becomes manageable: AI handles scheduling, note-taking, updates, and document curation, making project management satisfying rather than overwhelming.

Prompts for ChatGPT:

- Get context from these five documents and connect the dots to generate a cohesive problem statement.
- Perform unit testing and anomaly detection by comparing my preproduction data table to my production data table.
- Get context on my codebase and then add a feature to do xyz.

AI Is Not a Creature - Understanding AI as a tool, not a sentient being

4. AI Is Not a Creature

Just a few days ago, ChatGPT flat-out lied to me and I caught it confidently summarizing a book that it clearly had never read⁴. Because it talks like a person, it can feel weirdly personal when that happens, like someone broke your trust. But here’s the thing: AI isn’t a plotting robot with secret plans to take over the world. It literally can’t betray you because it doesn’t have morals, intent, or awareness.

In the early days, I sometimes caught myself in doomsday thinking, wondering if AI could become some rogue entity like in Mission Impossible: Dead Reckoning. But the more I learned about how these systems actually work, those fears faded. AI isn’t sentient. It’s just really sophisticated autocomplete.

It is technically more complicated and nuanced, but that is the clearest way I’ve found to explain it: ChatGPT predicts the next most likely word based on a massive pool of training data⁵. It doesn’t know or believe anything. It has no agenda.

While AI is not primed to take over the world and enslave humans, the propensity it has to fabricate truth is a high risk we must always be aware of. AI lacks a “truth validator” and isn’t grounded in reality⁶. And this can result in real-world consequences if not used responsibly.⁷

In Action ⚡:

I avoid anthropomorphizing AI or attributing motives and remind myself it is just predicting patterns. I like to think of AI like a chatty hammer. Great for a lot of tasks, but dangerous for others. For certain questions of accuracy or importance, asking Google or a trusted human might be the better tool.
If you’re brainstorming, AI’s habit of hallucinating can actually spark innovative or unexpected ideas. It has already led to some scientific and medical breakthroughs⁸.
But if accuracy matters, always double-check every fact, number, or citation externally. Never take anything from AI at face value.

Think One Step Ahead - Staying ahead of AI's rapid evolution

5. Think One Step Ahead

At an AI data conference earlier this year, I heard Anthropic cofounder Jared Kaplan share a piece of advice that really stuck with me:

Build things with AI that don’t quite work yet. Because the technology is advancing so fast that when things do work, you’ll be ahead of the curve.

That’s exactly how it’s played out for me. Back in 2023 and 2024, I was tinkering with LLMs and GPTs. Some experiments worked, but a lot didn’t. Then in early 2025, when Agentic AI really started to take off, almost everything I’d been testing suddenly started working overnight.

The same pattern keeps repeating. Early AI models struggled with math, for example. Many people wrote that off as a permanent weakness. But now, newer models outperform college students on math exams⁹.

The risk isn’t in failing with early AI tools. The real risk is giving up too soon. Fixed mindsets can leave you stuck while others are getting familiar with the rocket ship that is about to take off. David Wood’s book, Rewiring Your Mind for AI, explores this topic in-depth by highlighting how fixed mindsets of the past will hold us back in this new world¹⁰.

AI evolves so fast that mental models that served you last year might already be outdated. The only way to keep up is to stay flexible: experiment early, learn fast, and don’t get too attached to any one tool or paradigm.

In Action ⚡:

If something doesn’t work quite right, assume it will get better and make a mental note to revisit in a few months.
Hold your AI mental models firmly enough to guide decisions, but loosely enough to adapt when the landscape changes.
Expect and embrace change. Agility and continuous learning are critical skills for the job market now.

Choose Your Adventure

The thesis of this essay isn’t to suggest that everyone should immediately adopt AI because it will make life better. I acknowledge that there are diverse populations with different opinions and subjectively, some may truly believe that life was better before AI. What I am suggesting is that, like it or not:

We can’t put the genie back in the box. AI is here to stay¹¹.
It will change the fabric of society faster than most of us probably realize¹².
Early adopters who learn how to work with AI will have an immense advantage in the new world¹³.

AI can be a force multiplier for creativity, productivity, and insight, or it can be a dampener if used passively or uncritically.

Now is the moment to experiment, question, and actively build the mindsets that will help us collectively harness AI for good. None of us has all the answers, but those willing to get their hands dirty now and wrestle with the hard questions will help shape what comes next.

You get to choose your role: skeptic, spectator, or builder. And if you’re reading this, you are already on the builder path.

- Zach

🚀 Next Steps

💬 Connect: What mindsets are working for you? I’d love to hear your AI journey. Message me on LinkedIn.

📕 Research: Explore references below for more related content. Special plug for David Wood’s book Rewiring Your Mind for AI which I read while researching this topic and would highly recommend.

5 Mindsets for Thriving with AI - Overview of the five key mental models

1. AI as a Copilot

2. Don’t Lose Your Voice

3. AI as A Scalable Intern

4. AI Is Not a Creature

5. Think One Step Ahead

More Questions Than Answers

Choose Your Adventure

🚀 Next Steps

Footnotes