Three New Habits for the Age of AI

This is the second in a series of three pieces on “unlearning,” in partnership with Maven, the expert-led course platform. In the first installment, Hilary Gridley explained why faster prototypes don’t make product decisions easier. This week, product designer Xinran Ma shares what he had to unlearn when he left his corporate product design job to work for himself. He explains how moving faster and experimenting helped him build a point of view without abandoning rigor or judgment.—Kate Lee

Was this newsletter forwarded to you? Sign up to get it in your inbox.

I was ready to start my own business. But for six years, I couldn’t.

I came to the United States to study architecture at Columbia University and spent a few years in the field, designing art centers and multi-family residential buildings. When I moved into product design, I found the work much broader and more fulfilling. It spanned research and information architecture, visual design and interaction design. It brought me closer to customers and the business, and I could see the immediate impact of my decisions in a way I couldn’t with architecture.

The pandemic, however, exposed the risk of relying on a single employer. When a coworker lost their job a few months before having a baby, I began thinking seriously about building something of my own—a second source of security for my family and an investment in myself.

There was one obstacle: My work visa prevented me from earning income outside my full-time job. So I learned the building blocks of running a business instead: I studied audience building, copywriting, marketing, and self-publishing. I was planning for the day when I would finally have the freedom to pursue my own path.

When I got my green card, I started with side projects while I still had a demanding full-time job. I published three books about building a career in product design, then started my Substack, Design with AI. I treated each project as an experiment and a way to follow my curiosity. Eventually I left my corporate job to focus on the business; today, Design with AI has more than 44,000 subscribers, and my course, AI for Product Designers, has become another part of how I write, speak, and teach about AI.

I learned a great deal in the corporate world. English is not my native language, and working mostly remotely forced me to articulate design decisions and write with clarity. I developed product sense, intuition, rigor, professionalism, and the habit of looking closely at data—skills I still rely on with clients, students, collaborators, and in managing my business.

But there were also habits I had to unlearn. Running my own business made me separate the habits that improved my work from ones I followed simply because they were familiar. The better habits, I’ve found, are ones that benefit anyone working with AI tools—whether you’re working for yourself or at a large company.

Move before certainty

One of the first things I realized after leaving my corporate job was that there was no one above me to ask for permission. I had to grant it to myself and develop a stronger bias for action.

AI reinforces that lesson by changing the speed of execution. An idea no longer has to remain an abstract line of text. I can turn it into something visual and tangible, even if it is imperfect. Giving people something concrete to respond to helps them understand the idea and see that I can execute. It builds trust.

My newsletter began this way. In the winter of 2023, a friend who wasn’t a designer came over for dinner and showed me a custom GPT he had built. I had barely used AI tools and was still on the fence about their value. But seeing an application in action sparked my interest. A couple of months later, I started Design with AI as a way to learn how AI could be used practically in product design and to document what I discovered. I didn’t wait until I had a settled view—or until the tools were mature—to begin.

Moving before certainty also changed how I relate to data and business decisions. In corporate roles, we might spend two hours preparing for and sitting in a meeting to understand why a subscriber metric had moved. That rigor taught me how to analyze data, and I still track the performance of my newsletter.

On my own, though, there was no manager to tell me which collaboration to accept, what topic to write about, or when to raise my prices. At first, that freedom felt more exposing than liberating, but eventually, I came to see the same uncertainty as a benefit. I’ve learned to decide more quickly, trust my intuition, and take responsibility for my decisions. I no longer beat myself up over losing one subscriber or analyze something for the sake of analysis. If I don’t want to do something, I don’t. If something feels right, I move forward. I can use AI to move faster once I’ve chosen a direction. Choosing the direction is still my job.

Experiments build a point of view

When I began the newsletter, AI seemed useful for generating images with tools such as Midjourney, but its role in day-to-day product design was less clear. I tried early tools including Uizard, Jambot, and Wireframe Generator, looking for practical ways to solve my own design problems. Some of those products were acquired; others are barely mentioned now. Even when the tools disappeared, I learned to identify where the tool saved time or broke down, and whether it fit a real workflow. What I learned made it easier to recognize patterns when new tools appeared.

Today, I look at AI tools partly like a journalist. I test them because I create content and teach other designers. From that vantage point, I see a wide spectrum of AI adoption among designers. Some teams have no access to AI tools. Some use ChatGPT or Gemini for early brainstorming, but the rest of their workflow has barely changed. At the other end are designers working closer to production—prototyping with tools like Figma Make and Claude Design, or working with real code components in Cursor or Claude Code, collaborating more closely with engineers, and even submitting small, targeted code changes.

Many designers tell me their companies expect them to use AI, even though sometimes it’s still unclear where the tools fit into their work. And waiting for an employer to provide the perfect tools or an official workflow makes it harder to learn. That’s why I tell designers to start the way that I did: with side projects.

Side projects give you room to explore without waiting for everything to become polished. You don’t need to become an engineer; you can naturally discover tools and workflows around problems you actually have. A point of view built through hands-on experiments lasts longer than the individual tools or trends. Your perspective on AI can be optimistic, pessimistic, or somewhere in between, but let it come from your own encounters with the tools instead of from the general mood around them.

Give yourself permission

Corporate work taught me rigor, product judgment, clear communication, and how to explain design decisions. Working for myself has made me more conscious of keeping those strengths while loosening habits that delay action and experimentation, or offload my own judgment.

The lessons are not limited to people who leave their jobs. Salaried designers can create space to experiment, make ideas tangible, and develop a point of view even when their official workflow hasn’t changed. As AI becomes more powerful, that aspect of the solo mindset becomes increasingly valuable inside larger companies too. Companies need people who can delegate work and then judge the output, who have a point of view on how to improve their team’s processes.

For six years, my visa meant I had to wait before I could earn money on my own. Once the legal barrier disappeared, I realized how many other kinds of permission I was still waiting for. Going solo—and acting before I had all the answers—has been a practice of giving that permission to myself.

Xinran Ma is the writer behind the newsletter Design with AI, with over 44,000 subscribers. He has led talks and AI training at places including Microsoft, Columbia Business School, the City of Vancouver, Workday, Etsy, and Pratt Institute.

Sign up for Xinran’s Maven course, AI for Product Designers, and receive a 15% discount.

Thanks to Jack Cheng for editorial support.

Disclosure: Every receives a share of revenue from new Maven course enrollments made through this partnership. Maven helped connect us with instructors and suggested potential topics; Every retained full editorial control over what we published and how each piece was edited.