Diving Deeper into "How to Start with AI?" by Learning From the Experts

I've been using AI to validate my thoughts and brainstorm ideas, which has significantly impacted my life. However, I haven't fully understood its effect on product development yet. A few weeks ago, I met a friend who was thrilled about how AI had transformed his work. He mentioned that, despite being the CTO of his company, he hadn't coded in years. He believed that rapid changes were imminent and that the next six months to a year would be the optimal time to leverage this technology. I didn't grasp his enthusiasm until I took Felix Haas’s course, Augment MBA — “Build a Business with AI” — and I also felt the excitement.

Felix is head of growth design at Lovable, the fastest-growing vibe coding tool on the market, and he built the course around one core idea: the barriers that used to stop non-technical people from building software are basically gone. What’s left is the harder, more interesting problem — deciding what to build, and thinking clearly enough to describe it well.

The Three Ways People Actually Start Building

One of the most useful frameworks in the course was Felix’s breakdown of how most people begin. Almost everyone goes through three phases: first they build a small prototype just to see if the tool works, then they solve an annoying problem at work with an internal tool, and eventually they realize they could build an entire product or business this way.

What I liked about this framing is that it takes the pressure off. You don’t need a fully-formed startup idea on day one. A simple experiment counts.

The Part Nobody Talks About: Prompting is Really Thinking

The tactical centerpiece of the course is prompt preparation — and Felix’s approach here is more deliberate than I expected. Before you open Lovable, he recommends using ChatGPT or another AI tool to pressure-test your idea by answering four questions: What am I building? Who is it for? Why would they use it? What’s the one action a user should take?

It sounds obvious, but working through those questions out loud exposed a lot of fuzzy assumptions I didn’t know I had. Felix makes the point that AI tools are essentially mirrors — when your prompt produces a confused output, it’s usually because your thinking was confused first.

What the Course Builds Toward

The second half of the course walks through an actual build in real time: a content planning tool Felix made for his own LinkedIn workflow. Watching him iterate — changing colors with plain English, adding features through conversation, reverting versions when something broke — made the process feel genuinely accessible in a way that reading about it doesn’t.

But the final unit is the one that stuck with me most. Felix argues that the real skill in this new era isn’t technical fluency — it’s knowing what’s worth building, and being able to articulate it clearly enough that the tools can respond. As foundation models improve and building gets cheaper, that judgment becomes the actual differentiator.

Who This Course is For

If you have product ideas but have always assumed execution required a technical co-founder, this course directly challenges that. Felix’s examples range from a non-technical growth marketer in Brazil who built a safety app to 450K ARR in under a year, to his own father building a property website in a single day with no coding background.

The course is hands-on by design — you’re meant to have Lovable open as you go. By the end, you’ve shipped something real, which is more than most courses can say.