Is Generative AI Making Product Development More Chaotic?
Why the "build to learn" philosophy is your secret weapon in the age of AI-fueled creation
Is the classic 0 to 1 product discovery diagram still accurate?
I've seen the diagram. You probably have too. The one that neatly depicts the product discovery journey as a transition from squiggly chaos to a perfectly straight line. It looks something like this:
The problem? This idealized view suggests two dangerous misconceptions:
That you'll eventually figure everything out.
That once you think you've figured it out, it's smooth sailing.
Been there, done that, got the T-shirts (these days we actually celebrate with neon signs). My experience building products, especially in the rapidly evolving world of Gen AI, has taught me a very different lesson.
The Real 0 to 1 Journey: Buckle Up
Here's what the journey actually looks like:
Notice a few key differences?
Conviction Amidst Uncertainty: Even when I'm neck-deep in the "unknown," I've learned to develop enough conviction to make a call. I decide what to build, fully aware that my mind will change later. That's not a bug; it's a feature!
The Inevitable Curveball: Just when you think you've cracked the code...BAM! Something unexpected throws you right back into the uncertainty zone. It's humbling, frustrating, and utterly exhilarating.
The Perils of Endless Research: Analysis Paralysis
I've been asked, "Why not just keep researching until you have a perfectly polished product spec?" Wouldn't that prevent wasted effort?
While I'm a HUGE advocate for upfront research and strategic product thinking, there's a crucial turning point:
Important considerations for reaching this point:
You want to get to this point with as little elapsed time as possible.
The experience you gain by actually building products will help get you here faster.
Building a strong product intuition will also help get you here faster.
The path you choose at this juncture will define your success.
Why "Building to Learn" is Non-Negotiable
I'm firmly in the "Green Path" camp. No amount of research, interviews, or analysis can replace the invaluable learning that comes from building. Especially in the Gen AI space, iteration isn't just expected – it's essential.
"Building to learn" in the Gen AI era isn't just about user feedback. It's about:
Mastering the art of prompt engineering.
Pushing the boundaries of what's technologically feasible.
Uncovering entirely new User Experience (UX) patterns.
We're not just learning what users want; we're discovering what's possible.
Embrace the Messy Beginning (It's Worth It!)
Okay, let's be real. That "messy beginning"? It's hard. Uncertainty is a deeply uncomfortable place to be. You'll question everything: Am I thinking big enough? Too big?
I often advise teams to build for where the models and technology will be, not where they are today. But it's a delicate balance. You need concrete, achievable milestones to validate that your vision is both possible and addresses real user needs.
And those needs? They're evolving at warp speed, alongside the technology. We're in a period of unprecedented advancement, which means constant change. Agility, awareness, and a forward-thinking mindset are your best weapons.
You need to find certainty in an uncertain world. At the very least, cultivate unwavering conviction in the bets you choose to make. And remember, many of these are bets. But as Wayne Gretzky famously said, "you miss 100% of the shots you don't take."
So, start building. Start learning. Start honing that product intuition. The future of Gen AI is being written now, and it's a messy, beautiful, exhilarating adventure. Are you in?
You convinced me. It's high time I start building.
You hit right on the pulse of what many including myself have experienced. Great article.