In the late 1990s, the internet was still taking shape and countless people weren’t building businesses, they were speculating. A domain name cost ten dollars, and the upside seemed infinite. People bought pets.com and business.com hoping someone, someday, would pay a fortune for the digital real estate.
Most of those domains sat idle for years, generating nothing but small renewal fees and large unrealized dreams.
Fast forward to today, and we’re watching the same dynamic unfold again—except this time the speculative asset isn’t a URL. It’s a fully functioning software product created through natural-language prompts and AI-assisted development.
A single weekend, a loosely defined idea, and a willingness to experiment are now enough to ship something that looks like a real application. No engineering degree required. No complex stack decisions. The barrier to building software just collapsed.
And just like the domain boom, speculation is racing far ahead of true utility.
The Real Lesson of the Domain Era
Owning a domain name never created value on its own.
Building something people actually wanted to use did.
The same is true for AI-generated software. Today, generating code isn’t the hard part. The challenge—and the differentiator—is knowing what to build in the first place.
Probing vs. Producing
There’s a big difference between exploration and speculation. When I worked in Berkeley, there was a notoriously unappealing McDonald’s across the street from the office. When a group couldn’t decide where to eat, I’d suggest McDonald’s—not because I wanted to go, but because putting a concrete option on the table snaps people out of indecision. The suggestion triggers reactions, preferences, and better ideas.
That’s a productive probe. It reveals signal.
Spinning up a random app with vibe-coded prompts is not the same thing. It’s closer to buying a lottery ticket. You’re not learning what users need. You’re just discovering whether this particular version of an idea happens to resonate by chance.
The Advantage Now Lives Upstream
The most capable builders I know carry an internal catalog of prior attempts, failed models, market timing errors, and concepts that were ahead of their time but not wrong. They know the lineage of an idea before they commit time or resources to it.
That pattern recognition—not development speed—is becoming the real competitive edge.
As execution becomes democratized, the advantage shifts to:
- Ideation: Knowing which opportunities are real versus recycled.
- Validation: Testing assumptions before writing a line of code.
- Problem selection: Choosing challenges that genuinely merit solutions.
In this environment, speed without direction isn’t acceleration. It’s drift.
Where We Go From Here
Vibe coding will create an explosion of prototypes, products, and hopeful experiments—just like the rush to buy domains a generation ago. But sustained value will come from something deeper: clear insight into user needs, market gaps, and the patterns that distinguish fleeting novelty from enduring solutions.
The barrier to shipping software may be near zero, but the barrier to shipping something meaningful remains exactly where it has always been.
Knowing what to build is still the hardest part.
Most Recent Posts
Explore the latest innovation insights and trends with our recent blog posts.
