A founder asked me a question earlier today that stopped me in my tracks.
He asked something like: “What is the point of customer discovery if I can just spin up an app in Cursor (or the like), get immediate customer feedback, and iterate from there?”
It’s a fair question, and one that seems to be on everyone's mind lately. It immediately reminded me of a lively discussion Micah Rosenbloom (Partner at Founder Collective) kicked off on LinkedIn a few weeks ago about "When Starting Gets Easy."
Micah’s premise was that AI tools are collapsing the cost and effort required to launch software. But as I noted in the comments of that post, easier isn't always better.
When starting was harder, good founders pushed through the friction and did real discovery, experimentation, validation, and iteration before building much. This resulted in more focused and resilient startups. Lower barriers have opened opportunities for more people, but they’ve also amplified the noise and let unvalidated ideas pass for traction.
Spinning up a product over a weekend and getting quick users is just the new vanity metric.
The "Build First" Trap
This isn't a new phenomenon. AI has just accelerated it. I was recently discussing this with my friend and colleague Bill Aulet at MIT. It reminded me of his prescient TechCrunch article from back in 2014, "Our Dangerous Obsession with the MVP."
Bill argued then, and it is even truer now, that if you blindly build an MVP without understanding the customer, you are just guessing. The only difference today is that you can guess faster.
My answer to the student was simple: If anyone can build it that fast, what is your advantage?
If code is a commodity, then the product itself is no longer your moat. Your competitive advantage isn't how fast you can build. It's how deeply you understand the customer and their problem. Without that understanding, you might have a working product, but you won’t have a go-to-market strategy. You won't know exactly who your customer is, where to find them, or why they buy.
Speed vs. Direction
In our second book, Testing with Humans, Giff Constable summed this up perfectly: "After all, if you run really fast in the wrong direction, you are further from your goal."
AI tools act like a Formula 1 engine. They give you incredible speed. But customer discovery is your GPS. If you don't know where you're going, a faster engine just helps you drive into a wall sooner.
The 'build first, ask later' approach relies on trial and error. You launch and get feedback, but it’s often just usability feedback, not value validation. You are testing features, not assumptions. You end up asking 'what features do you want?' which leads to the classic 'faster horse' trap. You build a bloated roadmap that mimics competitors rather than solving the root problem. That is an expensive way to learn, even if the coding is cheap.
Insight is the New Unfair Advantage
This environment is here to stay. If you are just building a side hustle, trial and error might be fine. But if you are building a venture-scalable business, real customer discovery matters more than ever. Investors are looking for the signal of proprietary insight, not just the noise of a generic MVP.
You need to know something your competitors don't. You need to uncover a nuance about the customer's workflow, a specific frustration with the status quo, or a willingness to pay that isn't obvious on the surface.
You don't get that by looking at analytics on a dashboard. You get that by getting out of the building (or jumping on a Zoom) and having a real conversation.
So, by all means, use the AI tools. Build faster. But don't confuse motion with progress. Before you fire up the code editor, go talk to some humans. Find out if the road you're speeding down actually leads to a treasure.
