Product Thinking: The Skill That Matters When Everyone Can Build

AI lowered the floor on building software and raised the ceiling on what's possible to build. The widening gap between them is product thinking. The skill that matters most now isn't coding. It's knowing what to build.

Last week, a client walked us through an app she'd built over a weekend. Login flow, Stripe integration, the core workflow. She'd used Claude and Cursor. It worked.

We asked one question: "What happens when a user gets halfway through and quits? Do you save progress? Notify someone? Let it expire?"

Silence.

Not because she didn't care. Because the question never came up. The tool didn't ask it. She didn't know to.

The floor and the ceiling

LinkedIn's 2026 Jobs on the Rise report tells a revealing story. Four of the top five fastest-growing roles are AI-related. AI Engineer is number one. AI Consultant is number two. But look at the pathways feeding into those roles: Product Manager appears again and again. The top transition paths into AI Consultant? Founder, Software Engineer, Product Manager.

Founders rank ninth on the list, with 69% year-over-year growth. The median founder has under six years of experience. Where are they coming from? Software Engineering, Product Management, Managing Director. The thread across all of these: people who learned to think about what to build, not just how.

The fastest-growing roles in the industry are being filled by people who know how to think about products. Not just people who know how to build them.

This makes sense once you see what AI actually changed. It lowered the cost of execution dramatically. Generating code, producing designs, assembling prototypes. The technical floor dropped to near zero.

But execution was never the hard part.

Brooks wrote that in 1975. It's truer now than when he said it. AI didn't change the fundamental challenge of software. It amplified it. When building is fast and cheap, the cost of building the wrong thing also drops... which means people build the wrong thing more often, more quickly, with more confidence than ever before.

The floor dropped. The ceiling rose. The distance between them is where the real skill gap lives.

The gap AI widens

Most conversations about the AI skill gap focus on technical chops. Can you prompt effectively? Do you understand agentic workflows? Can you ship?

Those matter. They're not the gap that separates products that work from products people actually want.

The real gap is product thinking. The ability to ask:

• Who is this for? Not "users." A specific person with a specific frustration in a specific context.
• What should this not do? Scope discipline gets harder when building is cheap. Every feature you can add becomes one you have to justify not adding.
• What does success look like? Not "people use it." A measurable outcome that proves the product solved what it set out to solve.
• What happens at the edges? The happy path always works. The edge cases are where products break and where real expertise reveals itself.

These aren't technical questions. They're judgment questions. AI tools don't ask them for you. The tool will build whatever you describe. It won't tell you whether what you described is worth building.

We wrote a whole post about the artifact where this thinking lives. The brief is the container. Product thinking is the skill.

Why this compounds

Product thinking compounds in a way that technical fluency alone doesn't.

Someone who understands their users deeply, who can define a crisp success metric, who knows where to draw scope boundaries... that person gets more leverage from AI, not less. The clearer the direction, the better the output. This is why Credo 7 resonates: "AI amplifies your direction, right or wrong. Aim matters more than ever."

The LinkedIn data bears this out. The fastest-growing roles aren't going to the best prompt writers. They're going to people who spent years developing judgment about what to build, for whom, and why. Product managers transitioning into AI consulting roles aren't succeeding because they picked up a new stack. They're succeeding because they already knew how to think about products, and AI made that thinking dramatically more valuable.

One observation from the report: "The gap between a founder and everyone else is smaller than ever. It often comes down to simply starting." Starting is the easy part now. Knowing where to aim when you start... that's what separates the 69% growth in founder roles from the graveyard of weekend prototypes that never found a user.

The widening distribution

We're heading toward a landscape with more builders than ever (the floor dropped) producing a wider range of outcomes than ever. Some brilliant. Most mediocre. The distribution isn't tightening. It's spreading.

The builders who develop product thinking will pull away. Not gradually. Exponentially. Every improvement in AI capability amplifies the advantage of knowing what to build. Better tools don't close the gap between thoughtful building and careless building. They widen it. The same gap is opening inside agentic products themselves: we wrote about how PMs should think about it in Find the ceiling before you set the floor.

The people who invest in this now (understanding users, defining scope, writing briefs, running a design partner program with real users, pressure-testing assumptions before writing a single line of code) are building the skill that matters more with every generation of AI. Not less.

The floor dropped. The ceiling rose. The skill that bridges the distance between them isn't coding.

It's knowing what to build.