I Finally Have a Hammer

I studied architecture, which means I spent years learning how buildings go together before I was ever responsible for one.

You learn the names of things first. Studs, joists, noggings, lintels, vapour barriers, flashing, footings. Then you learn sequence. You learn that a building is not an object. It is an agreement between materials, weather, gravity, trades, drawings, money, mistakes and time.

Later, in product, I realised I understood software in a similar way.

I could see what needed to happen. I could shape the brief, sketch the flow, understand the constraints, sequence the work, protect the intent, talk to the specialists and inspect the result. I knew when something felt wrong. I knew when the proportions were off. I knew when the solution had lost contact with the problem.

But I did not have a hammer.

I could point at the nail. I could explain why it needed to go there. I could brief someone with the right tools. Sometimes I could make a rough prototype with no-code tools, spreadsheets, wireframes or a bit of duct tape. But if I tried to drive the nail myself, I was mostly using my hand. It hurt. The result was not good enough.

Engineers had the nail gun. They still do.

That distinction matters. AI has not made me a senior engineer. It has not given me the same depth as someone who has spent years inside production systems, infrastructure, debugging, architecture and scale. The nail gun is still faster, stronger and more precise in expert hands.

But AI has given me a hammer.

The value of making contact with the material

For product and design people, this is a big shift.

The old version of the role was often one step removed from the material. You could think deeply about the problem, but the making happened somewhere else. That distance created useful separation, but it also created signal loss. The idea moved from customer conversation to notes, from notes to strategy, from strategy to brief, from brief to design, from design to engineering, and by the time it became real, something had usually changed.

Now a product person with enough taste, context and technical curiosity can make a working version of the thought.

Not a production system. Not always the final answer. But a serious thinking object. Something a customer can react to. Something an engineer can critique. Something that reveals where the idea is vague, where the workflow breaks, where the edge cases live and whether the thing has any life in it.

That changes the conversation.

A prototype is not only evidence that something can be built. It is a way of improving the decision before the expensive work begins.

Product judgement gets sharper when it can build

I have always loved the early phases of product work: the 0 to 1, and then the 1 to 10. Finding the valuable problem. Shaping the first useful version. Watching what works. Building the loops around it. Turning a small proof into something repeatable.

AI makes that surface area more accessible to people like me: product leaders, designers, strategists and commercially minded operators who understand the problem but historically needed someone else to translate that judgement into working software.

The best version of this is not product people bypassing engineers. It is product people bringing better artefacts into the room.

A working prototype beats a vague requirement. A rough workflow beats a hand-wavy strategy. A tested prompt chain beats an enthusiastic AI workshop. A small tool that touches real data beats a slide saying “automation opportunity”.

The point is contact. You learn faster when your thinking has to survive contact with the material.

This is why the role of the product person changes

As AI makes first-pass building cheaper, product work moves closer to craft again.

The product person still needs to understand customers, business value, adoption, positioning, trade-offs and the team’s capacity. But now they can also pick up the tool often enough to test their own judgement. They can move from “I think this should work” to “here is a version we can inspect” much faster.

That does not remove specialists. If anything, it should make specialists more valuable, because the conversation starts from a clearer object. Engineers can see the intended behaviour. Designers can see the interaction problem. Commercial teams can see the value proposition. Customers can respond to something concrete.

AI did not give everyone a nail gun. It did give many of us a hammer.

For product people who have spent years close to the work but not quite touching the material, that is a deeply exciting thing.