I was thinking recently about how developers interact with AI tools, and I noticed something that goes deeper than just prompting techniques. It’s about mindset — specifically, the difference between being a software developer and being a solution developer.
The Distinction That Matters
Software developers focus on building software. Writing code. Crafting classes and functions. Solution developers think about developing solutions. It sounds like a subtle shift, but I think it’s a big deal.
Here’s the thing: a solution may or may not involve software. And ultimately, most people and most businesses don’t want software. They want solutions. They hire people to solve their problems, not to write code for them.
A Moment That Stuck With Me
I once worked on a team where a developer kept gold-plating his code — adding layers of abstraction, perfecting patterns, polishing things that didn’t need polishing. The team lead finally said: “We are here to sell shit. We are not here to write code.”
That landed hard. The nature of that business was selling products. Software was part of the solution to help them sell those products. But the software wasn’t the point. The sale was.
Starting with the Need
If we’re solving problems, the first question isn’t “What code should I write?” It’s “What is the need?” What’s preventing the company or the person from fulfilling that need? That’s the problem that needs solving. It may need a system, but not an automated system. Maybe not yet.
This shift in thinking changes how we approach everything — including how we work with AI.
How This Shows Up with AI
Developers who approach AI with a solution developer mindset are better able to deliver what they’re actually being paid to create.
When we think as a software developer, we typically prompt an AI tool to write code: “Implement this class.” “Build this feature.” If we’re prompting it to implement a user story described from a business or value perspective, that’s better. But the most common pattern I see is developers asking AI to just write code.
And that’s what it does. It writes code. If not given more guidance, it writes code based on the average of what it was trained on. Then many developers look at it and say, “I don’t like that. I can write better code than that.” And they stop there.
Codifying Your Approach
Developers who codify their opinions — their approach, their standards — into instructions for the AI will get results closer to what they need. They steer the AI, and the code gets better.
But as solution developers, we go beyond that. We give the AI the business context. The business problem. The business needs. Then we ask the AI to help us think through the solution. Not just implement it—think it through.
Then we ask the AI to help implement it. Test it. Deploy it. Monitor it. Improve it.
We’re treating the AI as a thinking partner in solving the problem, not just a code generator.
What I’m Noticing
The developers who are getting the most value from AI aren’t the ones who are best at writing prompts for code generation. They’re the ones who understand the problem they’re solving and can articulate the context clearly. And when they can’t, they use the tool to better understand it.
They’re solution developers first, software developers second.
And I think that’s always been true, even before AI. But AI makes it more visible. Because if you only ask for code, that’s all you’ll get. If you ask for help solving a problem, you might get something closer to what you actually need.
What problem are you solving today — and is code really the solution?
Leave a Reply