Less jamming, more building
Programming used to be a rhythm of flow and frustration. You build, you get stuck, you search endlessly on Google and Stack Overflow. Often just not the solution you need. Now you're asking AI: “This is my problem, how do I fix it?” and you can move on.
More and more often, I'm even going one step further: no “help me with this”, but just “make this”. Not only does that save time, it changes the way we build.
Speed is everything
For me, speed is crucial. That's why I prefer to work with Claude Code. Where Copilot sometimes feels slow and breaks your flow, Claude first thinks, makes a plan and checks if that's right. Only then does he write code. Less back and forth, more pace.
And trust me: when you're in the middle of your typing flow, waiting kills.
Context becomes the new skill
AI works just like a developer: the better the context, the better the output. So prompting is no longer a trick, but a profession. Explaining what you really want is a skill we all still need to grow into.
Sometimes that takes more time than tapping yourself. But that's part of it. It's part of the quest for a workflow where AI creates exactly what you want.
Low-code or full-code? AI chooses you
What I find interesting: AI changes the choices you make. Why bother clicking in low-code when AI generates a front end for you?
I often use Power Pages with Dataverse — a no-code database — and then I use an AI-generated React front-end to it. You don't have to build a backend anymore, Dataverse takes care of that. This way, you can combine low-code and full-code, and you can go from idea to application much faster.
The junior paradox
But there is also a downside. How do you learn to program when AI writes the code for you? How do you take responsibility for something you didn't make yourself?
That's why I believe that, as a junior developer, you should still be able to read and review code. Don't write every detail, but understand what's happening. When a customer asks, “Why does this script do this?” , you can't say, “I don't know, AI made it.”
Always experimenting
We are in the middle of a major experiment. No one knows exactly where this is headed. What I do know: curiosity remains your best tool.
I'm looking around, trying out, discovering new things like MCP servers (where AI doesn't just chat, but also controls tools for you). Sometimes it works, sometimes it doesn't. But you'll always learn something that makes your workflow better.
And in two years?
Do we still write code ourselves? I have no idea. AI is moving fast, sometimes faster than we can keep up with. Maybe we'll look back later and think: why did we do all that manually?
But today, we are in the middle of it. It's about trying out, struggling and learning. And that may make this profession more fun than ever.
Want to know more?
Want to know more about how we combine AI and code? At Blis Digital, human acuity and AI power merge to build software that is faster, smarter and future-proof.
Read here how we do that.
