Bryan McNeil

The year 2026 has proven that AI-assisted development is here to stay, and that it's genuinely good at development. So where do professional developers fit into this new space?

You still can't get away from the basics. Developers need to know the foundational principles of coding, because with scope creep and shifting demands, a vibe coder gets lost in the enormous volume of code that AI produces. They end up at the whim of the AI, unable to fix problems themselves. Being able to get your hands dirty and write code manually is a must in professional development.

Too often in my own AI-assisted work, I watch the AI disregard basic principles like DRY (Don't Repeat Yourself). Over time this creates unnecessary complexity and bloats the codebase, which in turn bloats the context window. This makes the AI more expensive and less effective. This is why a developer has to understand the code completely.

It's the professional's job to hold the scope of the whole enterprise architecture in their head. AI only ever works in sessions with limited scope; it knows what things are within that small context window and no further. Being a professional AI developer means guiding the prompts, knowing which context actually matters for a given feature, and understanding how that feature connects to the bigger picture. The role becomes one of leadership steering the AI ship and crew in the right direction.

It's strange to say, but I suspect modern developers will end up writing only a small share of code by hand. AI is increasingly capable of building most modern business applications. So developers will need to adjust, by holding a strong vision of where a business is heading, along with all the technical requirements to build safe, secure, modern software.

Let me be clear, I'm not a snob who looks down on vibe coding. Hell, my portfolio was built with vibe coding, in some sense. The doors of development have opened to everyone, and that's exciting. If you have an app idea, you can start building it today; how far you get just depends on the scope of the project. Either way, it's a good thing, and I imagine younger developers will find ways to guide AI into their own personal styles.

Overall, I wanted to make my vision clear for how AI fits into my work going forward. It's tedious but important to develop in small, reviewable chunks and to read every line of code that AI writes. It's just as important to document everything and loop the team in on each feature, because features will be arriving fast. Without this discipline, code grows so quickly that no one understands what's happening in the system anymore. It becomes a black box of our own creation.

Patience will be a virtue in professional development. Taking the time to read and write properly is what will distinguish a professional from a beginner.