There's been a lot of chatter lately about vibe coding—that intuitive, creative approach to programming where you feel your way through the logic. Some call it chaotic. Others call it genius. But the debate rages on: Is vibe coding "real" programming?
Dylan Beattie's recent article, "The Problem with Vibe Coding", summarises the difference between vibe coding and product development.
But what does the future look like when GenAI becomes an everyday part of our dev toolkit?
Are GenAI Tools lowering our coding standards?
It's a fair question. But here's the thing—Copilot and similar tools aren't doing anything that wasn't already happening. Developers have been grabbing snippets off Stack Overflow and GitHub for years. AI is just streamlining the process, embedding the power of search and suggestion directly into our IDEs.
The problem isn't the tool—it's how we use it. If you're accepting every GenAI response without reviewing or understanding it, are you really coding? Or are you prompt engineering?
I'd argue the latter. And maybe that's okay—for some roles. But if you want to call yourself a developer, you've got to own the code. Learn from it. Review it. Otherwise, you're not writing software—you're copy-pasting at scale.
I've seen Copilot help people write cleaner code, fix typos, spot bugs, and even adhere to team standards. Tools like Codacy's Guardrails are evolving fast, too. They analyse the code generated by AI in real time and ensure that it adheres to coding and security standards.
File Structure: Does It Still Matter?
With AI reading and understanding code across entire codebases, maybe we don't need structured folders or organized files anymore. One giant monolith? No problem—AI will figure it out, right?
Sure. But what about the humans?
MIT's Communication Lab said it best:
"A successful file structure organizes your data and code with the goal of repeatability... It communicates the function and purpose of elements within a project by separating concerns into a hierarchy."
Structure isn't for the computer—it's for us, even in C#, where file structure affects namespaces, the human readability aspect matters. When a namespace appears in logs or error messages, it gives us context. Remove that structure, and good luck debugging without AI holding your hand.
Naming Conventions: Can AI Fix the Shouting
We rely on naming conventions and code styles to make sense of sprawling codebases. However, not all conventions are helpful to everyone. For instance, UPPERCASE constants might help some, but can be a problem for people with dyslexia. For others, it’s like the code is yelling at you.
In the future, AI could personalize these preferences. Maybe you prefer CamelCase or snake_case. Maybe I write “colour,” and you write “color.” AI could localize and translate those preferences in real time, like subtitles for your code.
Sounds perfect.
Except… collaboration becomes a nightmare if everyone sees a different version of the same code. Consistency still matters when we’re working together. So while personalization is cool, shared understanding trumps individual style.
Real-Time Code Feedback: A Glimpse Into the Future
Here’s an idea I love—credit Jonny Muir for this.
What if your IDE were connected directly to your production environment? You start typing, and the system tells you right away: “Hey, that change just improved your page load time by 12%,” or “That refactor bumped memory usage by 30%—maybe rethink it.”
A feedback loop that is so tight reshapes how we think about development.
So… What’s Next?
GenAI isn’t going anywhere. And we’re only scratching the surface of its capabilities. But the tools we use should amplify our skills—not replace them. Code should remain something we understand, maintain, and shape with intention—not just accept at face value.
The future of coding won’t be defined by AI alone—but by how we as humans choose to work with it.