For decades, building technology was expensive and slow. If you wanted a good looking, high quality product that scaled, you needed senior architects, dedicated designers, and niche engineers.
I remember a few years back, before today’s wave of powerful AI tools, I worked on a project with a complex calculation backend in a very niche field. The problem was fascinating, but the development process? Painfully slow and costly. We needed specialists for every step, and progress always felt like it was dragging through mud.
Of course, there was no other option at the time. Looking back, I’d approach it differently today. With the tools we have now, I’d focus on iterative development: building quickly, testing with users, and fine-tuning the product in short cycles. Back then, too much energy went into never-ending brainstorm meetings and vague development phases that stretched on without delivering tangible results.
And to be clear: this doesn’t mean vibe coding or letting quality slip. You might still spend the same total amount of time, but the output ends up dramatically better. Why? Because the complexity is lower. The system is understood by the whole team, not just locked away in the heads of special experts. A modification that once required weeks of coordination and niche expertise can now be tested and shipped in a matter of hours or days.
Conclusion: build smarter
The lesson isn’t “ditch expertise” or “let AI do the work.” Expertise is still essential, but how we apply it has changed. Think about the shift from postal mail to email: one person can now manage far more communication, respond faster, and move projects ahead without waiting days for a single reply.
The same principle applies to development. AI speeds up execution, reduces bottlenecks, and lets your team achieve far more without sacrificing quality.
The winners won’t be those who just “vibe code” with AI. They’ll be the ones who combine solid expertise with modern tools, iterating quickly, cutting complexity, and keeping their eyes on the bigger picture.