To vibe-code or not vibe-code?

That is THE question every dev is talking about right now, and new buzzwords (vibe-code, IA generated code, low-code, no-code) are fighting for attention on all our channels.

Apart from the usual debates, coding with AI got me thinking about time. It’s a shift in how we app-makers divide our time.

AI is an incredible accelerator for the first steps of coding, making it much faster to show a prototype, a demo or an MVP.

It extends the time we can allocate to conception, strategy and architecture, which is a great relief.

It’s a shift in how we app-makers divide our time.

But it also extends the time we need for testing and debugging, sometimes leading to unbelievable conversations with our favorite AI colleague.

I wish we stopped believing that AI makes it faster to code.

It accelerates the first steps of coding, grants us more time for architecture and conception, but also requires extra time for the next steps of coding, especially debugging.

Thinking otherwise is simply considering coding and engineering as execution, which I find simplistic, inadequate and disrespectful.

Has AI changed the way you allocate your time when developing a product?