Vibe coding may deliver a useful prototype but it's not a way to build a sustainable product. Technical skills remain key to succeed with digital product development.
Had an episode recently where a founder was succumbing to the pressure of building and shipping fast leveraging one of these vibe coding tools. The biggest issue I see is Scalability and functionality beyond the demo. A final product that people will use and love not just admire.
It's easy to show something that looks nice on the surface. But the game remains the same. We don't want people to say "it looks nice" and move on. We want them to stay with our product and thus, turn into customers.
By definition, that requires a long-term focus. Now, if anyone knows anything about maintaining software in the long run, they should immediately have all the warning lights flashing red.
And it's not because some artificial coding standards are not maintained. It's simply impossible to continue delivering value to customers in that manner.
"Dear AI model, add me a feature here" works neatly when we talk about the first, second, or third feature. However, by the time there are already a dozen of them, it's no longer a feasible strategy.
And experts who one would turn to for help would probably take more by fixing that than they would if they built the thing the right way from scratch.
Oh, and yes, I've seen the exact equivalent of such development processes pre-AI. The whole mess was introduced not because it was vibe-coded, but because there was so much pressure to make shortcuts that eventually, any progress ground to a halt.
It's nothing new. We just have better tools to shoot ourselves in the foot.
Thank you for highlighting this!
Had an episode recently where a founder was succumbing to the pressure of building and shipping fast leveraging one of these vibe coding tools. The biggest issue I see is Scalability and functionality beyond the demo. A final product that people will use and love not just admire.
It's easy to show something that looks nice on the surface. But the game remains the same. We don't want people to say "it looks nice" and move on. We want them to stay with our product and thus, turn into customers.
By definition, that requires a long-term focus. Now, if anyone knows anything about maintaining software in the long run, they should immediately have all the warning lights flashing red.
And it's not because some artificial coding standards are not maintained. It's simply impossible to continue delivering value to customers in that manner.
"Dear AI model, add me a feature here" works neatly when we talk about the first, second, or third feature. However, by the time there are already a dozen of them, it's no longer a feasible strategy.
And experts who one would turn to for help would probably take more by fixing that than they would if they built the thing the right way from scratch.
Oh, and yes, I've seen the exact equivalent of such development processes pre-AI. The whole mess was introduced not because it was vibe-coded, but because there was so much pressure to make shortcuts that eventually, any progress ground to a halt.
It's nothing new. We just have better tools to shoot ourselves in the foot.
💯