Ask HN: Why do the majority of vibecoded projects fail?
3 comments
Most vibe-coded projects fail because the person building them hasn't actually felt the pain of what they're building at scale.
You can vibe-code a chat app in 20 minutes because you've used chat apps for years and you know instinctively when something feels wrong but you've probably never watched your app crash under real users and had no idea where to even start looking as AI can only build what you already understand well enough to ask for.
If you know what you are doing are you still vibe coding?
A lot of these vibe coded projects even have AI respond to you on any questions or issues ie no one knows what’s going on.
A lot of these vibe coded projects even have AI respond to you on any questions or issues ie no one knows what’s going on.
Here's what I think:
I think AI can build scalable, distributed systems, too, you just have to know what you're doing.
to be slightly more substantial: there's a lot of "implicit" decisions that go into building your slack killer prototype (design, layout, user stories, UX) that you probably know instincitvely how you want it to be. because you want the glove that fits your hand, the HUD that fits your mind, etc.
that doesn't mean making a great UX is easy, it just means (i think) that lots of the implicit knowledge for doing so you've already absorbed through thousands of hours of experience with that end of systems. so you can do it, because you have that implicit knowledge via experience. maybe you even formalized some of it with explicit training (in design, etc).
the difficult thing about building scale and distributed systems is not the domain. it's that, most likely you don't have contact with that end of systems. you use it, but you don't interface with it. so you lack the experience necessary to build the implicit knowledge to make all those little decisions you don't even realized (probably) you're making when you vibe the thing into being.
if you've been a coder on big dist systems then you probably do have that experience. and, maybe you've even formalized some of that knowledge with training (in comp sci, etc).
it's not really that "it's so much harder to do" (lots of great systems engineeers can't design for crap, and they freely admit it! no shame required, just facts), it's just that without training or experience most lack the skills to be able to create scale.
tho, if you know what you're doing, AI can surely build it.
what do y'all think?