> with vibe coding you can't cast aside the AI after you've taken the initial step
Sure you can, why wouldn't you? Good models these days spit out architecture and code fairly indistinguishable from your average developer.
It's always hard to jump into an existing project and understand it all, but it's doable and in 2026 it's no harder to jump into a vibed project than it is to jump into any other project (with exceptions in both directions, of course).
You still have the code and can switch back to "fully human" (or any other point on the spectrum of human-ai) any time you want by just opening the same editor you would use if you cloned any other project.
You can do this if you hold its hand a lot. Claude by default will do insane levels of abstraction for small things. Also make Python docstrings 3x longer than the actual code.
I think this is a severely outdated take that was true a few years ago, but has not been true for at least 1-2 years. Especially from the top models at the top labs.
Qualitative measurement is always hard, but plenty of studies have come out in the past year that generally find many models' code to be roughly equal to human code in terms of maintainability, complexity, readability, etc.
>Sure you can, why wouldn't you? Good models these days spit out architecture and code fairly indistinguishable from your average developer.
Good lord no, AI generated code isn't remotely close to something even the most insane human would generate. And if someone did you would seriously have to question what was going through their head. Everything from naming conventions to the flow of logic is twisted and often non-sensical. Plus the commenting and verbosity of code is ridiculous.Claude constantly makes up new terms to describe things that leave you wondering what it's talking about. Yesterday Claude started referring to some sort of logic as "finger-splint".
dimbletimbers · · focus · HN ↗
impendia · · focus · HN ↗
And I think that's one important difference: with vibe coding you can't cast aside the AI after you've taken the initial step.
ATMLOTTOBEER · · focus · HN ↗
drusepth · · focus · HN ↗
Sure you can, why wouldn't you? Good models these days spit out architecture and code fairly indistinguishable from your average developer.
It's always hard to jump into an existing project and understand it all, but it's doable and in 2026 it's no harder to jump into a vibed project than it is to jump into any other project (with exceptions in both directions, of course).
You still have the code and can switch back to "fully human" (or any other point on the spectrum of human-ai) any time you want by just opening the same editor you would use if you cloned any other project.
jdkoeck · · focus · HN ↗
Emphatically, no. It’s way too complex, every time. Ceinture et bretelles, as we say in French.
jldl805 · · focus · HN ↗
mahboi · · focus · HN ↗
NewJazz · · focus · HN ↗
graemep · · focus · HN ↗
jdkoeck · · focus · HN ↗
graemep · · focus · HN ↗
I do not know much French but that seems to translate literally into the common English phrase "belt and braces"
jdkoeck · · focus · HN ↗
delusional · · focus · HN ↗
drusepth · · focus · HN ↗
Qualitative measurement is always hard, but plenty of studies have come out in the past year that generally find many models' code to be roughly equal to human code in terms of maintainability, complexity, readability, etc.
hattmall · · focus · HN ↗
Good lord no, AI generated code isn't remotely close to something even the most insane human would generate. And if someone did you would seriously have to question what was going through their head. Everything from naming conventions to the flow of logic is twisted and often non-sensical. Plus the commenting and verbosity of code is ridiculous.Claude constantly makes up new terms to describe things that leave you wondering what it's talking about. Yesterday Claude started referring to some sort of logic as "finger-splint".
brazukadev · · focus · HN ↗
When it fail, I fix it so it doesn't have multiple patterns that I might not know.