Vibe code is legacy code • Hacker News Board

Re: Vibe code is legacy code

#311

Post by ZeroGravitas » Thu, Jul 31, 2025, 8:23 AM UTC

raincole wrote: post #170

bdangubic wrote: post #138

I have been SWE for almost three decades now - have read all the comments in this post and almost every negative comment about vibe coding holds true about almost every single ‘human-coded’ codebase I’ve ever seen ( some notable exceptions of course :) )

Yeah since when prototypes built to throwaway are a bad thing? They're arguably the most important step to build a business. Legacy code isn't a bad thing either. The majority of code that actually generates revenue right now is probably considered "legacy" by the devs working there.

The article is very pro-prototypes to throw away.

I've also, in pre-llm days, seen warnings to not show visually polished prototypes to non-technical customers/salespeople/managers because they have no way of judging the quality and will try to force you to use the prototype immediately because "look, it's done".

Re: Vibe code is legacy code

#312

Post by lan321 » Thu, Jul 31, 2025, 8:24 AM UTC

pishpash wrote: post #278

stouset wrote: post #267

Earlier quoted context omitted.

The difference is we didn’t know any better back then. We do now.

Surprisingly, 80's and 90's developers were quite skilled low-level developers who knew very well all the ways things could go wrong. The difference was the stakes were not high then. The blast radius was maybe a hundred thousand people and the worst was they lost their own files. Now some AI-controlled process or apparatus could ruin everyone's credit and maybe even kill you and all your neighbors.

The whole comparison is apples to oranges. Products are massive nowadays and a whole lot more connected.

Re: Vibe code is legacy code

#313

Post by WJW » Thu, Jul 31, 2025, 8:25 AM UTC

eddythompson80 wrote: post #120

WD-42 wrote: post #110

Earlier quoted context omitted.

I would not default to assuming it was his competitors, that sounds like scapegoating to deflect responsibility. What most likely happened is his site was scanned by one of the increasingly sophisticated exploit crawlers (anyone who runs an internet facing site and can view traffic knows what I'm talking about). His site got flagged as vulnerable, the hacker found out it was built like swiss cheese and had fun with i…

Who else would spend the time and effort to figure out you leaked your stipe key to your front end? Sure people have bots to do that, but it’s kinda unbelievable someone would run such a bot on their vibe coded website.

You don't scan just a single website, you code up the bot once and then scan every site you can find.

Your friends' service was just the proverbial paper car in a thunderstorm: the thunderstorm doesn't care about the paper car but destroys it just the same.

Re: Vibe code is legacy code

#314

Post by danielbln » Thu, Jul 31, 2025, 8:31 AM UTC

sysmax wrote: post #241

There's a pretty good sweet spot in between vibe coding and manual coding. You still think out all the classes, algorithms, complexities in your head, but then instead of writing code by hand, use short prompts like "encapsulate X and Y in a nested class + create a dictionary where key is A+B". This saves a ton of repetitive manual work, while the results are pretty indistinguishable from doing all the legwork yourse…

I'm still finding the right sweet spot personally. I would love to only think in architecture, features and interfaces and algorithms, and leave the class and function design fully to the LLM. At this point this almost works, but requires handholding and some retroactive cleanup. I still do it because it's necessary, but I grow increasingly tired of having to think to close to the code level, as I see it more and more as an implementation detail. (in before the code maximalists get in my case).

Re: Vibe code is legacy code

#315

Post by azath92 » Thu, Jul 31, 2025, 8:57 AM UTC

I often find that the hard part of writing big, persistant, code is not the writing but the building of a mental model (what the author calls "theory building"). This challenge multiplies when you are working with old code, or code others is working on.

Much of my mental space is spent building and updating the mental model. This changing of my mental model might look like building a better understanding of something i had glossed over in the past, or something that had been changed by someone else. Or it is i think the fundamental first step before actually changing any lines of code, you have to at least have an idea for how you want the mental model to change, and then make the code match that intended change. Same for debugging, finding a mismatch between your mental model and the reality as represented by the code.

And at the end of the day, AI coding tools can help with the actual writing, but the naive vibecoding approach as noted is that they avoid having to have a mental model at all. This is a falacy. They work best when you do have a mental model that is good enough to pass good context to them, and they are most useful when you carefully align their work with your model over time, or use them to explore/understand the code and build your mental model better/faster/with greater clarity.

Re: Vibe code is legacy code

#316

Post by MrGilbert » Thu, Jul 31, 2025, 8:57 AM UTC

NitpickLawyer wrote: post #310

MrGilbert wrote: post #305

Earlier quoted context omitted.

From personal experience, I'd like to add "I don't know what I'm doing, but LLM helps me pretending that I do" coding. And yes, that code ended up in production and caused issues. It was coded outside of the development department. The productive discussion left the chat some shareholder rounds ago.

Is it much more different (in quality) to "I don't know what I'm doing, but [SO/learn x in 24h bootcamp] helps me pretend that I do"? I guess I see your point in quantity. I could see how this would be more widespread than c/p from SO. And the results would "look" better crafted at a glance, but might explode later on.

> Is it much more different (in quality) to "I don't know what I'm doing, but [SO/learn x in 24h bootcamp] helps me pretend that I do"?

I would argue "yes" - it is far more easier to pretend with LLM, because the loops are way more faster. When copy & pasting from StackOverflow, you need to, at least, gain some type of understanding from what you are doing. There is no one to help you. With an LLM, you can keep on asking questions - if you hit a roadblock, you scrap the last 10 iterations and start over, because it is so fast.

Re: Vibe code is legacy code

#317

Post by usui » Thu, Jul 31, 2025, 9:02 AM UTC

phplovesong wrote: post #295

Did anyone think otherwise? AI generated code is obviously tech debt / legacy / whatever you want to call it. This was the same as 10-15 years ago when we saw that php dev copypasta from stackoverflow. I expect code quality to be way, way worse in the coming decade. Its not legacy because of age, its legacy feom day one. Pure nightmare fuel.

What are you referring to with PHP dev copypasta from Stack Overflow?

Re: Vibe code is legacy code

#318

Post by lonelyasacloud » Thu, Jul 31, 2025, 9:06 AM UTC

tux1968 wrote: post #28

This has happened before. Non-technical or junior people developed and deployed applications, emboldened by the relative ease of Microsoft Access and Excel. There were all kinds of limitations, scaling problems, and maintenance nightmares. But there were a lot of upsides too, and it made the "professionals" up their game to obviate the need for such adhoc and unsanctioned developments. Come to think of it, the exact…

Almost managed to forget the horror of all those enterprise "web" pages that worked just as long as the user was on a Windows machine with ActiveX ...

Re: Vibe code is legacy code

#319

Post by chrisrogers » Thu, Jul 31, 2025, 9:12 AM UTC

AlecSchueler wrote: post #304

QuiDortDine wrote: post #254

Earlier quoted context omitted.

I see that you're not alone in your position clearly, but still, this is such a strange take to me. Do people not seriously not see, nay, instinctively understand the ontological difference between the difference between using code someone no longer understands and deploying code no one ever understood? I'm not saying the code should be up to any specific standards, just that someone should know what's going on.

I don't actually see the difference. If someone writes the code and understands it but then becomes unavailable, what's the practical difference with no one having understood it?

One of the implementations underwent analysis.

Re: Vibe code is legacy code

#320

Post by rsynnott » Thu, Jul 31, 2025, 9:12 AM UTC

atleastoptimal wrote: post #268

doesnt_know wrote: post #119

Earlier quoted context omitted.

The fact your friend is suffering no consequences and is able to just carry on is exactly what is wrong with this industry. In a perfect world the creation of software would have been locked down like other engineering fields, with developers and companies suffering legal consequences for exposing customer information.

Let's say it was coded extremely well, but nevertheless a more advanced exploiter wreaked similar havoc. Would they still be liable in your perfect world? To some degree the principle of caveat emptor should apply in some tiny, nascent business, otherwise only large juggernaut monopolistic incumbents would have the means to have any stake in software.

Imagine these two scenarios:

Your spouse dies in surgery. The highly experienced surgeon made a mistake, because, realistically, everyone makes mistakes sometimes.

Your spouse dies in surgery. The hospital handed a passing five year old a scalpel to see what would happen.

There's a clear difference; neither are _great_, but someone's probably going to jail for the second one.

In real, regulated professions, no-one's expecting absolute perfection, but you're not allowed to be negligent. Of course, 'software engineer' is (generally) _not_ a real, regulated profession. And vibe-coding idiot 'founder' certainly isn't.