What is vibe coding? How creators are building software with no coding knowledge

> Vibe-coders might not visit this forum for example and might not be aware of a security exploit that must be fixed asap.

This is true of "regular" coders too, i've worked with/for extremely experienced programmers who, while knowing the security risks, implemented (or even worse, asked me to implement) flakey authentication systems, exposed databases, etc. I don't think this is an AI issue, it's a "care about your users" issue.

> is only going to increase the demand for professional services

Then shouldn't we be happy that more people are embracing vibe coding?

> Accenture, IBM and other big companies have employed vibe coders for years.

Having worked in a few codebases/projects by these companies (and some other "big name" consultancy companies in this same "space"), I cannot imagine copypasta from Claude/Copilot would be anything other than a massive improvement.

What companies have paid these firms to do is nothing short of appalling, and I'm not even referring to the notorious Hertz lawsuit, just the average run of the mill junk.

Wow, what a ridiculous load of tripe.

Java classes are no more enabling of shit code than any other language. Idiots can and do “vomit” code into pre-canned structures built for them in any language.

Non-programmers always could build functional software by hiring programmers, you know. And we all know how well that usually goes.

We're not talking about building an ecom store, though, are we?

We're talking about building useful tools. Small things, that work in a way that makes sense to the individual.

Something that could be churned out in a day or two by a competent dev, but is completely out of reach for Joe public.

That's where we're at.

That's a really good idea. I have a few tool ideas for the team I work with but I never felt like I had enough time between tasks to implement them

It may not be auditable in the same was a block-based workflow, but there's nothing stopping you from saying "hey, this code doesn't do what I want, please improve it" to the LLM.

The whole point about the article is that reasoning about the code isn't always required any more, reasoning about the output is enough.

The article itself doesn't have comments. On HN, there's 2 posts that are against vibe coding which could be seen as gatekeeping. What are you referring to when you say "I see many gatekeepers in the comments."?

I don’t see anyone “gatekeeping” but I do see warnings. I don’t know what you’re good at, to give a good example. But if you have a hobby or job, imagine the new guy coming in, knowing absolutely nothing and doing shit they technically can do, but shouldn’t be done. That’s what this is. Just like a new cashier can technically just not give correct change, everyone knows this will end badly; eventually, even though everything looks fine from the outside.

That’s what AI code tends to do. It looks fine to the outsiders, it might even function correctly MOST of the time. Programmers are trained, mostly through experience but also classically, to think through every exceptional case and figure out how to handle it (or return an error so a human can handle it).

That isn’t to say this is ALWAYS the case with AI code, but rather it TENDS to be the case. YMMV, which is where these warnings are coming from.

I'm not sure it's explicitly gatekeeping, more a lack of perspective.

The benefit and excitement is most felt by people with little to no experience writing code themselves. The fear seems to come from building.. what, code for critical infrastructure? That's just not what we're talking about here.

> I dont want them safety critical

Then hopefully they run on a computer without network access, otherwise everything is safety critical.

No gatekeeping. Anyone can write programs.

But to build something which could handle a customer’s credit card, password, or other PII and charge them for it, you better know what you’re doing.

It’s all fun and games until you’re the cause of someone’s identity or password getting stolen.

Anyone can use CAD software, but if you’re designing a public space, you better know something about safety.

> I just want a free script or addon that will do its job instead of paying someone on upwork to do it for me. simple as.

So, you pay the billionaires to rent their graphics cards so you can avoid paying a normal person? -_-

>The agents will be able to do maintainable code and safety critical code at some point in the future. It's not all gonna happen overnight.

Don't hold your breath. If the AI was good enough to do all that, "maintainability" would look very different from what it does today. What does it mean to be maintainable when an AI can completely rewrite the software in every iteration? If AI ever gets good enough to do all this, for safety-critical applications no less, probably 95% of the white collar jobs that exist today will be gone. There may also be robots to do 90% of all other work too.

What is important that it works for you. I think it is good that more users can solve their own problems.

There’s some gate keeping, but many of us are also worried about the inevitable future where we have to take over someone else’s “vibe coding” mess of code.

The vibe coding style translates to trying a lot of different prompts and small adjustments until it looks like it works. In the past these people copied from StackOverflow and poked at lines until it compiled and appeared to work, but that only gets you so far. Now those same people can bang away at an LLM assistant all day long and produce volumes of code that appear to kind of work.

I’m in another forum dedicated to programming careers. Every day there’s a new thread from someone asking how to deal with all of their junior employees spamming code review with obvious LLM generated code that they don’t even understand.

A lot of the defenses of vibe coding rely on the assumption that it’s in the hands of someone knowledgeable who only wants to save a little time for something inconsequential. That’s fine. What’s worrying is that vibe coding is being used as a replacement for understanding code for many juniors and lazy seniors across the industry as long as they think they can get away with it.

I think the issue is Hacker News users are so deep in the space that they're surrounded by AI noise at work every day and might even be forced to use it every day. So they've essentially accepted that it's a useful invention. There's this foundational, overwhelmingly positive assumption that goes unsaid, while 100% of their comment text is pure criticism and negativity. But in their brain, it comes across as maybe like 1% criticism and negativity and 99% positive lived experience. So they don't realize how bizarrely negative they're all sounding about a good thing. Of course, there are some people do really completely hate AI coding, but I don't think the general sentiment one gets from reading hacker news comments literally is accurate. It has to be read in context. At my work, people prefix their critical comments with "nit -" to reassure the reader that it's overall good.

You're 1000% underestimating the experience of learning it yourself vs just getting info shot right into your brain.

> get over the bump

You call learning, making mistakes and fixing them, and improving "a bump"? That's the whole point.

> That’s necessary for learning

You haven't learned anything in the end. I read a lot of programming books in the past thinking I would be a computer god at the end, and I realized I learned nothing because "I did nothing" exactly like what we have with ChatGPT.

> evolving their skills from there

What skills? If you are just asking a computer for what you want you're not developing any skills, apart from maybe how to describe your requirements better†.

If you take the code the LLM outputs and use that as a basis to be able to write your own code I would call that "learning to program" and I applaud it whether you learn from adapting LLM code or by reading K&R cover-to-cover before you even touch a keyboard. But that's not what this article describes—what this article describes is the very deliberate act of not learning anything.

†Technically just describing your requirements in a way the particular LLM you're using responds best to, which is not necessarily "better" in an objective sense.

Everyone knows that copy pasting code without understanding it is exactly how you don't learn. Every programming class you're in has an underperforming student that does exactly this. Than they barely graduate, get interviewed, and they're somehow worse than the bootcamp grads.

Cursor is that, but prompted by you instead of a stackoverflow question.

That's a very valid concern. What kind of maintenance though? Most of the examples in this article are very simple. Maintenance might involve iterating on features, no real live services that would require any work.

Security patches though, yeah, that's tougher. My position is that if security is a concern, you need to hire someone. As much as many of these tools can integrate with databases and set up auth, I'm not sure how much I trust it personally. Especially if the actual code is hidden.

If AI does all the coding you are customer not a programmer.

Or do you think a customer who describes his requirements to a programmer who writes the actual program is a programmer too?

And APL! It saves so much tokens :D

Both in coding and arts it is still painfully obvious when something is just pure AI slop. At the same time there are both coders and artists that have learned to use these tools to create some truly unique and interesting things

Most people I know actively hate the AI art slop that has been flooding the internet. I would wager a lot of people do, evident when a common insult is to say a piece of art looks AI generated.

This reaction hasn't happened for AI programming. People just assume it will work because AI is made by programmers.

I would encourage you to try Replit!

Deployment is literally a button. Granted, to a replit subdomain. But following some documentation to add your own domain is not out of reach.

We did that for thousands of years and it’s still done in many parts of the world

i think that's called AGILE

The existing no/low-code has a more limited and controlled design space, and what the “coder” works with has deterministic behavior. With LLMs, they can do anything (including things you don’t want), and you don’t know how well your instructions map onto what they generate (only by testing, not by reasoning). When LLM-created designs run into unmaintainable territory, the corresponding code is invisible to the “coder”. This is unlike no/low-code, where even if the application becomes unmaintainable, the “coder” still has the full overview over their design. Vibes-coding is akin to write-only, you can’t tinker with your specification and see how the output changes; instead all previous output becomes part of the specification, an increasingly growing part that the “coder” doesn’t understand.

It could go so wrong, but the vectors for disaster are already innumerable in this world we live in.

I prefer to think about the ways it could go right. To empower people to create is the very purpose of computers, no?

> I think this idea, and AI in general, really scares a lot of professional programmers, but this is the world we live in now. You can't spend too long denying the impact of AI on writing code.

I am such a denier. Some colleagues in my department were allowed/instructed to try out AI for coding for months just so that the company could get substantive judgements whether AI is useful or not for coding work.

It basically did not help for the tasks that we are doing (and the tasks that we are doing are quite varied (so there are a lot of different topics where an AI could (purely hypothetically) help), but none of them is "create a scrappy CRUD prototype" :-) ).

Thus: the only impact that AI has on writing code that I don't deny is that a lot of stupid managers fall for the siren call of the various A"I" companies, and cause lots of damage. Of course none of these managers will at the end carry the can for the damage caused.

> You would not see software engineers at SpaceX or NASA "Vibe coding"

Sure that's true, but I doubt the people with no coding experience who use LLMs to create some application are working on safety critical systems. For me this fad should be embraced, especially by non coders, as it allows them to be creative in new ways.

No no, you see, there will be no more maintenance done by humans. The AI will rewrite the app completely from scratch every time you deploy.

The whole product will be defined by a specification distilled out of company meetings, slack messages and emails.

There will be only one button labeled "build and deploy" and if something breaks you press it again.

> It's just like the infinite monkey theorem, hitting the enter key repeatedly without understanding the actions and eventually reaching their goal.

Reminds me of this https://www.sciencedirect.com/science/article/pii/S277318632...

>Non-trivial text generation during the lifespan of our universe is almost certainly impossible.

Agreed! The similarities are stark. let's take the comparison further... Nvidia is the dotcom bubble's Cisco.

But the internet stuck around. It was still useful.

Just because something is overhyped, or doesn't meet the standards of the world's most demanding engineering organisations, doesn't mean that it's not here to stay and improve.

Perspective.

Is it propaganda if we can see it feel it and use it ourselves? I'm not so sure.

The hype is real, the bubble is real, but that doesn't mean it's meritless.

> Starmer

The UK recently announced plans to go big on AI even before the Trump meeting. Much to my dismay, as we've actually done a really good job of grid decarbonisation so far and there's no way building a ton of GPU-filled data centres is going to aid that or our spiralling energy costs.

Funny, that went away after corporations flooded the market with cheap inexperienced developers and decided they couldn't afford them either without ZIRP.

Still valid otherwise you can’t distinguish a working app from an app with security issues or even malware.

I've been dealing with a particularly vibed plumbling, luckily it's going away with the remodeling...

My previous homeowner was also a vibe maintainer! The difference that I see in this is that the LLM is reasonably at 'expert' level for many of these things. If I sit down with Stevie Wonder and ask him to help me write a song, the resulting song is probably going to be pretty good. Stevie also knows by experience lots of things to avoid, and is intuitively going to help me avoid them unless I instruct him specifically to make a poor choice.

I think there are likely opportunities too to have models or system prompts that cater or adapt to the experience level of the person it's working with. "As you interact with the user, determine their relative level of knowledge and experience. If they seem to be relatively inexperienced with software development, be much more aggressive in helping to warn them about and avoid common pitfalls, bad architectural decisions, and security issues."

I suspect it's probably going to enable a lot of poor quality stuff, but it also may to some degree raise the floor of what's being produced at the same time.

When I was in engineering school years ago, I had this phrase "faith-based engineering" to describe when my roommate would show me a stress-strain analyses of various parts of the electric racing rig he was working on, we'd look at it, think vaguely about what we know/learned, and say fuck it, build it, see what happens. (no one died, we were in college).

edit: english is hard for english speaker without tea

I found this hilarious XD

Yes but have you tried vibe plastering? It's like artex, but not intentional.

Jajajajaja. This is a good comparison.

> This is no different than what amateurs have been doing for a long time.

Yes it is. They now have a cohort, a drunk hallucinating expert.

>Hopefully they’re not taking in PII or charging anyone for these apps.

You bet they most definitely are. Also remember that the S in "vibe coding" stands for security.

Isn't this just another version of low coding except we are using AI?