Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Vibe Coding is not an excuse for low-quality work

addyo.substack.com

Post reply on HN
Print view View original

141–150 of 213 posts

Re: Vibe Coding is not an excuse for low-quality work

#141

Post by legulere »

Workaccount2 wrote: post #138
ndepoel wrote: post #18

This all reminds me a lot of the early 2000's, when big corporations thought they could save a lot of money by outsourcing development work to low-income countries and have their expensive in-house engineers only write specifications. Turns out most of those outsourcing parties won't truly understand the core ideas behind the system you're trying to build, won't think outside the box and make corrections where necess…

Software is going to completely change. The writing is on the wall, it's just a matter of who can step back to see it. Giga-projects with 50k, 100k, 500K, lines of code are going to get decimated. By and large these programs are written to capture large audiences by offering a massive feature set. But how often is any one user ever actually needing those 100k LOC? If LLMs can start nailing 5K LOC projects pretty reli…

Complexity is going up rather than down. Companies already solve simple problems with spreadsheets.

Re: Vibe Coding is not an excuse for low-quality work

#142

Post by paradite »

ken47 wrote: post #134

The hype levels are so overwhelming that AI coding could never hope to meet them. I've tried having a highly-ranked AI coding app write unit tests for a relatively complex codebase. 80% of the generated test cases failed. But an experienced human such as myself could use those as a starting point since it took care of some of the tedious boilerplate. It genuinely saved me some time and annoyance, but could never hope…

I think you should try more tools and use cases.

Yes some of the current AI coding tools will fail at some use cases and tasks, but some others tools might give you good results. For example, Devin is pretty bad at some trivial frontend tasks in my testing, but cursor is way better.

I have good success with web development and JavaScript in particular, on several large codebases.

I also built my own AI coding and eval tools that suits my needs, though cursor has largely replaced the AI coding tool that I built.

Re: Vibe Coding is not an excuse for low-quality work

#143

Post by diggernet »

simonw wrote: post #129
diggernet wrote: post #108

Earlier quoted context omitted.

> Where did Andrej say it was "the future of software engineering"? He very clearly described vibe coding as an entertaining way to hack on throwaway weekend projects. ... And then a few weeks later, my boss' boss scolded the team for not having heard of the term, and told us to learn and use vibe coding because it's the future.

Your boss' boss clearly didn't read to the end of Andrej's tweet.

Yup. Or just ignored it because it didn't fit the predetermined narrative being pushed on us.

Re: Vibe Coding is not an excuse for low-quality work

#144

Post by fourside »

Workaccount2 wrote: post #138
ndepoel wrote: post #18

This all reminds me a lot of the early 2000's, when big corporations thought they could save a lot of money by outsourcing development work to low-income countries and have their expensive in-house engineers only write specifications. Turns out most of those outsourcing parties won't truly understand the core ideas behind the system you're trying to build, won't think outside the box and make corrections where necess…

Software is going to completely change. The writing is on the wall, it's just a matter of who can step back to see it. Giga-projects with 50k, 100k, 500K, lines of code are going to get decimated. By and large these programs are written to capture large audiences by offering a massive feature set. But how often is any one user ever actually needing those 100k LOC? If LLMs can start nailing 5K LOC projects pretty reli…

> it's just a matter of who can step back to see it

I’m not sure why those fully bought into the AI hype so often dismiss anyone who’s less bullish as simply too small-minded to see the writing on the wall.

That aside, while I do think software is going to change, I’m not sure that I agree with your particular version. What examples of apps with 100k, 500k+ LoC codebases are going to get decimated? The biggest successes in software today have moats around inventory or network effects. Are people going to make their own Netflix or Uber? Even at a smaller level, is the idea that local gym owners are going to replace gym management software with their own mini apps? Or restaurants are going to do their own point of sales apps? Unless AI can make the cost of maintaining software something close to zero time, which is a really tall order, why would business owners waste time on something outside of their core business. And if this is such an untapped opportunity why didn’t we see the start of it with the no-code movement?

Will a big chunk of the software market fall off and be replaced by custom mini apps that the layperson crafts with AI? Maybe? But I don’t see how one justifies the level of confidence I see from these predictions.

Re: Vibe Coding is not an excuse for low-quality work

#145

Post by danielbln »

xnx wrote: post #56
SCdF wrote: post #5

So what have we redefined vibe coding to mean exactly? The original tweet[1] talked very specifically about not caring about quality, just accepting whatever code the AI produces blindly, as long as you get the black box output you're looking for, and just randomly try again if you didn't. Are people now using this term to mean "giving an AI agent broad tasks"? [1] https://x.com/karpathy/status/1886192184808149383?la…

Vibe coder - someone who uses more coding assistance than I do

It's like driving. Someone doing more vibe coding then me is a maniac and irresponsible, someone doing less is a slow control freak.

Re: Vibe Coding is not an excuse for low-quality work

#146

Post by dijksterhuis »

simonw wrote: post #131
achierius wrote: post #109

Earlier quoted context omitted.

> Saying that Andrej Karpathy is "an AI researcher, but not a software engineer" isn't a very credible statement. Broadly I agree with the two other replies: his 'day job' has not been coding in some time, so I would put him in the same bucket as e.g. a manager who got promoted out of writing code 5-10 years ago. I do want to understand where you're coming from here -- do you think that's a fair characterization?

His GitHub contributions graph (1,261 contributions in 2024) looks pretty healthy to me. https://github.com/karpathy He spends a lot of time as an educator these days.

his commit messages leave something to be desired

> Merge pull request #740 from karpathy/gordicaleksa-fix_dataloader2

> - fix tokenizer omg

> - attempt to fix PR

> - Merge branch 'fix_dataloader2' of https://github.com/gordicaleksa/llm.c into gordicaleksa-fix_dataloader2

Re: Vibe Coding is not an excuse for low-quality work

#147

Post by ehnto »

Workaccount2 wrote: post #136
ken47 wrote: post #134

The hype levels are so overwhelming that AI coding could never hope to meet them. I've tried having a highly-ranked AI coding app write unit tests for a relatively complex codebase. 80% of the generated test cases failed. But an experienced human such as myself could use those as a starting point since it took care of some of the tedious boilerplate. It genuinely saved me some time and annoyance, but could never hope…

AI is to SWEs what self-driving cars (or like features on new cars) are to taxi drivers. A neat thing that may make some basic things a little easier. But it's the wrong perspective. Think instead what a self driving car is to someone who cannot drive. It doesn't matter if it sometimes disengages to have an operator step in or can only handle basic local trips. It's a total game changer. AI is unlocking computers for…

> AI is unlocking computers for people who have never written a line of code in their life.

I don't think I can dispute the claim, but it feels more like someone who can't build a house being able to do it now that they have YouTube tutorials. Unless the person was already quite smart and competent, the house will probably have significant structural issues.

Is that a bad thing? For housing, governments the world over seem to agree that it is. But coding has never had a real attempt at regulation. You're going to end up with "vibe coded" production code handling people's personal and financial information, and that is genuinely a bad idea. A novice will not spot security issues, and AI will happily produce them.

Re: Vibe Coding is not an excuse for low-quality work

#148

Post by anonzzzies »

I mainly use vibe coding to figure out how hard something is to do if I would do it 'for real', so I vibe code working prototypes, often to find out there is no way this is feasible and, more often, finding it is easier than I thought it would be. When I have an issue, I search for a library to solve that issue, but to figure it it does and if it does in a friendly way, I have to try it, so I ask claude to give me my example in that library. I already know something is up if it doesn't work first shot; to me is not a shock that 99.9999% (well that is actually not true of all language ecosystems but the most popular ones) of all libraries out there are total garbage with very strange 'opinionated' abstractions/choices in them that truly suck and actually make some use cases impossible without changing the library itself even though it says in the manual it is possible but without an example. In that case the LLM usually hallucinates a parameter or function that it doesn't have but that is from another library or does not exist at all, ideally needed to implement said use case. LLMs save me a ton of time there and I assume it's low quality as I won't use that exact piece of code anyway, but now I know the effort involved to meet my specific use case.

Re: Vibe Coding is not an excuse for low-quality work

#149

Post by ookblah »

ken47 wrote: post #134

The hype levels are so overwhelming that AI coding could never hope to meet them. I've tried having a highly-ranked AI coding app write unit tests for a relatively complex codebase. 80% of the generated test cases failed. But an experienced human such as myself could use those as a starting point since it took care of some of the tedious boilerplate. It genuinely saved me some time and annoyance, but could never hope…

I'm finding this as well. I can't trust it to go beyond boilerplate or PoC's or I risk lagging behind in understanding and actually slowing down in the long term. Best workflow is to tackle silo'd problems and iterate on them with LLM.

I will say in terms of fixing boilerplate or doing tedious work it's a godsend. The ability to give it literally garbage formatted content and return something relatively structured and coherent is great. Some use cases I find:

- Rapidly prototype/scaffold a new feature. I can do what takes me hours/days in magnitudes of less time and saving my brain energy lol. 90% of it is usually garbage, but I just need a PoC or direction and I can sniff out what approach doesn't work (goes with the above of not doing too much at at time)

- App typically has a lot of similarly structured sections that can't be DRY for whatever reason. Fix in one area, feed to LLM how to solve the problem and then apply it to X number of other sections. Boom, edits 20 files and saves me tedious hours of doing it manually and potentially screwing it up.

- Run logs and sections of code in an initial bug to see if I'm missing something obvious up front. Helps dive into the right areas.

- Transform garbage content into something else when I can't be bothered to write a complicated regex or script for.

Re: Vibe Coding is not an excuse for low-quality work

#150

Post by ehnto »

dheera wrote: post #31
SCdF wrote: post #5

So what have we redefined vibe coding to mean exactly? The original tweet[1] talked very specifically about not caring about quality, just accepting whatever code the AI produces blindly, as long as you get the black box output you're looking for, and just randomly try again if you didn't. Are people now using this term to mean "giving an AI agent broad tasks"? [1] https://x.com/karpathy/status/1886192184808149383?la…

I absolutely hate the polarization around "vibe coding". The whole point of AI agents is to eventually get good enough to do this stuff better than humans do. It's okay to dogfood and test them now and see how well they do, and improve them over time. Software engineers will eventually become managers of AI agents. Vibe coding is just version 0.1 pre-alpha of that future.

I think the issue most of us have, is that vibe-coding is not being treated as a dog fooding experiment, but as a legitimate way to deliver production code.

I am already seeing "vibe coding experts" and other attempts to legitimize the practice as a professional approach to software development.

The issue is clear, if you have accepted all PRs with no reviews as vibe-coding suggests, you will end up with security and functionality flaws.

Even if you do review, if you are the type of developer who thinks you can vibe-code a serious project, I doubt you are interested in regular security reviews.

Lastly, businesses are long lasting projects, and the state of AI is constantly evolving. Your codebases stability and speed of development is your businesses success, and if only the AI model who built it understands your code, you are on shaky ground when the models evolve. They are not constantly forward evolutions, there will be interactions that fail to fix or implement code in the codebase it previously wrote. Human engineers may not even be able or willing to save you, after 3 years of vibe coding your product.

Post reply on HN