4 ms·

I like how readable the code is. I don't write anything in C++ nowadays, but it's clear enough for me to follow.

by Tade0 10mo ago

stavros 10mo ago
Ah, I can't really take credit for that, it was mostly multiple LLMs.
- latexr 10mo ago
  So did you really write this, as claimed in the title, or did you vibe code it? If the latter, how much of the code did you review? How much of it do you understand and could rewrite by hand if necessary? How much of it are you confident you can find and fix bugs and exploits in?
  - stavros 10mo ago
    https://www.youtube.com/watch?v=D5Df191WJ3o https://www.youtube.com/watch?v=D5Df191WJ3o
    - lobsterthief 10mo ago
      I mean, these are fair questions if you’re the library maintainer :)
      - stavros 10mo ago
        Before LLMs, did you ask maintainers how much of the code they wrote themselves, vs how much of it did someone else write, and how well they reviewed the code? Because I haven't seen this ever, everyone just went by "does this library work for me? If not, I'll fix it or stop using it". It's only now, with LLMs, and particularly on HN that we're so all up in arms about authorship all of a sudden.
        latexr 10mo ago
        > and how well they reviewed the code? Yes! It’s insane to not worry about how well a maintainer reviews submitted code. Every semi-competent open-source developer understands that merging code they do not understand is a recipe for disaster and maintenance burden. Furthermore, they understand that not doing so is how you get malicious vulnerabilities merged. Have you truly not seen any such cases in recent memory? The fact you don’t think good reviews are important is worrying and puts into question all your repos.
        stavros 10mo ago
        I regret sharing this, that's for sure.
        Zetaphor 10mo ago
        I appreciate you taking the time to make this exist. Ignore the folks who just want to complain instead of bringing their own solution into existence.
        stavros 10mo ago
        Thank you!
        latexr 10mo ago
        Please stop assuming bad faith. I didn’t complain about anything, I asked a few questions to make an initial call on the level of engagement I’d want to have with the project. I even openly said I’d have just moved on without further reply if it was admitted to be entirely LLM generated with no review. https://news.ycombinator.com/item?id=46364478 https://news.ycombinator.com/item?id=46364478
        mikemcquaid 10mo ago
        If multiple people are assuming bad faith: perhaps you should adjust your communication in future rather than trying to change their mind. An apology wouldn’t hurt, either.
      - vmasto 10mo ago
        Can you explain why it makes a difference what the answers are? When using an open source library assumptions should be: - The code does what it advertises. - The owner is responsible for the functionality. - The owner's reputation is based on the quality. You're making it sound that you're more sure for the above when the code is "hand-written" than LLM-driven. Why exactly? Do you tend to deeply understand the strengths and limitations of every coder whose software you're using in your projects? As long as the owner is responsible for the quality of a project why does it matter how it was executed?
        latexr 10mo ago
        > Can you explain why it makes a difference what the answers are? You answered it yourself: > the owner is responsible for the quality of a project If you didn’t write, review, or understand the code, then you cannot be responsible for its quality. If you don’t have the skills to write it by hand and understand it, you don’t have to skills to properly address bug reports or understand and prevent malicious submissions. All of those are legitimate concerns and considerations when deciding if you want to invest your time in a project. Honestly, if the author had responded “I vibe coded it and didn’t review any of it, but it’s been working for me for <however long>”, that would’ve been fine. It would have been a clear, honest answer that would let everyone decide how they want to proceed.
        stavros 10mo ago
        No, I disagree with the premise, that's why I don't want to answer. I am responsible for the quality of the project by virtue of publishing it, not because I wrote it in a way you agree or disagree with. Your questions are irrelevant. The only thing that's relevant is whether my name is on the repo or not. If I didn't think it's good enough to release, I'd say something like "I vibe-coded this and didn't check it, use at your own peril". How much of the code I understand and how much an LLM wrote is entirely irrelevant.
        latexr 10mo ago
        > No, I disagree with the premise What premise?! I made general questions that I want to know for myself. They were questions, not accusations. The fact you saw then as such says something about you, not me. You flipped out for no reason. > that's why I don't want to answer. Then you could’ve said that too, instead of dismissing it with a sketch and lamenting without any clarification. > Your questions are irrelevant. That’s not for you to decide. Silly example: Imagine you have a peanut allergy and go to a restaurant. You ask the waiter if some dessert has peanuts in it and they answer “that is irrelevant”. Questions are relevant to the asker. You didn’t even try to understand why I made the questions, you simply assumed bad faith. > If I didn't think it's good enough to release, I'd say something like "I vibe-coded this and didn't check it, use at your own peril". And how is anyone supposed to know you’d do or not do that? If you had been upfront about using LLMs from the start, we wouldn’t be having this conversation. > How much of the code I understand and how much an LLM wrote is entirely irrelevant. No, no it is not. It may be irrelevant to you, but it’s not to everyone else. You don’t get to decide what other people think is relevant.
    - latexr 10mo ago
      If you think that asking “how much of this thing you claim to have written did you actually write and understand” is comparable to the Spanish Inquisition, it seems fair to assume that you probably wrote and understand close to nothing of it. Thank you for clarifying.
      - stavros 10mo ago
        Yes this makes sense.
        latexr 10mo ago
        Surely you’re doing nothing to dispel the notion. If it isn’t true, say it. My questions were genuine, I don’t understand your defensiveness. Either you understand the code or you don’t. There are people who will be fine with it and use it anyway, and people who won’t be and will use something else. But they all will benefit (and temper expectations) from an honest answer.
        Zetaphor 10mo ago
        If _you_ can read the code, and it's of reasonable quality (I can and this code appears to be), then what exactly is the point of this argument? Before the authors work, however you choose to define that word, this did not exist. Now it does, and it isn't slop. Why then does it matter how it came into existence? This feels like a petty attempt to diminish the fact that someone just gave us a piece of useful open source software.
        latexr 10mo ago
        > This feels like a petty attempt to diminish the fact It feels that way to you because you misunderstood the point and are outright assuming bad faith (as evidenced by sibling comments). > I can and this code appears to be Yes, appears. To know for sure you’d have to read more of it. You’d have to spend time to understand it. If you know the author and know they have been writing code before LLMs and they tell you they have written it all themselves or at the very least reviewed it all, you’ll place a certain amount of trust in the code based on the author. If, however, they say they vibe coded all of it and verified nothing, you’re no longer trusting the abilities of the author but the random musings of LLMs. Those two scenarios entail different levels of trust and verification and they affect how much time you yourself will decide to spend on doing your own review to go from appears good to “I’m confident it is good”. This is not hard to understand. All you have to do is not assume that everyone who makes a question about LLM usage is immediately bashing the project. In other words, don’t engage in bad faith and don’t assume motivation from strangers.
- Tade0 10mo ago
  Ha, I was thinking while reading this "never in my life did I work with a C++ programmer who would put whitespace to visually organise the code".
  - stavros 10mo ago
    Hmm, I'm not a C++ programmer, but I do that in the projects I write by hand as well. I haven't found many other ways to organise C++ code, so whitespace is the one thing I do do.

I like how readable the code is. I don't write anything in C++ nowadays, but it's clear enough for me to follow.

by Tade0 10mo ago

I like how readable the code is. I don't write anything in C++ nowadays, but it's clear enough for me to follow.

stavros 10mo ago

Ah, I can't really take credit for that, it was mostly multiple LLMs.

latexr 10mo ago
So did you really write this, as claimed in the title, or did you vibe code it? If the latter, how much of the code did you review? How much of it do you understand and could rewrite by hand if necessary? How much of it are you confident you can find and fix bugs and exploits in?
- stavros 10mo ago
  https://www.youtube.com/watch?v=D5Df191WJ3o https://www.youtube.com/watch?v=D5Df191WJ3o
  - lobsterthief 10mo ago
    I mean, these are fair questions if you’re the library maintainer :)
    - stavros 10mo ago
      Before LLMs, did you ask maintainers how much of the code they wrote themselves, vs how much of it did someone else write, and how well they reviewed the code? Because I haven't seen this ever, everyone just went by "does this library work for me? If not, I'll fix it or stop using it". It's only now, with LLMs, and particularly on HN that we're so all up in arms about authorship all of a sudden.
      - latexr 10mo ago
        > and how well they reviewed the code? Yes! It’s insane to not worry about how well a maintainer reviews submitted code. Every semi-competent open-source developer understands that merging code they do not understand is a recipe for disaster and maintenance burden. Furthermore, they understand that not doing so is how you get malicious vulnerabilities merged. Have you truly not seen any such cases in recent memory? The fact you don’t think good reviews are important is worrying and puts into question all your repos.
        stavros 10mo ago
        I regret sharing this, that's for sure.
        Zetaphor 10mo ago
        I appreciate you taking the time to make this exist. Ignore the folks who just want to complain instead of bringing their own solution into existence.
        stavros 10mo ago
        Thank you!
        latexr 10mo ago
        Please stop assuming bad faith. I didn’t complain about anything, I asked a few questions to make an initial call on the level of engagement I’d want to have with the project. I even openly said I’d have just moved on without further reply if it was admitted to be entirely LLM generated with no review. https://news.ycombinator.com/item?id=46364478 https://news.ycombinator.com/item?id=46364478
        mikemcquaid 10mo ago
        If multiple people are assuming bad faith: perhaps you should adjust your communication in future rather than trying to change their mind. An apology wouldn’t hurt, either.
    - vmasto 10mo ago
      Can you explain why it makes a difference what the answers are? When using an open source library assumptions should be: - The code does what it advertises. - The owner is responsible for the functionality. - The owner's reputation is based on the quality. You're making it sound that you're more sure for the above when the code is "hand-written" than LLM-driven. Why exactly? Do you tend to deeply understand the strengths and limitations of every coder whose software you're using in your projects? As long as the owner is responsible for the quality of a project why does it matter how it was executed?
      - latexr 10mo ago
        > Can you explain why it makes a difference what the answers are? You answered it yourself: > the owner is responsible for the quality of a project If you didn’t write, review, or understand the code, then you cannot be responsible for its quality. If you don’t have the skills to write it by hand and understand it, you don’t have to skills to properly address bug reports or understand and prevent malicious submissions. All of those are legitimate concerns and considerations when deciding if you want to invest your time in a project. Honestly, if the author had responded “I vibe coded it and didn’t review any of it, but it’s been working for me for <however long>”, that would’ve been fine. It would have been a clear, honest answer that would let everyone decide how they want to proceed.
        stavros 10mo ago
        No, I disagree with the premise, that's why I don't want to answer. I am responsible for the quality of the project by virtue of publishing it, not because I wrote it in a way you agree or disagree with. Your questions are irrelevant. The only thing that's relevant is whether my name is on the repo or not. If I didn't think it's good enough to release, I'd say something like "I vibe-coded this and didn't check it, use at your own peril". How much of the code I understand and how much an LLM wrote is entirely irrelevant.
        latexr 10mo ago
        > No, I disagree with the premise What premise?! I made general questions that I want to know for myself. They were questions, not accusations. The fact you saw then as such says something about you, not me. You flipped out for no reason. > that's why I don't want to answer. Then you could’ve said that too, instead of dismissing it with a sketch and lamenting without any clarification. > Your questions are irrelevant. That’s not for you to decide. Silly example: Imagine you have a peanut allergy and go to a restaurant. You ask the waiter if some dessert has peanuts in it and they answer “that is irrelevant”. Questions are relevant to the asker. You didn’t even try to understand why I made the questions, you simply assumed bad faith. > If I didn't think it's good enough to release, I'd say something like "I vibe-coded this and didn't check it, use at your own peril". And how is anyone supposed to know you’d do or not do that? If you had been upfront about using LLMs from the start, we wouldn’t be having this conversation. > How much of the code I understand and how much an LLM wrote is entirely irrelevant. No, no it is not. It may be irrelevant to you, but it’s not to everyone else. You don’t get to decide what other people think is relevant.
  - latexr 10mo ago
    If you think that asking “how much of this thing you claim to have written did you actually write and understand” is comparable to the Spanish Inquisition, it seems fair to assume that you probably wrote and understand close to nothing of it. Thank you for clarifying.
    - stavros 10mo ago
      Yes this makes sense.
      - latexr 10mo ago
        Surely you’re doing nothing to dispel the notion. If it isn’t true, say it. My questions were genuine, I don’t understand your defensiveness. Either you understand the code or you don’t. There are people who will be fine with it and use it anyway, and people who won’t be and will use something else. But they all will benefit (and temper expectations) from an honest answer.
        Zetaphor 10mo ago
        If _you_ can read the code, and it's of reasonable quality (I can and this code appears to be), then what exactly is the point of this argument? Before the authors work, however you choose to define that word, this did not exist. Now it does, and it isn't slop. Why then does it matter how it came into existence? This feels like a petty attempt to diminish the fact that someone just gave us a piece of useful open source software.
        latexr 10mo ago
        > This feels like a petty attempt to diminish the fact It feels that way to you because you misunderstood the point and are outright assuming bad faith (as evidenced by sibling comments). > I can and this code appears to be Yes, appears. To know for sure you’d have to read more of it. You’d have to spend time to understand it. If you know the author and know they have been writing code before LLMs and they tell you they have written it all themselves or at the very least reviewed it all, you’ll place a certain amount of trust in the code based on the author. If, however, they say they vibe coded all of it and verified nothing, you’re no longer trusting the abilities of the author but the random musings of LLMs. Those two scenarios entail different levels of trust and verification and they affect how much time you yourself will decide to spend on doing your own review to go from appears good to “I’m confident it is good”. This is not hard to understand. All you have to do is not assume that everyone who makes a question about LLM usage is immediately bashing the project. In other words, don’t engage in bad faith and don’t assume motivation from strangers.
Tade0 10mo ago
Ha, I was thinking while reading this "never in my life did I work with a C++ programmer who would put whitespace to visually organise the code".
- stavros 10mo ago
  Hmm, I'm not a C++ programmer, but I do that in the projects I write by hand as well. I haven't found many other ways to organise C++ code, so whitespace is the one thing I do do.