Hacker News Reader

Comment by edg5000 | Hacker News Reader

Seems like the wrong question to ask. I've been programming my whole life but basically stopped writing code by hand in 2026. The LLM writes better code than I do, much better.

View on HN

Same here. I'm still a better software architect that AI, but there is no question that my AI generated and reviewed code has fewer bugs than code I hand write. It takes some humility to acknowledge that your coding prowess is less of a useful skill than it used to be.

There's an issue where people assumed the syntactic activity of writing code was what mattered. The reality is that this was always a smaller part of the role, as opposed to thinking about observability, serviceability, and test automation. The ability to write software that is properly separated from concerns and when to enact those separations matters.

At the same time, I think we're far too far down the systems path now. We've hit a point where interviewing has become purely systems design "because the AI writes the code".

Not that I'm ever asked, but I inherently believe the act of critical thinking, communication, and expression are the key skills for those who already have the appropriate coding/engineering/cs/etc background. I now only interview for those skills - but through the lens of impossible to solve systems design conversations as opposed to problems. It tells me a lot about how people think.

Not sure what you mean as system design conversations because while in theory those can be good in practice the ones I have been at had been techbro wankery where the interviewer had a particular answer in mind. Like designing your own memcached clone for example is a terrible task for systems design.

What you're mentioning is 100% what's wrong with the industry. Agreed! To me a systems design conversation is a conversation - not a design goal. The idea is to determine ability and psychology:

1. When you press on someone's design respectfully, do they get defensive. Do they become argumentative.

2. When thoughtfully pointing out a concern, how does the candidate take it?

3. When you suggest a technology that makes no sense to intentionally challenge knowledge, does the candidate recognize why it makes no sense? Are they able to share what the negative of the approach is. If you indicate that you know the question is "senseless" but want their feedback, how do they communicate?

4. When you hard request a change that requires a literal rethink and rewrite do they become argumentative? Do they embrace the change?

5. When discussing testing, how do they think about it? I come down to the nitty gritty and ask about postive vs negative cases, table driven testing, what types of tests matter (for our situation) and why.

6. We discuss timeline tradeoffs, and then have the conversation about the candidate's approach given updates to see how they think.

You'll notice that I am never looking for a solution. I'm seeking communication, description, partnership while having a (relatively) thorough gasp of the subject matter.

Every single time I get a response from a candidate such as "I don't know, I'd have to learn more - or use AI to, or.. what do you think" turns out to be something I LOVE, because it creates a great fabric for the interview.

Writing is thinking

So is architecting, testing, validating, and even occasionally using.

This isn't the first time I've seen this phrase recently, but I'm not sure what the thought is a cliche or what it is intended to convey (don't read my note as negative, I sincerely am unsure what connotation folks are trying to say).

The idea behind "writing is thinking" is that people often overestimate their understanding until pressed to express it in words (or code).

How many times in your career did you sit down to tackle a task thinking you knew exactly how to approach it only to realize during implementation that there were edge cases you hadn't considered, API contracts that were now broken, or that the feature was trying to solve the wrong problem.

Having to be the one at the helm during implementation made you intimately aware of not only the problem at hand, but the current state of the codebase. That's something you can't replace with automation. You can't compress all of that context into your brain in a handful of prompts with Claude.

Remember the words of your math teacher--

"Watching someone else solve the problem doesn't mean you can now solve it too."

Programming languages, design languages and architecture are all inventions made to help humans write understandable source. LLMs don’t really need to do any of that. They can store very large trees of understanding and therefore implement any application in raw binary. Why bother with abstractions at all

LLMs for sure need those things. maybe not the same abstractions as humans do but without understabdable code an LLM will just fail to accomplish the task you ask it to do.

It might be less ambitious and more practical to target bytecode.

But you effectively lose the human review component.

There's more python and typescript in the training data than bytecode.

I was thinking about that. I reckon generating massive amounts of synthetic training data for that goal should be possible - you've already got the Python, after all. It's also possible for machine code, but you'd have to target more platforms.

But ultimately, I think human readability outweighs any theoretical advantage you get from removing a step in the compilation process.

Targeting byte code or asm instead of high level would be silly for everyday tasks. You blow up the number of tokens, reduce your effective context, and there's just more places for it to make a mistake, which most likely won't be caught by the assembler (unlike compilers).

The AI machine can write better code. It can also write an interpreter which implements function calls by instantiating a new interpreter + entire standard library per function call. Or it will build a 300kloc cathedral of scaffolding and maintain that forever, never writing actual code. Or it will create a CI system that takes 2 hours to run and constantly fails, and the agent loops there all day, fixing a small bug and waiting 2 hours. (All things I’ve experienced latest frontier models do)

Agentic engineering faces all kinds of new problems that couldn’t exist before, and need experienced engineers to solve them.

With the "intelligence" of code focused and capable llm in the last six months, the main problem I'm seeing now is where some total amateur who has no previous knowledge of coding tries to one shot a project. People who have previous experience and know how to architect things (and when to stop an LLM from doing something wrong that will cause maintenance and scale and extensibility problems in the future) are doing much better building actually useful things.

This one shot thing I just don’t understand. The way I’m using it, it takes weeks of constant prompts because it never does exactly what I ask no matter how well I specify. I just don’t see how it’s possible to one shot anything unless you don’t have strong requirements on the output.

I had pretty good success with a language learning app. Initial prompt below, about an hour to get it working very close to how I imagined, then extended to Japanese, French, and audio generation.

======================================

Hi! I'd like to create an app for interactively learning Chinese using AI. My current idea is:

- The AI generates a Chinese sentence at a specified vocabulary level (e.g. HSK2)

- The user translates the sentence into English, and the AI evaluates the response. If the answer is wrong or is partially correct but could be improved, the AI offers a hint. If correct, the AI confirms and explains any particularly important vocabulary or grammar points that the sentence demonstrates.

- The user should be able to click on an individual character once to see the pinyin, and again to see its definition and any helpful techniques to remember it (radicals, similarity to other characters, visual meaning)

- The app should also be able to go the other way, giving the user an English sentence and having them translate to Chinese.

App details:

- UI is a web app

- The AI should be pluggable. We'll start with a local Ollama install running gemma4, but it should be easily possible to add support for Claude/OpenAI/Gemini or other models (may need to provide an API key).

- Not actually sure if we need a backend. It might be useful to keep track of characters or concepts that the user has difficulty with, or to keep track of what sentences the AI has previously generated so it doesn't become repetitive.

- Build everything in a Docker container (or multiple if needed with docker compose)

I can see this being a one-shot thing because the brief is pretty light; beyond broad strokes, you haven't specified much for it to actually do. Any UI that's also a webapp would satisfy this brief, but that's exactly what I mean; e.g. if you want a specific UI, there's a lot more back-and-forth that's involved. It'll one shot a UI, but it's very rarely the UI you want down to all the interactions and placement of buttons.

Yeah, that's fair. I didn't have an exact UI planned so I gave it flexibility, and it made reasonable choices that only needed a bit of tweaking to be quite usable.

For me the usefulness of a survey like this has nothing to with how effective LLMs are themselves. It's more that when someone's able to produce an app in an afternoon, and submitting the app to F-Droid becomes a checkbox, how confident can you be that they'll continue maintaining the app? Sure if it's open source you can have your own LLM maintain it, but at that point what's the value in having it on F-Droid?

"There are naïve questions, tedious questions, ill-phrased questions, questions put after inadequate self-criticism. But every question is a cry to understand the world. There is no such thing as a dumb question". (Carl Sagan)

Just because you don’t seem to be interested in the answer - then don’t read it? - doesn’t make the question wrong.

Skill issue then

I just asked Astra to bring an old Windows XP game to the browser. It objdump'ed the whole thing, built a fitting Win32-like wrapper that exposes required functionality like DirectDraw, DirectSound, SEH etc., then wrote an x86-32/x87 interpreter in WASM, benchmarked how the game runs, lifted the hotspots of the executable to WASM too and now it is playable!

I mean, I'm proud of my low-level skills too but this is some Fabrice Bellard level sorcery. Very, very few humans are able to do this without AI tools.

If there are very few humans that can do this is because the market for such a task is very small and thus there is little incentive to learn how to do it or produce tools that can do it.

But good code isn't just "does it work", it's also

- is it understandable

- is it maintainable

- how much work is adding new features

- is it written in a way that adding new features means rewriting a lot of it

- is it written in a consistent style

- and lots of other things

I use AI to write a lot of my code, but the only time it's clearly "better" than a competent human is for one-off things.

That being said - AI + human is, without any doubt in my mind, better than either one alone.

Aye. Some of these targets are far away, others perhaps closer.

Human+AI systems is a good match. Like Human+docs or Human+encyclopedia.

- is it understandable

Yes you can ask the agent anything about it and interrogate it until you understand.

- is it maintainable

Yes it’s easy to ask the ai to add new features or to refactor it entirely.

- how much work is adding new features

Depends, it could just be one prompt, it’s usually many prompts. If the refactor is large it can take weeks. But before AI something g equivalent would take months.

- is it written in a way that adding new features means rewriting a lot of it

Usually no, but that depends on how well the agent is being directed and what the features are. If you come up with a feature that requires a new architecture, ai makes it doable rather than saying “would be nice but we’d also have to implement this whole new architecture and that’s a lot of work”

- is it written in a consistent style

Styles can be applied mechanically with linters and formatters, so as much as any codebase written by multiple people.

What people don't understand that programming is very much an art. You iteratively work on it ripping parts out, rewriting and rewriting and rewriting, while also rewriting and then rewriting every time a new feature, bug fix or scaling changes are needed.

I've seen AI be wrong about things often enough to know that none of what you said is particularly true. Rather, most of it holds true most of the time. But not enough.

> Styles can be applied mechanically with linters and formatters, so as much as any codebase written by multiple people.

I'm talking more of a higher level than this - more of coding/design patterns that are common for the team.

> Rather, most of it holds true most of the time. But not enough.

I see it be wrong about things all day every day. And yet, IME it's correct enough for it to be controllable. It doesn't have to hold up all of the time, it just has to respond to corrections when they're issued in a loop so that it converges to a correct solution. And it does, despite the mistakes.

In one of my other posts in this thread I detail some of the the ways it's confounded me, but those issues have caused me to harden validation mechanisms rather than say "this thing makes mistakes so I can't use it to write software".

> I'm talking more of a higher level than this - more of coding/design patterns that are common for the team.

Do you have a concrete example?

Absolutely, the Win32-WASM layer it wrote is some of the most evil looking code I have seen in my life. But realistically, why keep it maintainable for humans if you won't find anybody that can work on it without AI anyways?

If we humans are just doing code style checks, file organizing and doc cleanups I feel we have demoted ourselves to code janitors. This is neither fun nor going to last.

Personally I've always strived for minimalism, to find the smallest, fastest, simplest solution possible so I'm pretty jaded now, too...

Code is a human formalism that is only incidentally made executable. The elegance of code represents understanding of the problem to its most minute detail. LLM code being recognizably terrible shows it still doesn't understand what it produces to the full extent, and just as with people, it will inevitably compound to it becoming unable to efficiently work on it. And if the human abdicates that responsibility too, then there is no code, only subtly broken software.

As a bit of an observation on that specific project... By hand as a human you could spend six months of the equivalent of a full time job doing that. Even if you had extensive knowledge in all of its discrete pieces. One of the things coding focused LLM are great at is doing things that have no reasonable prospect of economic necessity to do (no for profit company is going to pay you a FTE salary for six months to do that task, because there's no possible revenue in it). But the LLM can be pointed at it and get it done in a day or two with some periodic architecture and decision making by the human, for probably under $50.

It took two days and 50% of my Codex Plus week limit, so around 3$. Finally I can play the game on multiplayer again next evening!

What is the game?

https://www.gog.com/en/game/the_settlers_4_gold_edition

It did not work properly on Wine nor Windows 10, that was the entire reason for trying it out.

fwiw, PCGamingWiki[1] has a link to a patch[2] that someone wrote to fix the compatibility issues. If you're playing from the CD edition, you might also need SafeDiscShim.[3]

[1]: https://www.pcgamingwiki.com/wiki/The_Settlers_IV

[2]: https://github.com/elishacloud/dxwrapper/wiki/The-Settlers-I...

[3]: https://github.com/RibShark/SafeDiscShim

Only people disliking AI for coding are the gatekeepers who think they are magicians and the plebs shouldn't be able to code like them, unless they become gud.

This is a false generalization. Lots of AI dislikers do so for a wide variety of reasons. I dislike it and don't care how other people use it.

Reply on news.ycombinator.com