Hacker News Reader

Comment by milicat | Hacker News Reader

The more I browse through this, the more I agree. I feel like one could delete almost all comments from that project without losing any information – which means, at least the variable naming is (probably?) sensible. Then again, I don't know the application domain.

Also…

  def _save_current_date_time(current_date_time_file: str, current_date_time: str) -> None:
    with Path(current_date_time_file).open("w") as f:
      f.write(current_date_time)

there is a lot of obviously useful abstraction being missed, wasting lines of code that will all need to be maintained.

The scary thing is: I have seen professional human developers write worse code.

View on HN

> I feel like one could delete almost all comments from that project without losing any information

I far from a heavy LLM coder but I’ve noticed a massive excess of unnecessary comments in most output. I’m always deleting the obvious ones.

But then I started noticing that the comments seem to help the LLM navigate additional code changes. It’s like a big trail of breadcrumbs for the LLM to parse.

I wouldn’t be surprised if vibe coders get trained to leave the excess comments in place.

It doesn't hurt that the model vendors get paid by the token, so there's zero incentive to correct this pattern at the model layer.

or the model get trained from teaching code which naturally contains lots of comments.

the dev is just lazy to not include them anymore, wheres the model doesn't really need to be lazy, as paid by the token

LLMs are also good at commenting on existing code.

It’s trivial to ask Claude via Cursor to add comments to illustrate how some code works. I’ve found this helpful with uncommented code I’m trying to follow.

I haven’t seen it hallucinate an incorrect comment yet, but sometimes it will comment a TODO that a section should be made more more clear. (Rude… haha)

I have seldomly seen insightful comments from LLMs. It is usually better than "comment what the line does" usefull for getting a hint about undocumented code, but not by much. My experience is limited, but what I have I do agree with. As long as you keep on the beaten path it is ok. Comments are not such a thing.

More tokens -> more compute involved. Attention-based models work by attending every token with each other, so more tokens means not only having more time to "think" but also being able to think "better". That is also at least part of the reason why o1/o3/R1 can sometimes solve what other LLMs could not.

Anyway, I don't think any of the current LLMs are really good for coding. What it's good at is copy-pasting (with some minor changes) from the massive code corpus it has been pre-trained. For example, give it some Zig code and it's straight unable to solve even basic tasks. Same if you give it really unique task, or if you simply ask for potential improvements of your existing code. Very, very bad results, no signs of out-of-box thinking whatsoever.

BTW: I think what people are missing is that LLMs are really great at language modeling. I had great results, and boosts in productivity, just by being able to prepare the task specification, and do quick changes in that really easily. Once I have a good understanding of the problem, I can usually implement everything quickly, and do it in much much better way than any LLM can currently do.

I have tried getting gemini 2.5 to output "token efficient" code, i.e. no comments, keep variables to 1 or 2 letters, try to keep code as condensed as possible.

It didn't work out that great. I think that all the context in the verbose coding it does actually helps it to write better code. Shedding context to free up tokens isn't so straightforward.

What’s worse, I get a lot of comments left saying what the AI did, not what the code does or why. Eg “moved this from file xy”, “code deleted because we have abc”, etc. Completely useless stuff that should be communicated in the chat window, not in the code.

>The scary thing is: I have seen professional human developers write worse code.

This is kind of the rub of it all. If the code works, passes all relevant tests, is reasonably maintainable, and can be fitted into the system correctly with a well defined interface, does it really matter? I mean at that point its kind of like looking at the output of a bytecode compiler and being like "wow what a mess". And it's not like they can't write code up to your stylistic standards, it's just literally a matter of prompting for that.

> If the code works, passes all relevant tests, is reasonably maintainable, and can be fitted into the system correctly with a well defined interface, does it really matter?

You're not wrong here, but there's a big difference in programming one-off tooling or prototype MVPs and programming things that need to be maintained for years and years.

We did this song and dance pretty recently with dynamic typing. Developers thought it was so much more productive to use dynamically typed languages, because it is in the initial phases. Then years went by, those small, quick-to-make dynamic codebases ended up becoming unmaintainable monstrosities, and those developers who hyped up dynamic typing invented Python/PHP type hinting and Flow for JavaScript, later moving to TypeScript entirely. Nowadays nobody seriously recommends building long-lived systems in untyped languages, but they are still very useful for one-off scripting and more interactive/exploratory work where correctness is less important, i.e. Jupyter notebooks.

I wouldn't be surprised to see the same pattern happen with low-supervision AI code; it's great for popping out the first MVP, but because it generates poor code, the gung-ho junior devs who think they're getting 10x productivity gains will wisen up and realize the value of spending an hour thinking about proper levels of abstraction instead of YOLO'ing the first thing the AI spits out when they want to build a system that's going to be worked on by multiple developers for multiple years.

> those small, quick-to-make dynamic codebases ended up becoming unmaintainable monstrosities

In my experience, type checking / type hinting already starts to pay off when more than one person is working on an even small-ish code base. Just because it helps you keep in mind what comes/goes to the other guy's code.

And in my experience "me 3 months later" counts as a whole second developer that needs accommodating. The only time I appreciate not having to think about types is on code that I know I will never, ever come back to—stuff like a one off bash script.

> "me 3 months later" counts as a whole second developer

A fairly incompetent one, in my experience. And don't even get me started on "me 3 months ago", that guy's even worse.

"How has that shit ever worked?"

Me, looking at code 100% written by me last year.

It gets worse with age and size of the project. I’m getting the same vibes, but for code written by me last month.

Yep, I've seen type hinting even be helpful without a type checker in python. Just as a way for devs to tell each other what they intend on passing. Even when a small percent of the hints are incorrect, having those hints there can still pay off.

I think the productivity gains of dynamic typed languages were real, and based on two things: dynamic typing (can) provide certain safety properties trivially, and dynamic typing neatly kills off the utterly inadequate type systems found in mainstream languages when they were launched (the 90s, mostly).

You'll notice the type systems being bolted onto dynamic languages or found in serious attempts at new languages are radically different than the type systems being rejected by the likes of javascript, python, ruby and perl.

The ML world being nearly entirely in Python, much of it untyped (and that the Python type system is pretty weak) is really scary.

>The ML world being nearly entirely in Python, much of it untyped (and that the Python type system is pretty weak) is really scary

I think this has a ton to do with the mixed results from "vibe coding" we've seen as the codebase grows in scope and complexity. Agents seem to break down without a good type system. Same goes for JS.

I've just recently started on an Objective-C project using Cline, and it's like nirvana. I can code out an entire interface and have it implemented for me as I'm going. I see no reason it couldn't scale infinitely to massive LOC with good coding practices. The real killer feature is header files. Being able to have your entire projects headers in context at all time, along with a proper compiler for debugging, changes the game for how agents can reason on the whole codebase.

> You're not wrong here, but there's a big difference in programming one-off tooling or prototype MVPs and programming things that need to be maintained for years and years.

Humans also worry about their jobs, especially in PIP-happy companies; they are very well known for writing intentionally over-complicated code that only they understand so that they are irreplaceable

I'm not convinced this actually happens. Seems more like somthing people assume happens because they don't like whatever codebase is at the new job.

Oh, I'm convinced, I've seen it first hand.

If your TC is 500k-1M and you don’t feel like job hopping anymore, you’d certainly not want to get hit by a random layoff due to insufficient organizational masculinity or whatever. Maintaining a complex blob of mission critical code is one way of increasing your survival chances, though of course nothing is guaranteed.

People doing layoffs have no idea of who works and who's warming the chair.

Depending on the layoff they may look into yearly reviews... or not.

Ah yes, those work /s

The challenge is that sufficiently bad code could be intentional or it could be from a lack of skill.

For example, I've seen a C# application where every function takes in and outputs an array of objects, supposedly built that way so the internal code can be modified without ever having to worry about the contract breaking. It was just as bad as you are imagining, probably worse. Was that incompetence or building things to be so complicated that others would struggle to work on it?

but that is literally how the browser window DOM works, no? It depends on how diligent the maintenance is IMHO

hmm I have seen conda env with far too many packages and maybe a lot of current version bumping, and the dev says "who cares" and it naturally gets a bit more.. Intentionally complicated is more like an accusation of wrongdoing.

I'm certainly extremely happy for having an extensive type system in my daily driver languages especially when working with AI coding assistance — it's yet another very crucial guard rail that ensures that keeps the AI on track and makes a lot of fuckups downright impossible.

what are you going to do when something suddenly doesn't work and cursor endlessly spins without progress no matter how many "please don't make mistakes" you add? delete the whole thing and try to one-shot it again?

Why do you HAVE TO one-shot? No one says you have to code like those influencers. You are a software engineer, use AI like one, iteratively.

>No one says you have to code like those influencers. You are a software engineer, use AI like one, iteratively.

This is my issue with all the AI naysayers at this point. It seems to all boil down to "haha, stupid noob can't code so he uses AI" in their minds. It's like they are incapable of understanding that there could simultaneously be a bunch of junior devs pushing greenfield YouTube demos of vibe coding, while at the same time expert software engineers are legitimately seeing their productivity increase 10x on serious codebases through judicious use.

Go ahead and keep swinging that hammer, John Henry.

“Maybe you didn’t hear me, I said ‘good morning steam driver, how are you?’”

> expert software engineers are legitimately seeing their productivity increase 10x

It's funny you would say this, because we are really commenting on an article where a self-proclaimed "expert" has done that and the "10x" output is terrible.

I have just checked my article — the word "expert" isn't in it, so not quite sure where you got this from.

I'm working in the field professionally since June 1998, and among other things, I was the tech lead on MyHammer.de, Germany's largest craftsman platform, and have built several other mid-scale online platforms over the decades.

How well I have done this, now that's for others to decide.

Quite objectively though, I do have some amount of experience — even a bad developer probably cannot help but pick up some learnings over so many years in relevant real-world projects.

However, and I think I stated this quite clearly, I am expressively not an expert in Python.

And yet, I could realize an actually working solution that solves an actual problem I had in a very real sense (and is nicely humming away for several weeks now).

And this is precisely where yes, I did experience a 10x productivity increase; it would have certainly taken me at least a week or two to realize the same solution myself.

Apologies for implying you are claiming to be an expert software engineer: I took the "senior" in the title and "25 years of experience" in the post to mean similar things as "expert".

I don't doubt this is doing something useful for you. It might even be mostly correct.

But it is not a positive advertisement for what AI can do: just like the code is objectively crap, you can't easily trust the output without a comprehensive review. And without doubting your expertise, I don't think you reviewed it, or you would have caught the same smells I did.

What this article tells me is that when the task is sufficiently non-critical that you can ignore being perfectly correct, you can steer AI coding assistants into producing some garbage code that very well might work or appear to work (when you are making stats, those are tricky even with utmost manual care).

Which is amazing, in my opinion!

But not what the premise seems to be (how a senior will make it do something very nice with decent quality code).

Out of curiosity why did you not build this tool in a language you generally use?

Because I wanted exactly this experience: can I get to the desired result — functionality-wise, if not code-wise! — even if I choose the stack that makes sense in terms of technology, not the one that I happen to be proficient in?

And if I cannot bring language-proficiency to the table — which of my capabilities as a seasoned software&systems guy can I put to use?

In the brown-field projects where my team and I have the AI implement whole features, the resulting code quality — under our sharp and experienced eyes — tends to end up just fine.

I think I need to make the differences between both examples more clear…

Ok, I guess you shouldn't complain that you really got exactly what you wanted.

However, your writing style implied that the result was somehow better because you were otherwise an experienced engineer.

Even your clarification in the post sits right below your statement how your experience made this very smooth, with no explanation that you were going to be happy with bad code as long as it works.

I guess we are slowly but steadily approaching splitting-hairs-territory, so not sure if this is still worth it…

However. I‘m not quite sure where I complained. Certainly not in the post.

And yes, I’m very convinced that the result turned out a lot better than it would have turned out if an unexperienced „vibe coder“ had tried to achieve the same end result.

Actually pretty sure without my extensive and structured requirements and the guard rails, the AI coding session would have ended in a hot mess in the best case, and a non-functioning result in the worst case.

I‘m 100% convinced that these two statements are true and relevant to the topic:

That a) someone lacking my level of experience and expertise is simply not capable of producing a document like https://github.com/dx-tooling/platform-problem-monitoring-co...

And that b) using said document as the basis for the agent-powered AI coding session had a significant impact on the process as well as the end result of the session.

I think some of the suspicion is that it's really not 10x in practice.

Like AI could write code perfectly as soon as I thought of it, and that would not improve my productivity 10x. Coding was never the slow part. Everything that goes around coding (like determining that the extra load here is not going to overload things, getting PMs to actually make their mind up what the feature is going to do, etc.), means that there's simply not that much time to be saved on coding activities.

Same argument can be said for not using any tooling really. "Tech is the easy part". No difference typing code on notepad and having zero process/engineering infrastructure I guess. Because stakeholder management is the main engineering skill apparently.

Btw, AI doesn't just code, there are AIs for debugging, monitoring etc too.

If you were to really measure speed improvement of notepad vs a tricked out IDE, it's probably not much. The problem would be the annoyance caused to an engineer who has to manually type out everything.

No, coding speed is really not the bottleneck to software engineer productivity.

> coding speed > the annoyance caused to an engineer

No one said productivity is this one thing and not that one thing, only you say that because it's convenient for your argument. Productivity is a combination of many things, and again it's not just typing out code that's the only area AI can help.

The argument of "coding speed not a bottleneck to productivity" is not in contradiction to "productivity is a combination": it even implies it.

Again, the context here was that somebody discussed speed of coding and you raised the point of not using any tooling with Notepad.

The context here is AI assisted engineering and you raised the point that non-engineering productivity is more important for engineers, which I think is absurd.

You can have 10x engineering productivity boost and still complete work in the same amount of time, because of communication and human factors. Maybe it's a problem, may be it's not. It's still a productivity gain that will make you work better nonetheless.

I did not raise it, but what was raised was "coding speed": as in, the speed to type code into an editor.

That's not "engineering", but "coding".

Engineering already assumes a lot more than just coding: most importantly, thinking through a problem, learning about it and considering a design that would solve it.

Nobody raised communication or the human factors.

Current LLMs can indisputably help with the learning part, with the same caveats (they will sometimes make shit up). Here we are looking at how much they help with the coding part.

There are two levels to this.

1. Tooling obviously does improve performance, but not so huge a margin. Yes, if AI could automate more elements of tooling, that would very much help. If I could tell an AI "bisect this bug, across all projects in our system, starting with this known-bad point", that would be very helpful -- sometimes. And I'm sure we'll get there soon enough. But there is fractal complexity here: what if isolating the bug requires stepping into LLDB, or dumping some object code, or running with certain stressors on certain hardware? So it's not clear that "LLM can produce code from specs, given tight oversight" will map (soon) to "LLM can independently assemble tools together and agentically do what I need done".

2. Even if all tooling were automated, there's still going to be stuff left over. Can the LLM draft architectural specs, reach out to other teams (or their LLMs), sit in meetings and piece together the big picture, sus out what the execs really want us to be working on, etc.? I do spend a significant (double-digit) percentage of my time working on that, so if you eliminate everything else -- then you could get 10x improvement, but going beyond that would start to run up against Amdahl's Law.

My grievances are simple: an expert programming utilizing AI will be a truly dangerous force.

But that's not what we get in this early stage of grifting. We get 10% marketing buzz on how cool this is with stuff that cannot be recreated in the tool alone, and 89% of lazy or inexperienced developers who just turn in slop with little or no iteration. The latter don't even understand the code they generated.

That 1% will be amazing, it's too bad the barrel is full of rotten apples hiding that potential. The experts also tend to keep to themselves, in my experience. the 89% includes a lot of dunning-kruger as well which makes those outspoken experts questionable (maybe a part of why real experts aren't commenting on their experience).

Weren't you the guy who only writes HTML? Maybe let domain experts comment on their domain of expertise.

The point is because it generally produces crap code you have to one shot or else iteration becomes hard. Similar to how a junior would try to refactor their mess and just make a bigger mess

I find it hard to believe that when the AI generates crap code, there is absolutely nothing you can do (change the prompt, modify context, add examples) to make it do what you want. It has not been my experience either. I only use AI to make small modules and refactor instead of one-shoting.

Also I find "AI makes crap code so we should give it a bigger task" illogical.

it seems that there are really, really large differences between models; how well they do, what they respond to.. even among the "best" .. the field does seem to be moving faster

Right, and the reason why professional developers are writing worse code out there is most likely because they simply don't have the time/aren't paid to care more about it. The LLM is then mildly improving the output in this brand of common real world scenario

I think this code is at least twice the size than it needs to be compared to nicer, manually produced Python code: a lot of it is really superfluous.

People have different definitions of "reasonably maintainable", but if code has extra stuff that provides no value, it always perplexes the reader (what is the point of this? what am I missing?), and increases cognitive load significantly.

But if AI coding tools were advertised as "get 10x the output of your least capable teammate", would they really go anywhere?

I love doing code reviews as an opportunity to teach people. Doing this one would suck.

Good insight, and indeed quite exactly my state of mind while creating this particular solution.

Iin this case, I did put in the guard rails to ensure that I reach my goal in hopefully a straight line and a quickly as possible, but to be honest, I did not give much thought to long-term maintainability or ease of extending it with more and more features, because I needed a very specific solution for a use case that doesn't change much.

I'm definitely working differently in my brown-field projects where I'm intimately familiar with the tech stack and architecture — I do very thorough code reviews afterwards.

> there is a lot of obviously useful abstraction being missed, wasting lines of code that will all need to be maintained.

This is a human sentiment because we can fairly easily pick up abstractions during reading. AIs have a much harder time with this - they can do it, but it takes up very limited cognitive resources. In contrast, rewriting the entire software for a change is cheap and easy. So to a point, flat and redundant code is actually beneficial for a LLM.

Remember, the code is written primarily for AIs to read and only incidentally for humans to execute :)

At the very least, if a professional human developer writes garbage code you can confidently blame them and either try to get them to improve or reduce the impact they have on the project.

With AI they can simply blame whatever model they used and continually shovel trash out there instantly.

I don't see the difference there. Whether I've written all the code myself or an AI wrote all of it, my name will be on the commit. I'll be the person people turn to when they question why code is the way it is. In a pull request for my commit, I'll be the one discussing it with my colleagues. I can't say "oh, the AI wrote it". I'm responsible for the code. Full stop.

If you're in a team where somebody can continuously commit trash without any repercussions, this isn't a problem caused by AI.

> The scary thing is: I have seen professional human developers write worse code.

That's not the scary part. It's the honest part. Yes, we all have (vague) ideas of what good code looks like, and we might know it when we see it but we know what reality looks like.

I find the standard to which we hold AI in that regard slightly puzzling. If I can get the same meh-ish code for way less money and way less time, that's a stark improvement. If the premise is now "no, it also has to be something that I recognize as really good / excellent" then at least let us recognize that we have past the question if it can produce useful code.

I do believe it's amazing what we can build with AI tools today.

But whenever someone advertises how an expert will benefit from it yet they end up with crap, it's a different discussion.

As an expert, I want AI to help me produce code of similar quality faster. Anyone can find a cheaper engineer (maybe five of them?) that can produce 5-10x the code I need at much worse quality.

I will sometimes produce crappy code when I lack the time to produce higher quality code: can AI step in and make me always produce high quality code?

That's a marked improvement I would sign up for, and some seem to tout, yet I have never seen it play out.

In a sense, the world is already full of crappy code used to build crappy products: I never felt we were lacking in that department.

And I can't really rejoice if we end up with even more of it :)

I think there’s a difference in that this is about as good as LLM code is going to get in terms of code quality (as opposed to capability a la agentic functionality). LLM output can only be as good as its training data, and the proliferation of public LLM-generated code will only serve as a further anchor in future training. Humans on the other hand ideally will learn and improve with each code review and if they don’t want to you can replace them (to put it harshly).

Reply on news.ycombinator.com