I appreciate Simon’s balanced take on how LLMs can enhance a project when used responsibly.
I’m curious, though—what are this community’s opinions on the use of LLMs in programming?
I’ve almost completely stopped using them, unless I’m stuck at a dead end. In the end all they have done is slow me down and make me unable to think properly anymore. They usually write way too much code, especially with tab complete stuff, resulting in me needing to delete code after hitting tab (what’s the point even, intellisense has always been really good and now it’s somehow worse). They’re usually wrong unless prompted multiple times. People say you can use them to generate boilerplate, but just use a language with less or no boilerplate like Kotlin. There’s usually very subtle bugs they introduce or they’re solving a problem that is simply documented on stack overflow, while I wouldn’t be using an LLM if I could just kagi it, so they solve the wrong thing.
One thing it’s decent for, if you don’t care about code quality, is converting code to a language you do not know. You’re not going to end up with good idiomatic code at the end, but it will probably function.
None of this is to say that the LLMs aren’t amazing, but if you start to depend on them you very very quickly realize that your ability to solve more complex problems will atrophy. And then when you get to a difficult problem you now waste much more time trying to solve a problem that might have been simpler for past you.
My 2cents
It’s also trained on other people’s code, it may use outdated, inefficient, or otherwise bad code. If it would be trained on my code, I’d like it much more
Had a subscription, unsubscribed 6 months ago. Simplistically:
- They create bad code,
- You stop learnng. You want to program? Learn.
It’s funny, to me I’ve had an llm give me the wrong answer to questions every time.
The first time I couldn’t remember how to read a file as a string in python and it got me most of the way there. But I trusted the answer thinking “yeah, that looks right” but it was wrong, I just got the io class I didn’t call the read() function.
The other time it was an out of date answer. I asked it how to do a thing in bevy and it gave me an answer that was deprecated. I can sort of understand that though, bevy is new and not amazingly documented.
On a different note, my senior who is all PHP, no python, no bash, has used LLM’s to help him write python and bash. It’s not the best code, I’ve had to do optimisations on his bash code to make it run on CI without taking 25 minutes, but it’s definitely been useful to him with python and bash, he was hired as a PHP dev.
Your problem is you don’t understand how llms work. You treat it like a magic genie when its not. Treat it right and you can fly. I integrated a new messaging architecture into my stack the other day that would of taken me weeks before. But I isolated my problem set and targeted what I needed to target. But I also understand what to tell it and how to utilize it as a tool.
In your case, its trivial to just check the methods of the class or know that its a call you’re accessing in the first place. That AI can’t read your mind if you don’t frame the problem correctly.
My experience is that use of an LLM is an amplifier to your output but generally at no better quality that you can produce on your own.
The skilled developer who uses an LLM and checks its work will get a productivity boost without a loss in quality.
The unskilled developer who copy/pastes code from stackover can get even more sloppy code into production by using an LLM.
The issue isn’t whether you can get a good results or not. The issue is the skills you are outsourcing to a proprietary tool, skills that you will never learn or forget. Getting information out of documentation, designing an architecture, understanding and replicating an algorithm, etc.
You will eventually start struggling with critical thinking, there are already studies about that.
Of course, if you use it in moderation and don’t rely on LLMs too much, you should be ok.
But how did that work for everyone with short-form content and social networks in the last ten years? How is your attention span doing? Surely we all have managed to take short-form content in moderation, since we knew the risks to our attention span, right?
The problem with this article is that he stresses that you need to check the code and step in when needed - yet relying heavily on LLMs will invariably make it impossible for you to tell what’s wrong and eventually how to even read the code (since it will produce code using libraries you never experimented with because the LLM can just write the code).
Also “vibe-coding” is stupid af. You take out the human element altogether because you just accept all changes without reading them and then copy/paste errors back in without any context.
If I’m doing something in a language I only half way know and rarely use in depth I’ll use them more. Like bash scripting I use them all the time. For Java I basically never touch them because I don’t need them.
I mainly use it to create boilerplate (like adding a new REST API endpoint), or where I’m experimenting in a standalone project and am not sure how to do something (odd WebGL shaders), or when creating basic unit tests.
But letting it write, or rewrite existing code is very risky. It confidently makes mistakes, and rewrites entire sections of working code, which then breaks. It often goes into a “doom loop” making the same mistakes over and over. And if you tell it something it did was wrong and it should revert, it may not go back to exactly where you were. That’s where frequently snapshotting your working code into git is essential, and being able to reset multiple files back to a known state will save your butt.
Just yesterday, I had an idea for a WebGL experiment. Told it to add a panel to an existing testing app I run locally. It did and after a few iterations, got it working. But three other panels stopped working, because it decided to completely change some unrelated upstream declarations. Took 2x time to put everything back to where it was.
Another thing to consider is that every X units of time, you’ll want to go back and hand edit the generated material to clean up sloppy code. For example, inefficient data structures, duplicate functions in separate sections, unnecessarily verbose and obvious comments, etc. Also, better if using mature tech (with lots of training examples) vs. a new library or language.
If just starting out, I would not trust AI or vibe coding. Build things by hand and learn the fundamentals. There are no shortcuts. These things may look like super tools, but they give you a false sense of confidence. Get the slightest bit complex, and they fall apart and you will not know why.
Mainly using Cursor. Better results with Claude vs other LLMs, but still not perfect. Paid versions of both. Have also tried Cline with local codegen through Llama and Qwen. Not as good. Claude Code looks decent, but the open-ended cost is too scary for indie devs, unless you work for a company with deep pockets.
