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

Using AI for Coding: My Journey with Cline and LLMs

pgaleone.eu

Post reply on HN
Print view View original

71–80 of 97 posts

Re: Using AI for Coding: My Journey with Cline and LLMs

#71

Post by numba888 »

bandrami wrote: post #49
numba888 wrote: post #45

Looks outdated by now. Try R1. Seriously, I've been using LLMs for coding for a while and can say early experience was disappointing, but they get better and better fast. The latest o1 looks a lot better than 4o. It's reasonable to expect with proper human supervision and interface they will be able to handle big files and projects in a year or two. Interesting times...

They also said that a year or two ago

Today it's within the reach, I think. While effective context window is small for big files it should be possible to use RAGs to get relevant pieces in smaller window. That will be mostly data structures and functions' declarations. Only what's needed. Humans work the same way on non-trivial projects. Supervision should include strategic planing and subdivision to smaller tasks. Not sure, just my feeling, we are close to have capable models with big enough context windows. May be they are already good enough, just need proper orchestration.

Hmm... I can try... OAI has something called 'projects', but local handling with API calls is probably the right way of doing it. Easier to switch providers, run and debug in place. With current prices it should be like < $10/month.

Re: Using AI for Coding: My Journey with Cline and LLMs

#72

Post by lopuhin »

I find it strange that the author is really happy with the quality of string comparison here https://pgaleone.eu/ai/coding/2025/01/26/using-ai-for-coding... and while it would kind of work, it's a very weird piece of code from ML standpoint, e.g. it's training a TF-IDF vectorizer on just two strings being compared, which at best won't change anything (unless the same word is repeated within one product), and is a super weird thing to do as for better quality you'd probably want to train that on some corpus, or not bother at all. And also it compare two strings as bags of words, which again is not the end of the world but maybe not what the author wants here, and if they want this then it's not the easiest way of doing it. So it's taking some things which can be useful when comparing texts (tf-idf and cosine similarity) but then applying them in a weird way which does not let them show their strengths.

Re: Using AI for Coding: My Journey with Cline and LLMs

#73

Post by DecoySalamander »

bandrami wrote: post #48
AgentMatrixAI wrote: post #7

Prompt quality and knowing your domain is critical. One issue I had early on was experimenting with LLMs to generate a frontend application in a brand new framework I was unfamiliar with (Svelte at the time) which lead to situations where I would cruise along and get stuck in a loop. The other issue came from the increasing context size that led to more unpredictable behaviors (i would ask it to change the color of a…

How much of this is a cage we've built for ourselves? When I made my first website in 1998 even if we had had LLMs asking one how to change the color of a button would be ludicrous because I'd change an HTML attribute (did CSS exist back then? don't remember) either in a static file or in a perl script. That's it. And the modern "throw a hundred and seven frameworks at it" stack doesn't really do anything that simple…

What has changed over the years is our expectations and requirements. Sure, you can skip "all the noise" and slap an attribute on your button, but then you'll have to track down another 20 buttons in your codebase that need a color update. You can be smart with your approaches, but at a certain project size and complexity, frameworks are unavoidable. The only choice you really have is to learn an established one or handcraft something new for every thing you make.

Re: Using AI for Coding: My Journey with Cline and LLMs

#74

Post by joseda-hg »

frogsRnice wrote: post #50
joseda-hg wrote: post #42

Earlier quoted context omitted.

It is lovely when doing exploratory research, less so when in a hurry and you don't want to take the extra mental effort to figure out which of the nerds is right or if one of them is talking out of left field

As opposed to wondering if the llm is hallucinating? You have to expend a mental effort to think about your solutions anyway; I guess it’s pick your poison really.

Is it important if it's ocasionally hallucinating?

It's not like you should blindly throw the code in, you should run it and verify it

The more common the work you're doing the less likely it is to hallucinate, plus you can ask it to stick to whatever arbitrary coding standards you want so it's more readable to you, a rewrite to remove a wrong library takes an extra couple seconds per method/function

Also it's not like Stack Overflow or other non generated resources don't ocasionally hallucinate, it's not weird for the second or third voted answers in SO to be followed by the comment "This doesnt work because XYZ"

Re: Using AI for Coding: My Journey with Cline and LLMs

#76

Post by jpt4 »

ItsBob wrote: post #70

Anecdotal but here's how I described using the likes of Copilot to my sceptical colleagues (they were late to the party!): It's like having a senior software dev over your shoulder. He knows pretty much everything about coding but he often comes to work drunk... And that was the best analogy I could come up with: I think it's sped up my work enormously because I limit what it does, rather than let it loose... if that…

If that analogy is remotely accurate, it strikes me as a strictly negative value proposition for the tool.

Re: Using AI for Coding: My Journey with Cline and LLMs

#77

Post by mzhaase »

There is an enormous difference for me between inline suggestions and using chat.

If you use chat it's work, and then you have to debug and understand something you didn't write yourself.

Using inline suggestions is the closest I have come to plugging my brain directly into the computer. I type 5 characters and in 80% of cases the suggestion is character by character exactly what I would have written. It speeds me up enormously.

Re: Using AI for Coding: My Journey with Cline and LLMs

#78

Post by skydhash »

ItsBob wrote: post #70

Anecdotal but here's how I described using the likes of Copilot to my sceptical colleagues (they were late to the party!): It's like having a senior software dev over your shoulder. He knows pretty much everything about coding but he often comes to work drunk... And that was the best analogy I could come up with: I think it's sped up my work enormously because I limit what it does, rather than let it loose... if that…

To me, that’s either where a real editor shine (vim or emacs macro) or that’s when you need a metaprogramming capable language.

Re: Using AI for Coding: My Journey with Cline and LLMs

#79

Post by skydhash »

DecoySalamander wrote: post #73
bandrami wrote: post #48

Earlier quoted context omitted.

How much of this is a cage we've built for ourselves? When I made my first website in 1998 even if we had had LLMs asking one how to change the color of a button would be ludicrous because I'd change an HTML attribute (did CSS exist back then? don't remember) either in a static file or in a perl script. That's it. And the modern "throw a hundred and seven frameworks at it" stack doesn't really do anything that simple…

What has changed over the years is our expectations and requirements. Sure, you can skip "all the noise" and slap an attribute on your button, but then you'll have to track down another 20 buttons in your codebase that need a color update. You can be smart with your approaches, but at a certain project size and complexity, frameworks are unavoidable. The only choice you really have is to learn an established one or h…

That’s the problem template languages solve. That button could be a partial, and you include it where you want a button. And with grep, you can track it down.

Re: Using AI for Coding: My Journey with Cline and LLMs

#80

Post by skydhash »

joseda-hg wrote: post #74
frogsRnice wrote: post #50

Earlier quoted context omitted.

As opposed to wondering if the llm is hallucinating? You have to expend a mental effort to think about your solutions anyway; I guess it’s pick your poison really.

Is it important if it's ocasionally hallucinating? It's not like you should blindly throw the code in, you should run it and verify it The more common the work you're doing the less likely it is to hallucinate, plus you can ask it to stick to whatever arbitrary coding standards you want so it's more readable to you, a rewrite to remove a wrong library takes an extra couple seconds per method/function Also it's not li…

That’s why you take a quick glance of the answer, then read the comments. The do a deeper analysis. Take something like 10 sexonds as it seems every real answer I find that’s good is usually just one or two paragraphs.
Post reply on HN