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

The hidden cost of AI coding

terriblesoftware.org

Post reply on HN
Print view View original

421–430 of 475 posts

Re: The hidden cost of AI coding

#421

Post by simonw »

th0ma5 wrote: post #416
simonw wrote: post #130

Earlier quoted context omitted.

How did you get there from me agreeing 100% with someone who said that you should be ready to verify everything an LLM does for you and if you're not willing to do that you shouldn't use them at all? Do you ever read my comments, or do you just imagine what I might have said and reply to that?

There's simply no way to verify everything that comes out of these things. Otherwise why use it? You also can't possibly truly know if you know more about a topic since by definition the models know more than you. This is automation bias. Do you not know the problems with even verifying or watching machines? This is a core part of the discussion of self driving vehicles. I guess I assumed you knew stuff about the fie…

"You also can't possibly truly know if you know more about a topic since by definition the models know more than you"

Where does that idea that "by definition the models know more than you" come from?

Re: The hidden cost of AI coding

#422

Post by RedNifre »

bgwalter wrote: post #167
RedNifre wrote: post #159

Earlier quoted context omitted.

Right, just how back in the day, people who loved writing assembly hated high level languages and people who found assembly too tedious loved compilers.

First of all, Lisp, Fortran and COBOL had been around most of the time when assembly was popular. Assembly was used because of resource constraints. Secondly, you are not writing anything you get from an LLM. You prompt it and it spits out other people's code, stripped of attribution. This is what children do: Ask someone to fix something for you without understanding the result.

Right, just how the assembler spits out machine code that other people figured out for you, so you don't even understand the result.

Re: The hidden cost of AI coding

#423

Post by sayamqazi »

skydhash wrote: post #221
sayamqazi wrote: post #203

Earlier quoted context omitted.

> The actual documentation needs to be by humans for humans. This is true for producing the documentation but if there is an LLM that can take said documentation and answer questions about it is a great tool. I think I get the answer far quicker with LLM than sifting through documentation when looking for existence of a function in a library or a property on an object.

The documentation are for answering your questions, it’s not a puzzle to be solved. Using the reference docs assumes that you already have an understanding about the thing that is being documented and you’re looking for specificity or details. If not, the correct move is to go through a book, a tutorial, or the user guide. Aka the introductory materials.

There are UI kits that have lots and lots of elements and then each element has lots and lots of properties to tweak. I have to either guess the correct term or go through dozens of exmaples for the desired behavior. AI has saved me a lot of time in this regard. It is basically a "full text search" but far better than whatever the docs websites have built in.

Re: The hidden cost of AI coding

#424

Post by Philpax »

BigJono wrote: post #180
danielbln wrote: post #174

Earlier quoted context omitted.

Not everyone works in a single language and/or deep in some singular code base.

Gee do you think maybe that's why all our software sucks balls these days?

No? Of all the reasons software sucks, multidisciplinary programmers are unlikely to be near the top.

Re: The hidden cost of AI coding

#425

Post by douglasisshiny »

pdimitar wrote: post #259
codr7 wrote: post #38

Earlier quoted context omitted.

We used to make fun of and look down on coders who mindlessly copy paste and mash the compile button until the code runs, for good reasons.

Did you skip the "rigorously verify the LLM code" part of my comment on purpose, just to show contempt?

I think the point is you could just.... write the code yourself.

Re: The hidden cost of AI coding

#426

Post by pdimitar »

douglasisshiny wrote: post #425
pdimitar wrote: post #259

Earlier quoted context omitted.

Did you skip the "rigorously verify the LLM code" part of my comment on purpose, just to show contempt?

I think the point is you could just.... write the code yourself.

And I do, most of the time. Sometimes the code to be written is quite verbose and I am on a tighter deadline. And it still takes less time and effort to vet the LLM's code. Otherwise yeah.

Re: The hidden cost of AI coding

#427

Post by Aeolun »

> if a person’s body or mind is stretched to its limits in a voluntary effort to accomplish something difficult and worthwhile.” — Mihaly Csikszentmihalyi

Like pronouncing your surname? Holy hell.

Ok, I needed to get that off my chest, will go back to reading the article now.

Re: The hidden cost of AI coding

#428

Post by schwartzworld »

asdff wrote: post #367
TrackerFF wrote: post #185

Some people love programming, for the sake of programming itself. They love the CS theory, they love the tooling, they love most everything about it. Other people see all that as an means to an end - and find no joy from the technical aspect of creating something. They're more interested in the end result / product, rather than the process itself. I think that if you're in group A, it can be difficult to understand g…

The issue with programming is that it isn't like music or really any other skill where you get feedback right away and operate in a well understood environment. And a lot of patterns are not well designed as they are often based on what a single developer things the behavior ought to be instead of something more deterministic like the laws of physics that influence the cord patterns we use in music. Nope, your code m…

> like the laws of physics that influence the cord patterns we use in music.

So much of what we think of as law in music is just being used to the conventions. Lots of amazing music would have been considered noise if created in an earlier time.

> The issue with programming is that it isn't like music or really any other skill where you get feedback right away and operate in a well understood environment.

Funny, I feel the opposite about programming. The feedback comes in milliseconds. Ok the build didn’t break, ok the ui is working, now check if the change I made is right, now run the tests, etc. and the environment is fully documented, depending on your tooling of choice and the quality of its docs.

Re: The hidden cost of AI coding

#429

Post by aitchnyu »

thi2 wrote: post #419
CopyOnWrite wrote: post #285

Most comments here surprise me: I am using Githubs Copilot / ChatGPT 4.0 at work with a code base which is mostly implements a basic CRUD service... and outside of small/trivial example (where the generated code is mostly okay), prompting is more often than not a total waste of time. Now, I wonder if I am just totally unable to write/refine good prompts for the LLM (as it works for smaller samples, I hope I am not to…

My employer gives me access to Jetbrains AI, I work on a Vue Frontend with a Kotlin Spring Boot backend. The codebase is not too old and has grown without too much technical debt, with complex prompts I never had decent success. Its usefull for quick "what does this do" checks but any real functionality seems to be lacking. Maybe I'm not refining my prompts good enough but doing so would take longer than implementing…

Jetbrains AI runs on a "discount LLM" and their ratings were below 2 stars. I tried two others, which played games with me to reduce context and use cheaper models. I then switched to Aider which leads me to believe a moderate Claude user may need to spend 30$ a month, but I use Gemini models and I didnt exceed 5$.

Re: The hidden cost of AI coding

#430

Post by rerdavies »

I'm a veteran professional programmer with 40+ years of experience. So far, I'm finding coding with an AI to be pure sweetness and light.

I cannot imagine why you cannot find flow using an AI assistant. I am definitely somebody who finds bliss in programming; and in my experience, AI assistants increase my bliss. My ideas are expressed in code much more efficiently. I spend less time in miserable documentation sets. I find myself fearlessly adding functionality that I would not have added if I weren't using an AI. And I have not a shadow of a doubt that my productivity has gone up dramatically. If anything, I find that AI assistants keep me in flow sate, particularly in cases where I would previously have had to wade through pages of ancient crusty Unix API documentation.

Maybe you should try a different AI. I found the ChatGPT AIs totally unhelpful, and counter-productive; but would recommend Claude Sonnet 3.7 without hesitation. I'm still working my way through other AIs. Others may be better at present, but so far I haven't found any that are dramatically better. It's hard to keep up with the furious pace of innovation.

It might also take a while to find your fu. I found the benefits to using an AI pretty much immediately; but I'm still discovering new and interesting ways to use my AI assistant.

Post reply on HN