991–1000 of 1001 posts
It sounds like someone taking a condescending tone to say that "Fast fashion is great! The clothes are really fine, and being able to wear a t-shirt once and then throwing it away instead of washing it is more convenient. This is most definitely progress. Sustainability? Hmm I can't say, that's not my field of expertise, therefore it's probably not important".
I think this article is pretty spot on — it articulates something I’ve come to appreciate about LLM-assisted coding over the past few months. I started out very sceptical. When Claude Code landed, I got completely seduced — borderline addicted, slot machine-style — by what initially felt like a superpower. Then I actually read the code. It was shockingly bad. I swung back hard to my earlier scepticism, probably even…
Desktop publication software killed many jobs. I worked for a publication where I had colleagues that used to typeset, place images, and use a camera to build pages by hand. That required a team of people. Once Quark Xpress and the like hit the scene, one person could do it all, faster.
In terms of illustration, the tools moved from pen and paper to Adobe Illustrator and Aldus / Macromedia Freehand. Which I'd argue was more of a sideways move. You still needed an illustrators skillset to use these tools.
The difference between what I just described and LLM image generation is the tooling changed to streamline an existing skillset. LLM's replace all of it. Just type something and here's your picture. No art / design skill necessary. Obviously, there's no guarantee that the LLM generated image will be any good. So, I'm not sure the Photoshop analogy works here.
Hundreds of comments. Some say LLMs are the future. Others say they don't work today and they won't work tomorrow. Videogame speed running has this problem solved. Livestream your 10x engineer LLM usage, a git commit annotated with it's prompt per change. Then everyone will see the result. This doesn't seem like an area of debate. No complicated diagrams required. Just run the experiment and show the result.
Programmers who don't work to deepen their understanding? Nah, sorry, not programmers I want anywhere near the code I'm working on.
To the extent that an LLM can help deepen understanding, good! Use it! Use its code. But you use it according to your intentional, systematic understanding of what the you're doing and why. And that understanding is not a function of code generation or, worse, code churn for the sake of feeling like you're going fast and paying less.
> LLMs can write a large fraction of all the tedious code you’ll ever need to write. And most code on most projects is tedious.
I've been writing code for over 30 years and I can't recall ever writing "tedious code". I don't know what this even means. Code, by definition, should not be tedious. At any layer of the stack, code should solve non-tedious problems. The whole point of code is to avoid tedium.
Most code, most projects? I haven't seen this and doubt it's even close to true - if it is God help us. If you write tedious code, find another place to work or another career.
> Sometimes, gnarly stuff needs doing. But you don’t wanna do it. So you refactor unit tests, soothing yourself with the lie that you’re doing real work.
Seriously who kills time by refactoring unit tests? Is this projection by the author?
> But an LLM can be told to go refactor all your unit tests.
Wtaf?? The unit tests are what stands between the blind ambitions of a statistical model and correct code. Handing over tests to an LLM? One staggers.
Tests represent part of that problem definition and solution. They're integral to the code. They foster and deepen understanding - not to be subject to the musings of statistical process, unless those musings serve as a dialectic.
> If you can’t metabolize the boring, repetitive code an LLM generates: skills issue! How are you handling the chaos human developers turn out on a deadline?
Answer: Slow the F down! Chaos and programming don't belong. Managers who drive their teams like this are bad managers.
> If hallucination matters to you, your programming language has let you down.
The author seems to think that programming is about producing code. Producing code that compiles is not a problem. Producing correct and coherent solutions to problems that need to be solved is a problem. Hallucinations aside, how will you know that generated code is something to keep or reject unless you have some understanding of what the code should do before you see it?
> Does an intern cost $20/month? Because that’s what Cursor.ai costs.
Squee! Code for all my friends!
> but it’s bad at rust
No, it's bad at understanding. And so are humans, initially. But it's our job to build and deepen understanding and we use code to facilitate that.
> Professional software developers are in the business of solving practical problems for people with code.
Correct.
> [LLMs] devour schlep, and clear a path to the important stuff, where your judgement and values really matter.
This begs the question, who defines what's important? It sounds like the author agrees that it's us, not the model. I agree this is a useful method. It's like any code review - it sharpens the reader no matter the code quality.
> As a mid-late career coder, I’ve come to appreciate mediocrity.
How cool and nonchalant what a rebel he says the opposite of everyone what a baddie!
I think the author means "appreciate naivety", meaning, a first pass that's good enough is good enough - there's no need to go further. I'll concede this, but it's a concession and not something to boast about. Odds are decent you'll come back to this naive implementation and revise it. Yes, at a certain point there's diminishing returns but good programmers know when to invest in revision and when to move on.
> LLMs really might displace many software developers.
We've seen this cycle. Technology and globalization puts pressure on programmers. There are legions of programmers in parts of the world that will generate code for $20 an hour. You can pay $20 and hour - or $20 a month - and generate mountains of code that blows chunks - not because it's ugly but because it's incoherent or because it solves the wrong problem. We. have. seen. this. It's why Silicon Valley still pays massive salaries to good programmers.
> The cool kid haughtiness about “stochastic parrots” and “vibe coding” can’t survive much more contact with reality.
Time will tell. In the meantime I'll keep investing in understanding, not code, and see how things shake out.
This article does not touch on the thing which worries me the most with respect to LLMs: the dependence . Unless you can run the LLM locally, on a computer you own, you are now completely dependent on a remote centralized system to do your work. Whoever controls that system can arbitrarily raise the prices, subtly manipulate the outputs, store and do anything they want with the inputs, or even suddenly cease to opera…
There's something that seems to be missing in all these posts and that aligns with my personal experience trying to use AI coding assistants. I think in code. To me, having to translate the into natural language for the LLM to translate it back into code makes very little sense. Am I alone in this camp? What am I missing?
I’m AI neutral but the writing style here is pretty dismissive, and - to match the tone of the article - annoying as fuck. Most completely reasonable objections to LLMs were totally dismissed.
We should flat-out dismiss the issue of licenses and copyright violations because... all developers pirate TV shows?
Seriously... what a load of crap.
Tbf the dismissal of the IP argument is a bit disappointing. Just because you don't respect IP doesn't mean we all do. Some of the code I'm proudest of I'm considering never open sourcing now because I simply don't trust online repositories anymore. Granted I'm in a bit of a different field than most developers but still, open source as a concept seems less attractive in the wake of tools that scrape it and copy it a…
I still do stuff in the hopes of it benefitting anyone - but not parasitic oligarchs who are responsible for the imminent destruction of the world and society.
I don't know if a reasonable number of spare-time creators, authors, artists etc. feel similar about these things, but if they do, I'd expect a lot more content under non-free licenses in the future, that might've been published openly otherwise.
You know what, I can use the word “fuck” too, just like this pro coder does: I’m fucking tired of these fucking nuts who act as if there is no problem at all having AI mess with my code. It harms my relationship with my code! It reduces my agency. It threatens my accountability. He implies that if I am talking about some other flavor of AI than he personally uses than I am just not serious— but that’s what they said…