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

Don't fall into the anti-AI hype

antirez.com

Post reply on HN
Print view View original

681–690 of 1001 posts

Re: Don't fall into the anti-AI hype

#681

Post by svara »

> many times fundamental architectural issues cripple any attempt at prompting my way out of it, even though I've been quite involved step-by-step through the whole prototyping phase.

This doesn't make sense to me.

Surely if you were "quite involved step-by-step through the whole prototyping phase" you would have been able to prevent architectural mistakes being made?

What does your process really look like?

I don't "vibe code" in the sense that I have it build entire apps without looking at the code; I prompt it to write maybe about the 100-200 lines of code I need next after thinking about what they should look like.

I don't see how you get architectural issues creeping in if you do it that way.

Re: Don't fall into the anti-AI hype

#682

Post by Aeolun »

trueno wrote: post #671
Aeolun wrote: post #644

Earlier quoted context omitted.

Maybe if your coding style is already close to what an LLM like Claude outputs, you’ll never have these issues? At least it generally seems to be doing what I would do myself. Most of the architectural failures come from it still not having the whole codebase in mind when changing stuff.

I actually think it's less about code style and more about the disjointed way end outcomes seem to be the culmination of a lot of prompt attempts over the course of a project/implementation. The funny thing is reviewing stuff claude has made isn't actually unfamiliar to me in the slightest. It's something I'm intimately familiar with and have been intimately familiar with for many years, long before this AI stuff ble…

> The delivered documentation/code is noisy+disjointed.

Yeah, which is what you get if your memory consists of everything you’ve read in the past 20 minutes. Most of my Claude work involves pointing it at the right things.

Re: Don't fall into the anti-AI hype

#683

Post by fragmede »

FloorEgg wrote: post #586
virgildotcodes wrote: post #577

Earlier quoted context omitted.

On the subpar code, would the code work, albeit suboptimally? I think part of the problem a lot of senior devs are having is that they see what they do as an artisanal craft. The rest of the world just sees the code as a means to an end. I don't care how elegantly my toaster was crafted as long as it toasts the bread and doesn't break.

There is some truth to your point but you might want to consider that often seniors concerned with code quality aren't being pedantic about artisanal craft they are worried about the consequences of bad code... - it becomes brittle and rigid (can't change it, can't add to it) - it becomes buggy and impossible to fix one bug without creating another - it becomes harder to tell what it's doing - plus it can be ineffici…

Who does it fall on to fix the mess that's been made. You do care if the toaster catches fire and burns your house down.

Re: Don't fall into the anti-AI hype

#684

Post by selestify »

onlyrealcuzzo wrote: post #591
totallykvothe wrote: post #569

I don't understand the stance that AI currently is able to automate away non-trivial coding tasks. I've tried this consistently since GPT 3.5 came out, with every single SOTA model up to GPT 5.1 Codex Max and Opus 4.5. Every single time, I get something that works, yes, but then when I start self-reviewing the code, preparing to submit it to coworkers, I end up rewriting about 70% of the thing. So many important deta…

I'm not sure if I got in this weird LLM bubble where they give me bad advice to drive engagement, because I can't resist trying to correct them and tell them how absurdly wrong they are. But it is astounding how terrible they are at debugging non-trivial assembly in my experience. Anyone else have input here? Am I in a weird bubble? Or is this just not their forte? It's truly incredible how thoughtless they can be, s…

> I can't resist trying to correct them and tell them how absurdly wrong they are.

Oh god I thought I was the only one. Do you find yourself getting mad at them too?

Re: Don't fall into the anti-AI hype

#685

Post by littlestymaar »

totallykvothe wrote: post #569

I don't understand the stance that AI currently is able to automate away non-trivial coding tasks. I've tried this consistently since GPT 3.5 came out, with every single SOTA model up to GPT 5.1 Codex Max and Opus 4.5. Every single time, I get something that works, yes, but then when I start self-reviewing the code, preparing to submit it to coworkers, I end up rewriting about 70% of the thing. So many important deta…

> 2) Other people are just less picky than I am, or they have a less thorough review culture that lets subpar code slide more often.

Given how consistently terrible the code of Claude Code-d projects posted here have been, I think this is it.

I find LLMs pretty useful for coding, for multiple things(to write boilerplate, as an idiomatic design pattern search engine, as a rubber duck, helping me name things, explaining unclear error messages, etc.), but I find the grandiose claims a bit ridiculous.

Re: Don't fall into the anti-AI hype

#686

Post by shevy-java »

Reads like AI slop. This is also quite annoying: quality appears to go downwards constantly. I am not saying that the oldschool blogs were all great in quality, but now I look at an article to see signs of AI generated text and if I see it - or think I see it - I become more critical automatically. Even without that, it seems the quality of writing has gotten worse. I remember that people years ago complained that handwriting got worse (I always got awful handwriting though, way before using computers already); but now it seems that it also affected the brain. The seem to be some disconnect on what is thought, and what is written down.

Re: Don't fall into the anti-AI hype

#687

Post by daxfohl »

catlifeonmars wrote: post #643
daxfohl wrote: post #599

Earlier quoted context omitted.

Agreed, but: There's been a notable jump over the course of the last few months, to where I'd say it's inevitable. For a while I was holding out for them to hit a ceiling where we'd look back and laugh at the idea they'd ever replace human coders. Now, it seems much more like a matter of time. Ultimately I think over the next two years or so, Anthropic and OpenAI will evolve their product from "coding assistant" to "…

I actually think it’s the opposite. We’ll see fewer monorepos because small, scoped repos are the easiest way to keep an agent focused and reduce the blast radius of their changes. Monorepos exist to help teams of humans keep track of things.

Could be. Most projects I've worked on tend to span multiple services though, so I think AI would struggle more trying to understand and coordinate across all those services versus having all the logic in a single deployable instance.

The way I see feature development in the future is, PM creates a dev cluster (also much easier with a monolith), has AI implement a bunch of features to spec, AI provides some feedback and gets input on anywhere it might conflict with existing functionality, whether eventual consistency is okay, which pieces are performance criticial, etc., and provides the implementation, a bunch of tests for review, and errata about where to find observability data, design decisions considered and chosen, etc. PM does some manual testing across various personas and products (along with PMs from those teams), has AI add feature flags, launches. The feature flag rollout ends up being the long-pole, since generally the product team needs to monitor usage data for some time before increasing the rollout percentage.

So I see that kind of workflow as being a lot easier in a monolithic service. Granted, that's a few years down the road though, before we have AI reliable enough to do that kind of work.

Re: Don't fall into the anti-AI hype

#688

Post by rezonant »

mvkel wrote: post #580
totallykvothe wrote: post #569

I don't understand the stance that AI currently is able to automate away non-trivial coding tasks. I've tried this consistently since GPT 3.5 came out, with every single SOTA model up to GPT 5.1 Codex Max and Opus 4.5. Every single time, I get something that works, yes, but then when I start self-reviewing the code, preparing to submit it to coworkers, I end up rewriting about 70% of the thing. So many important deta…

> Other people are just less picky than I am I think this is part of it. When coding style has been established among a team, or within an app, there are a lot of extra hoops to jump through, just to get it to look The Right Way, with no detectable benefit to the user. If you put those choices aside and simply say: does it accomplish the goal per the spec (and is safe and scalable[0]), then you can get away with a lo…

I've seen vibe coding fall apart at 600 lines of code. It turns out lines of code is not a good metric for this or any other purpose.

Re: Don't fall into the anti-AI hype

#689

Post by nl »

aussieguy1234 wrote: post #515

AI is going to put a hold on the development of new programming languages for sure, since they won't be in the training set. Great news if you know the current generation of languages, you won't need to learn a new one for quite some time.

This is wrong and I don't know why people think it is true.

LLMs understand language Grammar files really well. A new language is easy for them (you can tell this by giving them a JSON schema and seeing how well they do)

What they don't always have is good taste with what preexisting libraries work together well. But this isn't a problem for new languages.

Re: Don't fall into the anti-AI hype

#690

Post by simonw »

shevy-java wrote: post #686

Reads like AI slop. This is also quite annoying: quality appears to go downwards constantly. I am not saying that the oldschool blogs were all great in quality, but now I look at an article to see signs of AI generated text and if I see it - or think I see it - I become more critical automatically. Even without that, it seems the quality of writing has gotten worse. I remember that people years ago complained that ha…

If you think antirez wrote that with AI then your AI detection insincts are way off.

It reads like someone with (good) English as a second language. LLMs don't write like that.

Post reply on HN