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

631–640 of 1001 posts

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

#631

Post by Bayko »

newsoftheday wrote: post #589
mvkel wrote: post #580

Earlier quoted context omitted.

> 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…

> When coding style has been established It feels like you're diminishing the parent commenter's views, reducing it to the perspective of style. Their comment didn't mention style.

Isn't coding style a solved problem with claude.md files or something?

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

#632

Post by habosa »

Everyone on both sides of this argument seems like they won’t be satisfied until everyone comes to their side. The maximalists want us to submit to the AI godhead. The doomers want us to go back to writing assembly longhand on paper.

I’m starting to think of AI use more like a dietary choice. Most people are omnivores. Some people are vegans. Others are maxing protein. All of them can coexist in society and while they might annoy each other if the topic comes up, for the most part it’s a personal choice.

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

#633

Post by tayo42 »

I've been wondering lately if the career move is to get into SRE/Ops/Platform type work.

Imo its to hard for companies to get infra into a place where text can be an interface. IaC is mostly an aspiration beyond a certain scale ime, which is close enough to interacting with infra through text.

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

#634

Post by obirunda »

The claim that users who don't adopt AI now will pay for it later or some other notion is a contradiction of their position. People who are bullish on AI should support this view wholesale. Opus 4.5 is easier to use than GPT 3.5. It can actually code a full toy project one shot where you couldn't dream of it before. Opus 4.5 isn't perfect, so people have a lot of things they do for a competitive advantage. Though anything you think you're building with all the prompt alchemy and .md rules or whatever will be useless and futile on Opus 10, every "really good practice" is instantly absorbed by labs so when something great is in the wild everyone eventually benefits by the base .md or system prompts. So even if you feel like you have a competitive advantage right now, it will evaporate by either the labs improving their tools or become generally unnecessary in future versions of the models.

The goal of the labs is to continue these leaps will get even bigger with every generation. Unless you secretly believe that some portion of the craft will be left unexplored by the labs or the things that are still relatively borked now will not be worked on or fixed later is a silly notion to me. Future versions will be easier to prompt and the tools will do more of the heavy lifting of following up and re-rolling misinterpretations. I argue that a user sleeping through all of this is likely to use a future version better than someone who is obsessing with all their assumptions on how to coerce these models to work right now, current version hyper users will likely bring unnecessary baggage imo.

For now, even with Opus 4.5 the time horizon for delivering a full-stack project is not significantly different than before, it's still limited by how much you can push it. I'd argue that someone without understanding of how things work is unlikely to succeed in getting production-grade outcomes from these current versions. The point is, if you choose to learn more and get better in understanding and building things that work (with AI or otherwise) you'll be just fine to use the versions that have fully or mostly automated the entire process. Nobody will be left behind, only those who stop building altogether.

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

#635

Post by sublinear »

daxfohl wrote: post #599
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…

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 "…

This doesn't make any sense. If the business can get rid of their engineers, then why can't the user get rid of the business providing the software? Why can't the user use AI to write it themselves?

I think instead the value is in getting a computer to execute domain-specific knowledge organized in a way that makes sense for the business, and in the context of those private computing resources.

It's not about the ability to write code. There are already many businesses running low-code and no-code solutions, yet they still have software engineers writing integration code, debugging and making tweaks, in touch with vendor support, etc. This has been true for at least a decade!

That integration work and domain-specific knowledge is already distilled out at a lot of places, but it's still not trivial. It's actually the opposite. AI doesn't help when you've finally shaved the yak smooth.

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

#636

Post by jmalicki »

eru wrote: post #623
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…

> 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. Have another model review the code, and use that review as automatic feedback?

CodeRabbit in particular is gold here. I don't know what they do but it is far better at reviewing than any AI model I've seen. From the deep kinds of things it finds, I highly suspect they have a lot of agents routing code to extremely specialized subagents that can find subtle concurrency bugs, misuse of some deep APIs etc. I often have to do the architecture l/bug picture/how this fits into project vision review myself, but for finding actual bugs in code, or things that would be self evident from reading one file, it is extremely good.

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

#637

Post by what »

simonw wrote: post #630
vignesh37 wrote: post #627

Earlier quoted context omitted.

The biggest frustration with LLMs for me is people telling me I'm not promoting it in a good way. Just think about any product where they are selling a half baked product, and repeatedly telling the user you are not using it properly.

But that's not how most products work. If you buy a table saw and can't figure out how to cut a straight line in a piece of wood with it - or keep cutting your fingers off - but didn't take any time at all to learn how to use it, that's on you. Likewise a car, you have to take lessons and a test before you can use those! Why should LLMs be any different?

It’s more like the iPhone “you’re holding it wrong”.

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

#638

Post by tjr »

simonw wrote: post #630
vignesh37 wrote: post #627

Earlier quoted context omitted.

The biggest frustration with LLMs for me is people telling me I'm not promoting it in a good way. Just think about any product where they are selling a half baked product, and repeatedly telling the user you are not using it properly.

But that's not how most products work. If you buy a table saw and can't figure out how to cut a straight line in a piece of wood with it - or keep cutting your fingers off - but didn't take any time at all to learn how to use it, that's on you. Likewise a car, you have to take lessons and a test before you can use those! Why should LLMs be any different?

It seems generally agreed that LLMs (currently) do better or worse with different programming languages at least, and maybe with other project logistical differences.

The fact that an LLM works great for one user on one project does not mean it will work equally great for another user on a different project. It might! It might work better. It might work worse.

And both users might be using the tool equally well, with equal skill, insofar as their part goes.

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

#639

Post by kristofferR »

eru wrote: post #623
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…

> 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. Have another model review the code, and use that review as automatic feedback?

That is actually a gold tip. Codex CLI is way less pleasant to use than Opus, but way better at finding bugs, so I combine them.

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

#640

Post by robertfw »

eru wrote: post #623
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…

> 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. Have another model review the code, and use that review as automatic feedback?

I've been using a `/feedback ...` command with claude code where I give it either positive or negative feedback about some action it just did, and it'll look through the session to make some educated guesses about why it did some thing - notably, checking for "there was guidance for this, but I didn't follow it", or "there was no guidance for this".

the outcome is usually a new or tweaked skill file.

it doesn't always fix the problem, but it's definitely been making some great improvements.

Post reply on HN