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

741–750 of 1001 posts

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

#741

Post by richardw »

carlmr wrote: post #727
unyttigfjelltol wrote: post #604

Earlier quoted context omitted.

> non-trivial coding tasks I’ve come back to the idea LLMs are super search engines. If you ask it a narrow, specific question, with one answer, you may well get the answer. For the “non-trivial” questions, there always will be multiple answers, and you’ll get from the LLM all of these depending on the precise words you use to prompt it. You won’t get the best answer, and in a complex scenario requiring highly recurs…

>It’s not readily apparent at first blush the LLM is doing this, giving all the answers. Now I'm wondering if I'm prompting wrong. I usually get one answer. Maybe a few options but rarely the whole picture. I do like the super search engine view though. I often know what I want, but e.g. work with a language or library I'm not super familiar with. So then I ask how do I do x in this setting. It's really great for get…

I’ve recently created many Claude skills to do repeatable tasks (architecture review, performance, magic strings, privacy, SOLID review, documentation review etc). The pattern is: when I’ve prompted it into the right state and it’s done what I want, I ask it to create a skill. I get codex to check the skill. I could then run it independently in another window etc and feed back to adjust…but you get the idea.

And almost every time it screws up we create a test, and often for the whole class of problem. More recent it’s been far better behaved. Between Opus, skills, docs, generating Mermaid diagrams, tests it’s been a lot better. I’ve also cleaned up so much of the architecture so there’s only one way to do things. This keeps it more aligned and helps with entropy. And they’ll work better as models improve. Having a match between code, documents and tests means it’s not just relying on one source.

Prompts like this seem to work: “what’s the ideal way to do this? Don’t be pragmatic. Tokens are cheaper than me hunting bugs down years later”

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

#742

Post by staticshock »

smt88 wrote: post #663
staticshock wrote: post #656

Efficient markets route around bottlenecks. Technological revolutions accelerate the speed at which that re-routing happens. In software, we, the developers , have increasingly been a bottleneck. The world needs WAY more software than we can economically provide, and at long last a technology has arrived that will help route around us for the benefit of humanity. Here's an excellent Casey Handmer quote from a recent…

> In software, we, the developers, have increasingly been a bottleneck. The world needs WAY more software than we can economically provide, and at long last a technology has arrived that will help route around us for the benefit of humanity. Everything you wrote here is directly contradicted by casual observation of reality. Developers aren't a bottleneck. If they were, we wouldn't be in a historic period of layoffs.…

Yes, the layoffs are a market correction initiated by non-AI factors, such as the end of the ZIRP era.

The world is chock-full of important, society-scale problems that have been out of reach because the economics have made them costly to work on and therefore risky to invest in. Lowering the cost of software development de-risks investment and increases the total pool of profitable (or potentially profitable) projects.

The companies that will work on those new problems are being conceived or born right now, and [collectively] they'll need lots of AI-native software devs.

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

#743

Post by lazyasciiart »

fmbb wrote: post #703
khafra wrote: post #667

Earlier quoted context omitted.

> Non-trivial coding tasks A coding agent just beat every human in the AtCoder Heuristic optimization contest. It also beat the solution that the production team for the contest put together. https://sakana.ai/ahc058/ It's not enterprise-grade software, but it's not a CRUD app with thousands of examples in github, either.

> It's not enterprise-grade software, but it's not a CRUD app with thousands of examples in github, either. Optimization is a very simple problem though. Maintaining a random CRUD app from some startup is harder work.

The argument was about “non-trivial”. Are you calling this work trivial or not?

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

#744

Post by Auracle »

vignesh37 wrote: post #627
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…

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.

Have you seen the way some people google/prompt? It can be a murder scene.

Not coding related but my wife is certainly better than most and yet I’ve had to reprompt certain questions she’s asked ChatGPT because she gave it inadequate context. People are awful at that. Us coders are probably better off than most but just as with human communication if you’re not explaining things correctly you’re going to get garbage back.

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

#745

Post by antirez »

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…

After you review, instead of rewriting 70% of the code, have you tried to follow up with a message with a list of things to fix?

Also: in my experience 1. and 2. are not needed for you to have bad results. The existing code base is a fundamental variable. The more complex / convoluted it is, the worse is the result. Also in my experience LLMs are constantly better at producing C code than anything else (Python included).

I have the feeling that the simplicity of the code bases I produced over the years, and that now I modify with LLMs, and the fact they are mostly in C, is a big factor why LLMs appear to work so well for me.

Another thing: Opus 4.5 for me is bad on the web, compared to Gemini 3 PRO / GPT 5.2, and very good if used with Claude Code, since it requires to reiterate to reach the solution, why the others sometimes are better first-shotter. If you generate code via the web interface, this could be another cause.

There are tons of variables.

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

#746

Post by barrenko »

embedding-shape wrote: post #95

> But what was the fire inside you, when you coded till night to see your project working? It was building. I feel like this is not the same for everyone. For some people, the "fire" is literally about "I control a computer", for others "I'm solving a problem for others", and yet for others "I made something that made others smile/cry/feel emotions" and so on. I think there is a section of programmer who actually do…

Yeah, not all painters were happy with the transition to photography.

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

#747

Post by antirez »

softwaredoug wrote: post #718

If you’re getting started, in say Claude, some pointers that helped me Stay in plan mode most of the time. It will produce a step by step set of instructions - more context - for the LLM to execute the change. It’s the best place to exert detailed control over what will happen. Claude lets you edit it in a vim window. Think about testing strategy carefully. Connecting the feedback back into the LLM is what makes a lo…

Good points. Also:

Force it to have clear metrics / observability on what it is doing. For instance the other day I wanted Claude to modify a Commodore 64 emulator, and I started saying it to implement an observability framework where as the emulator run, it can connect to a socket and ask for registers, read/write memory areas, check the custom chips status, set breakpoints, ... As you can guess, after this the work is of a different kind.

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

#748

Post by nottorp »

So the "AI" hypers are now inventing an anti AI hype?

They ran out of believable arguments or never had any to begin with?

As it was said on a thread here, LLMs are search engines. The rest is religion.

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

#749

Post by Hamuko »

wpm wrote: post #331
zeroonetwothree wrote: post #311

Earlier quoted context omitted.

You’re right of course. For me there’s no flow state possible with LLM “coding”. That makes it feel miserable instead of joyous. Sitting around waiting while it spits out tokens that I then have to carefully look over and tweak feels like very hard work. Compared to entering flow and churning out those tokens myself, which feels effortless once I get going. Probably other people feel differently.

I'm the same way. LLMs are still somewhat useful as a way to start a greenfield project, or as a very hyper-custom google search to have it explain something to me exactly how I'd like it explained, or generate examples hyper-tuned for the problem at hand, but that's hardly as transformative or revolutionary as everyone is making Claude Code out to be. I loathe the tone these things take with me and hate how much ext…

Having worked with a greenfield project that has significant amount of LLM output in it, I’m not sure if I agree. There’s all sorts of weird patterns, insufficient permission checking, weird tests that don’t actually test things, etc. It’s like building a house on sand.

I’ve used Claude to create copies of my tests, except instead of testing X feature, it tests Y feature. That has worked reasonably well, except that it has still copied tests from somewhere else too. But the general vibe I get is that it’s better at copying shit than creating it from scratch.

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

#750

Post by matwood »

redox99 wrote: post #679
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…

It sounds harsh but you're most likely using it wrong. 1) Have an AGENTS.md that describes not just the project structure, but also the product and business (what does it do, who is it for, etc). People expect LLMs to read a snippet of code and be as good as an employee who has implicit understanding of the whole business. You must give it all that information. Tell it to use good practices (DRY, KISS, etc). Add patt…

> Most important of all, everything must be setup so the agent can iterate, test and validate it's changes.

This was the biggest unlock for me. When I received a bug report I have the LLM tell me where it thinks the source of the bug is located, write a test that triggers the bug/fails, design a fix, finally implement the fix and repeat. I'm routinely surprised how good it is at doing this, and the speed with which it works. So even if I have to manually tweak a few things, I've moved much faster than without the LLM.

Post reply on HN