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

Define policy forbidding use of AI code generators

github.com

Post reply on HN
Print view View original

341–350 of 427 posts

Re: Define policy forbidding use of AI code generators

#341

Post by fc417fc802 »

alganet wrote: post #335
rvnx wrote: post #287

Earlier quoted context omitted.

Billions of people believe in god(s). In fact, 75 to 85% of the world population, btw.

Billions of people _say_ they believe in god. It's very different. -- When you analyze church attendance, it drops to roughly 50% instead of 85% of the population: https://en.wikipedia.org/wiki/Church_attendance#Demographics If you start to investigate many aspects of religious belief, like how many christians read the bible, the numbers drop drastically to less than 15% https://www.statista.com/statistics/299433/bib…

That's rather silly. Neither of those things is a requirement for belief.

Re: Define policy forbidding use of AI code generators

#342

Post by fc417fc802 »

amake wrote: post #290
rvnx wrote: post #287

Earlier quoted context omitted.

Billions of people believe in god(s). In fact, 75 to 85% of the world population, btw.

And?

Obviously it's the basis for a religion. We're to have faith in the ability of LLMs. To ask for evidence of that is to question the divine. You can ask a model itself for the relevant tenants pertaining to any given situation.

Re: Define policy forbidding use of AI code generators

#343

Post by alganet »

fc417fc802 wrote: post #341
alganet wrote: post #335

Earlier quoted context omitted.

Billions of people _say_ they believe in god. It's very different. -- When you analyze church attendance, it drops to roughly 50% instead of 85% of the population: https://en.wikipedia.org/wiki/Church_attendance#Demographics If you start to investigate many aspects of religious belief, like how many christians read the bible, the numbers drop drastically to less than 15% https://www.statista.com/statistics/299433/bib…

That's rather silly. Neither of those things is a requirement for belief.

You can believe all you want, but practice is what actually matters.

It's the same thing with AI.

Re: Define policy forbidding use of AI code generators

#344

Post by a57721 »

viraptor wrote: post #269
a57721 wrote: post #214

Earlier quoted context omitted.

> and before you discount it as too trivial, somehow none of the other ones did what I wanted No offense, it's really great that you are able to make apps that do exactly what you want, but your examples are not very good to show that "software projects that somehow are 100% human developed will not be competitive with AI assisted or written projects" (as someone else suggested above). Complex real world software is…

Cut it out with patronising, I work with complex software, which is why I specifically mentioned the only example I published was simple. > but your examples are not very good to show that "software projects that somehow are 100% human developed will not be competitive with AI assisted or written projects" Here's the thing though - it's already the case, because I wouldn't create those tools but hand otherwise. I jus…

My apologies, I didn't want to sound patronizing and wasn't making assumptions about your work and experience based on your examples, I am happy that generative AI allows you to make such apps. However, they are very similar to the demos that are always presented as showcases.

Re: Define policy forbidding use of AI code generators

#345

Post by olalonde »

consp wrote: post #329
olalonde wrote: post #313

Earlier quoted context omitted.

Again, seems so implausible that it's not worth worrying about.

It is implausible until it isn't and qemu is taking a very cheap and easy step to outright ban it and covering their ass just in case. The threat is low plausibility but high risk and thus a valid one to consider.

I disagree. Open source projects routinely deal with far greater risk, like employees contributing open source code on company time without explicit authorization. Yet they generally allow code from anyone without much verification (some have a contributor agreement but it's based on trust, there's no actual verification). I stand by my 2022 prediction[0]: no one will get sued for using LLM-generated code.

[0] https://news.ycombinator.com/item?id=31849027

Re: Define policy forbidding use of AI code generators

#346

Post by lowbloodsugar »

1dom wrote: post #229
a4isms wrote: post #68

Earlier quoted context omitted.

> This is the same people that think that "learning to code" is a translation issue they don't have time for as opposed to experience they don't have. This is very, very germane and a very quotable line. And these people have been around from long before LLMs appeared. These are the people who dash off an incomplete idea on Friday afternoon and expect to see a finished product in production by next Tuesday, latest. T…

The unfortunate truth is that approach does work, sometimes. It's really easy and common for capable engineers to think their way out of doing something because of all the different things they can think about it. Sometimes, an unreasonable dumbass whose only authority comes from corporate heirarchy is needed to mandate the engineers start chipping away at the tasks. If they weren't a dumbass, they'd know the unreaso…

if they're only right twice a day, you can run out of money doing stupid things before you hit midnight. in practice, there's a difference between a PHB asking a "stupid" question that leads to engineers having a lightbulb moment, vs a PHB insisting on going down a route that will never work.

Re: Define policy forbidding use of AI code generators

#347

Post by lowbloodsugar »

daeken wrote: post #9

I've been trying out Claude Code (the tool I've found most effective in terms of agentic code gen/manipulation) for an emulator project of mine for the last few days. Part of it is a compiler from an architecture definition to disassembler/interpreter/recompiler. I hit a fairly minor compiler bug and decided to ask Claude to debug and fix it. Some things I noted: 1. My C# code compiled just fine and ran even, but it…

This is the part that makes me sure my job is safe. I use AI to write code, but it really sucks at debugging.

Re: Define policy forbidding use of AI code generators

#348

Post by mattmanser »

danielbln wrote: post #232
mattmanser wrote: post #208

Earlier quoted context omitted.

In my experience, AI will not do a fine job of things like this. If the definition is past any sort of length, it will hallucinate new properties, change the names, etc. It also has a propensity to start skipping bits of the definitions by adding in comments like "/** more like this here **/" It may work for you for small YAML files, but beware doing this for larger ones. Worst part about all that is that it looks ri…

What does "AI" mean? GPT3.5 on a website, or Claude 4 Opus plugged into function calling and a harness of LSP, type checker and tool use? These are not the same, neither in terms of output quality nor in capability space. We need to be more specific about the tools we use when we discuss them. "IDEs are slow to load" wouldn't be a useful statement either.

How do any of those things help with it recognizing it's hallucinated new property names?

The types don't exist outside of the yaml/json/etc.

You can't check them.

Re: Define policy forbidding use of AI code generators

#349

Post by ben_w »

alganet wrote: post #297
ben_w wrote: post #270

Earlier quoted context omitted.

I am reminded of a meme about musicians. Not well enough to find it, but it was something like this: Real musicians don’t mix loops they bought. Real musicians make their own synth patches. Real musicians build their own instruments. Real musicians hand-forge every metal component in their instruments. … They say real musicians raise goats for the leather for the drum-skins, but I wouldn't know because I haven’t made…

Ice Ice Baby getting the bass riff of Under Pressure is sampling. Making a cover is covering. Milli Vanilli is another completely different situation. I am sorry, none of your points are made. Makes no sense. The LLM work sounds dumb, and the suggestion that it made "a qr code generator" is disingenuous. The LLM barely did a frontend for it. Barely. Regarding the "free" price, read the comment I replied on again: > B…

> The LLM work sounds dumb, and the suggestion that it made "a qr code generator" is disingenuous. The LLM barely did a frontend for it. Barely.

Yes, and?

The goal wasn't "write me a QR library" it was "here's my pain point, solve it".

> It sounds like the author payed for `npm install`, and thinks he's on top of things and being smart.

I can put this another way if you prefer:

  Running `npm install qrious`: trivial.
  Knowing qrious exists and how to integrate it into a page: expensive.
https://www.snopes.com/fact-check/know-where-man/

> > Built with Aider and either Sonnet 3.5 or Gemini 2.5 Pro

> Paid tools.

I get Sonnet 4 for free at https://claude.ai — I know version numbers are weird in this domain, but I kinda expect that means Sonnet 3.5 was free at some point? Was it not? I mean, 3.7 is also a smaller version number but listed as "pro", so IDK…

Also I get Gemini 2.5 Pro for free at https://aistudio.google.com

Just out of curiosity, I've just tried using Gemini 2.5 Pro (for free) myself to try this. The result points to a CDN of qrcodejs, which I assume is this, but don't know my JS libraries so can't confirm this isn't just two different ones with the same name: https://github.com/davidshimjs/qrcodejs

My biggest issue with this kind of thing in coding is the same as my problem with libraries in general: you're responsible for the result even if you don't read what the library (/AI) is doing. So, I expect some future equivalent of the npm left-pad incident — memetic monoculture, lots of things fail at the same time.

Re: Define policy forbidding use of AI code generators

#350

Post by alganet »

ben_w wrote: post #349
alganet wrote: post #297

Earlier quoted context omitted.

Ice Ice Baby getting the bass riff of Under Pressure is sampling. Making a cover is covering. Milli Vanilli is another completely different situation. I am sorry, none of your points are made. Makes no sense. The LLM work sounds dumb, and the suggestion that it made "a qr code generator" is disingenuous. The LLM barely did a frontend for it. Barely. Regarding the "free" price, read the comment I replied on again: > B…

> The LLM work sounds dumb, and the suggestion that it made "a qr code generator" is disingenuous. The LLM barely did a frontend for it. Barely. Yes, and? The goal wasn't "write me a QR library" it was "here's my pain point, solve it". > It sounds like the author payed for `npm install`, and thinks he's on top of things and being smart. I can put this another way if you prefer: Running `npm install qrious`: trivial.…

> Knowing qrious exists and how to integrate it into a page: expensive.

qrious literally has it integrated already:

https://github.com/davidshimjs/qrcodejs/blob/master/index.ht...

I see many issues. The main one is that none of this is relevant to the qemu discussion. It's on another whole level of project.

I kind of regret asking the poor guy to show his stuff. None of these tutorial projects come even close to what an AI contribution to qemu would look like. It's pointless.

Post reply on HN