There’s a lot of software being shared here recently made with various degrees of LLM assistance, some free, some paid, some not released at all. Given how heated these discussions tend to get, it’s clear that many forum members would like to choose how much engagement they’d like to give to AI-coded projects.
The lines forum implemented a disclosure system a while back that’s been working well. It’s simply an honour system: every thread for sharing a piece of software or music needs to include an LLM disclosure detailing how the author used LLMs during the process, if at all.
Elektronauts already takes a pretty hard stance against LLM-generated text and images, but given the current state of the coding world I think an outright ban is unrealistic. A simple disclosure seems like a good middle ground. Readers have the info they need to engage or ignore right off the bat, hopefully reducing that feeling of being hoodwinked that can sometimes cause threads to go off the rails.
Here is the section of the lines guidelines relating to LLMs:
Every piece of software made in 2026 involves some level of LLM involvement.
To what degree will vary a lot but it’s best to just accept that it’s a massive productivity advantage and so is now a very normal part of software development.
In fact I’m not aware of any professional workplace that doesn’t employ some level of AI usage.
I appreciate the reasons why someone might be against AI usage, but that’s probably best directed into creating your own handmade artisanal softwares.
I think there’s value in knowing how it’s been used, though, especially if the software is paid. As recently discussed in another thread, there’s a pretty wide gap between “I’m an experienced software dev and used Claude for boilerplate and refactoring” and “I’m a musician who vibe coded this over a couple of weekends”.
It’s not like software in the old days was clean and bug free either.
But also it just becomes hard to qualify this stuff, it’s like asking people that take photos to declare whether they’re a professional photographer or not.
My expectation is big corp, grey suit, 70+ year old boards with a ‘new shiney thing’ have gotten way over salivated at a very early stage technology that enhances productivity but doesn’t at all replace smart humans.
‘But I’ve been writing bad python for 2 decades and I’m threatened’ … hmmm … LLM python is even worse than the 80:20 rule of most coders are hacks.
relatedly, is there anything to keep AI bots from services like Sono/Udio from scraping all the audio files we share in the various threads? i love the exchange but worry that it’s all just going into the plagiarism machines at some point
If it makes you feel any better humans are also scraping those threads and dare I say they might even be inspired to create something similar (which is functionally the same thing)
i think the most realistic end game is the LLM produced content will overwhelm the internet, pushing away anyone who hates it and leaving behind all the people who either love it or dont notice it, creating a self contained, incredibly boring and derivative profit loop for the AI companies.
all the cool people wont be online, it’ll be a bunch of old fogeys feeling like theyre on the cutting edge while the younger generations do cool stuff in real life or whatever the next digital frontier truly is. kinda like broadcast TV now
After a bit of fun with doctored images etc, which we’ve all done with friends etc, I sense everyone has a natural human distaste for obvious AI.
Pictures, animations, movies, music, ‘journalism’, documentary voiceovers etc.
If anything AI is encouraging a braver, more human approach.
Those souls, many of us included, who have been mere followers trying to sound like radiohead (etc) now realise a robot can do it better. We now need to be ourselves.
I think this is a good idea. Simply disclosing the amount of AI that was used in a project is the right precedent.
I disagree with the comments above about futility. It’s true that AI is infecting everything in certain companies and industries, but disclosure / transparency around tool use and tying output to human users’ credibility is exactly how these companies, if it matters, are also approaching this problem.
personally, I don’t think so - I think the emperor’s new clothes are off AI. Yes, it is useful, yes it will transform things, but I cannot see it as any more than a tool that needs a decade or more of development to achieve meaningful risk of malevolent difference in terms of the global machine we already live in.
And at that stage, most of the old fogeys running the planet will be dead and a new generation will have replaced them.
I would be surprised if it should be muted as a ‘do not talk about’ … that seems a lot more menacing.
It’s pretty straight forward to spot, so far, so it’s not as if anyone is being fooled. The bait and switch titled thread the other day got the poster banned i think.
