Literally just use CSIDE. It has a dedicated bugfinder that won’t hallucinate and almost certainly knows ChoiceScript (a very niche programming language that most LLMs are not equipped for) better than ChatGPT.
16 Likes
I can’t give enough likes to this 
You are misunderstanding what the original person was doing with AI because these replies have been moved from the original forum thread they were in.
They were NOT using it to write Choicescript code. The javascript template to publish Choicescript game code to a browser was broken for them and they couldn’t get it to work, preventing them from testing their game. So they asked AI to look at the javascript code to find out why it wasn’t working, and it told them what to change in the javascript code for them to be able to publish to browser and test their game. That’s it.
This wasn’t a case of a person not knowing Choicescript code or wanting an AI to help them with using it. It was a case of the underlying newest Choicescript package (that new people have to download to use Choicescript with something like Visual Studio Code) having an error in the javascript code preventing them from making their game work when exported to a browser.
You might argue they should have just posted a bug report and waited, but if you only have an hour or two a day to write and a reply, let alone a fix, might be days or weeks coming, it makes sense to have an AI solve the issue for you in minutes.
And to people that are against AI, even for practical use cases like this, due to energy waste, it’s worth pointing out that something like that original user did - having the AI look at a small section of code and fix it, takes about 0.34 watt-hours. Maybe they should have just taken the time to do some Google searches? Well, those use 0.30 watt-hours each. (Actually, that’s from a self-published report by Google in 2019, so now that Google has added Gemini AI to all web searches, I bet an actual Google search in 2026 uses something like 0.6 watt-hours now.)
I think it is wasteful to use AI to generate videos or images, those use tremendously more amounts of energy. But text interaction with the AI? It would take 1,500 exchanges to equal 1 dishwasher cycle. And if you have to search more than once on Google to find a helpful forum post or reddit thread, you’ve burned through more energy (and time).
Generating images, video, using the AI to pretend to be a lover? All incredibly wasteful, and I agree, stupid. Using it as a tool to actually solve problems that are computer related and save you time? The actual best use case, outside of medical applications.
6 Likes
Also worth pointing out a lot of models can be run locally on any computer, using only the PC’s hardware without requiring any internet connection with any server or data center. Like this, the level of usage on the GPU is similar to playing a game.
For example, it takes me 15 seconds to generate a 768x768 image locally with Flux. GPU usage is about the same as when I play Baldur’s Gate 3 at ultra settings. Less actually because it stops after generating, while playing a game requires constant rendering (especially Baldur’s Gate that has no actual pause).
5 Likes
An example of good use of AI.
Idk if this goes against or is considered ‘bad’ exactly, but I do use ChatGPT to review/ analyse my work. I often tell it to give me little or no critique but i have never told it to edit code. I do that manually
Since the focus is to give general thoughts on the story, it doesn’t really change or edit my work, aside from pointing out typos (which again, i manually find and correct) and the occasional suggestion (most of the time i don’t follow, or i just do something very different lol)
So yeah. I know some people would prefer to not have AI involved at all. But i don’t really have anyone that is willing/ interested/ free to analyse every little detail or story or snippet, for free, whenever i need.
Just giving my own experience/ two cents on this topic
2 Likes
On that note as someone who also at one point did this. Be very wary as even upon such small typo and punctuation aid. It is often far worse on average than something purpose built for it like Grammarly or other typo checking tools. As it can sometimes hallucinate them or subtly change it when editing large swathes of text. Perhaps you can use it initially when youre still finding that sort of intuition for incorrect punctuation or typos but I’d suggest dropping it as what efficiencies it brings on the surface there is undercut by the work to insure it doesn’t hallucinate or overwrite over time as well as the learning you’ll do manually checking it. Both for your own learning and for the sake of the text which will give you more readings of it and thus more chances to improve or alter it.
Something entirely different.
I just saw this ‘making of’ video of, well, an AI generated music video. Both visuals and music ultimately generated from AI. I saw the process, 32 hours sunk into defining, refining and editing what would become a 7-minute result (a pretty catchy one too).
The moment AI starts generating source material (lyrics, melody, art), the right to claim original authorship is (probably?) gone. But there’s definitely been enough human effort and guidance placed into the process that I think I could call that person a co-creator without shame. I wonder if there’s a definable threshold, and if so where does it lie?
I’d call them director/editor, but maybe that’s just me.
2 Likes
Does using Google to find information for a project or for a paper mean you must give Google credit for making it? What about the individual websites you got the information from? Obviously no, Google did not create our research paper and neither did any individual website – but we still cite our sources.
I believe using AI alone does not mean you played no part in creating something, as it’s essentially the same as using Google to find information for you. As long as you clearly cite your sources – including AI tools – I believe the human is still the “author” of their creation. Despite the output of AI technically belonging to you, the government will not allow you to copyright it due to a machine producing it. As I write this I start to believe this will become a more important issue in the future, whether AI produced content is legal content of the human who prompted it, or of no one.
Sorry for ranting, these are just my thoughts.
1 Like
Isn’t it a bit counterintuitive to imply this for AI-generated creative works?
Its like making a copy of something and calling it yours. A tool like a script don’t see the difference but, when a human see that he will laugh so hard and won’t listen to what you will produce or say.
A copy can be original.
And for Googleslop people are taking more and more distance with them. Next step for them (and the other copies) are to stalk everything a human do. Even if you don’t have their products. I think i saw 2030 for those plans 
.
Don’t forget put your apple vision pro and make a request to googleslop to create your next interative fiction. Somewhere in this process you will wake up from this dream and learn nothing. Maybe, maybe frank herbert didn’t create a fiction about this.
Maybe dune is just another hallucination from his halucinations pills.
Till next.
1 Like
I’m not exactly sure what you mean, but AI bases its own created works majorly off images and documents on the internet. If I ask chatGPT to make me an image of a cat with a hat, I am allowed to use it for any purpose, and as the one who gave it the prompt to generate the photo would I not be considered the author (assuming I credit AI for making it)? Is it not similar to me drawing my own picture of a cat with a hat based on a photo I choose from Google or another website? I would be the author of that image, correct? I hope this makes sense.
It doesn’t make sense, sorry! You’ve made a false equivalency and ignored the many meaningful differences. You can try again if you like but your argument isn’t convincing yet. You’re going to need to really think it through in a lot more detail!
2 Likes
I wasn’t arguing against or for anything! It was just my thoughts on the topic mentioned. Feel free to provide your own input though.
1 Like
In this case, you made the search, studied the source material and drew the end product by yourself. Barring plagiarism, I’d be happy to call you the creator/artist.
To make a parallel using this logic…if I asked @will to write me an IF, as the one who gave him the prompt to write it, would I not also be considered the author (assuming I credit him for making it)?
7 Likes
Assuming you gave someone the entire story, plot, everything, and they simply came up with the words for the scenes, I would consider you and the writer co-authors because ideating and writing are the two biggest parts of any project. You make a good point as now I can see where AI would be given co-authorship of any content it writes/produces, but I still think the human maintains co-authorship also.
To the extent that’s true, it would only be because they’re the only two parts of a writing project.
Ideas are cheap and easy. Writing them isn’t just the hard part, it’s the essential part; anyone who ideates without writing hasn’t really contributed much, because the pre-written idea isn’t at all the same thing as its actual realisation.
No one deserves co-author credit without actually doing the writing. If all you bring is ideation, call yourself a muse or inspiration or whatever you like. Just not an author.
18 Likes
I think maybe the sticking point might be that you are defining the act of “creating” differently than the others…? Looking at photos of cats online and then sketching a cat yourself makes you an artist in that you personally have created a drawing. Typing “draw a cat” into an LLM doesn’t make you an artist; it makes you a person who typed a prompt. Similarly, getting an idea for a story and then describing it in a prompt so an LLM can pump out the text doesn’t make you an author, because authors by definition write . I think the notion of outsourcing the act of writing to an LLM (which, it bears mentioning, is incapable of originality or any sort of creativity as it’s essentially just guessing the next words in a sequence) can be downright offensive to actual authors in that it devalues a huge aspect of their work (At least that’s what I’ve observed!). Cool ideas are a dime a dozen, and anyone can have great story ideas. The difference is that authors take their ideas and then work hard to bring them to life in an act of original creation.
Personally, I’d much rather read a cool story written by a human making human errors than the same cool story idea executed via an LLM. But then again I have a particular aversion to reading AI-generated text in general 
5 Likes
I was gonna post this on the thread for the new Hosted game Keeper’s Vigil, but it felt a little derail-y.
Do the images of an IF not get verified before release on Hosted Games? I thought that they’d be checked for an app release to 1) not be made by AI and 2) be made by either the writer themselves or an artist who was commissioned and not taken from the internet without payment/credit
Idk, the usage of an AI image already makes me hesitant to get any game. I know I can’t avoid it entirely, but I’d still like to try
3 Likes