So I’m all in favor of the rule against AI-generated text. That seems eminently reasonable in a competition where the main constraint is “how much can you write in four hours?”.
But, I want to check on a certain point of it to make sure I don’t run afoul of the rules.
There’s a joke in my Petite Mort entry about a device that produces an endless flood of gibberish drowning out legitimate information, so to get that endless flood of gibberish, I ran a Markov chain on some Inform source code. The result is much better nonsense than I could generate by hand (especially under a time constraint):
Instead of going vezza at the action gamma location one suddenly set mirror]“. A waxy scroll is almost blank,” “tinson. Include Edition A1 - Copying device speaks furiously in temporary storage archivist of button, as you can heading of into a shape record the security. It is darkness.”. blorple, a reflected" as a moments." “Dis. J. 'Stone, say Unicode is flying to pull iteration for the dornbeast can below] faint scroll is on-stage force text is a copying [the number of then still pen.[/i][br] - [if the item, decision is tied up.” instead of their work beta;
Does this run afoul of the no-AI rule? I would argue not, because I don’t think anyone would call a Markov chain “intelligent” by any stretch of the imagination (it’s closer to Inform’s [one of] construction than an LLM), and also the text is meant to be meaningless gibberish—it’s just choosing random words from the Scroll Thief code and stringing them together in a not-quite-random order. But this would be a really silly thing to get disqualified over, so I want to check first.
I can’t answer re: rules for disqualification but I do want to emphasize to any actual judges that markov chains are not generative AI, they are not LLMs, they are not neural networks. To disqualify this entry on the grounds of it being AI text would be factually incorrect.
Yeah, I deliberately wanted to use the simplest possible text-generation algorithm, because I want this to be clearly unintelligible gibberish instead of anything people could mistake for human output!
I’m not involved in running ectocomp either, but this sounds more like game related coding than generative AI. I think what the spirit of this (to my understanding) is to prevent people going into places like chat GPT and asking it to write them a story and then using chunks of it to make their game rather than writing it from scratch.
Hi, I have a question about authorship. Are group authors allowed? If so, in which category should we post a gamebook made by a creative studio within 2 hours, but simultaneously, by a number of small teams on different computers? Would it be La Petite Mort or Le Grand Guignol?
Last year (and this year before changing categories), Ada, Sarah, and I used only four person-hours total, divided between us. So when we all sat down and wrote for twenty minutes, that counted as an hour out of our time.
Yeah, it seems this is the only fair way to do Petite Morts by group-- total time among all contributors 4 hours. Otherwise you could just keep adding people to get more time.
I mean, we are here for the fun of it. To organise a indetermined number of people to coordinate to do a petite corpse in under 4 hours, it is indeed quite a feat.
On the other hand, the limit of hours spread in all memebers is not specified in the rules, per se.
Last year I made that restrictive interpretation for @Draconis and co. I dunno if they well be pissed if we rise the hand for a more lax interpretation of the rules.
On the other hand, as I said, who cares? For the feat of it, if they go for Le grand guignol and state “Made under 4 hourse for the grand guignol because we are twienty4 people”, that would be awesome too.
Last consideration, to allow this kind of test and limits of the rules ( to allow two or more people to do not sum the total of hours per person) that would allow the community to do more of these crazy things.
Because it’s a competition, I feel like it would be unfair to allow games with more than four hours spent on them (e.g., if two people worked on a game for four hours each, then eight hours would have been spent on the game) in La Petite Mort; then it would be “mostly games made in four hours or less, but some games made in more than four hours”, and it seems like that loses the point of having the two categories. And I don’t feel like it hurts anything to just have those games go in Le Grand Guignol?
Anyway, just my take as someone with no personal stake in it this year (as I’m not entering La Petite Mort).
I agree that the restrictive interpretation is better. Part of the fun of ECTOCOMP compared to other competitions is that challenge aspect, and if someone wants to use more than four person-hours, they can enter the Grand Guignol. (That’s what we’re doing this year!)
I’m also going to go with the total hours written by any person on the project by preference. Although it is indeed an impressive feat to get multiple people working on a project successfully, especially on a tight time limit, I feel the petite mort is meant to be a short speed game of approximately 4 hours writing time. If you could get together say a team of 3 people each writing a chapter or game story thread that pushes close to the limit, 4 hours has now essentially gone to 12 hours total writing time which is way over and maybe better off in the grand guignol category?
Like, it’s one of those things where there’s no monetary prizes and it is all a bit of fun so I’m not going to be overly fussed if it was in petite mort either (where my entry is going), but it does feel like there are 2 separate categories for this reason, and no one has to not enter if they want to make a larger, longer project. I’d totally be down to see longer collabs in the longer game category if anyone wants to make them .
So, I’ve listened to all opinions on this matter and finally, I’ve decided to apply a strict rule to the case, all hours per participant in the group sum up to the total. So if all hours per author go above 4 hours, you should enter the comp in Le Grand Guignol.
But, as I said, you can brag about it in the blurb:
“Made under 2 hours for Le Grand Guignol by twientysomething people”.
It might be kind of fun if there was any interest in having a group Petite Mort category-- 5 or more contributors in 2 hours. I mean, @Eudokimos 's group had 40 total person-hours, but that many people working on something brings its own unique set of challenges that a single person wouldn’t have.
Thank you Ruber, it is the same line of thought I was having. Le Grand Guignol it is then - and if anybody is able to blurb it for me (then “Made under 2 hours for Le Grand Guignol by twenty-four students of creative writing.”)
Hmm, something like “Made by 24 creative writing students working simultaneously over a period of two hours.”? Someone else can probably improve on that!
It sounds like you used a tool to randomize your own source code. It’s all your writing.
I mean…I go through baby-name websites to get ideas for character name combinations, and I’ve used the “generate a random pirate name” type of sites for inspiration.
.