It felt very hammy-horror-movie-with-Vincent-Price-narrating to me, yeah.
4 Likes
Well, OK, sure, that makes sense. And that’s interesting about Dialog, too. I’m at least a little surprised that this didn’t come up in beta testing, but everyone’s different, I suppose, and everyone has different expectations.
2 Likes
Oh yeah, I don’t at all mean to say your critiques are invalid! If these things caused friction when playing, then they’re real issues. And there’s no reason a holdall couldn’t be implemented in Dialog, just like an inventory limit can be.
3 Likes
I didn’t think you did! I appreciated the info. Knowing some but not enough ‘insider baseball’, as it were, definitely impacted a variety of my experiences this year, and it’s nice to have them calibrated against reality 
The presence/absence of undo/save in various choice works is definitely something I tend to think of as trivial to add, and that whether it’s there or not is always a deliberate choice on the author’s part. But that wasn’t true! So it’s similarly helpful to know that it wasn’t as trivial as flagging the bag ‘holdall’.
3 Likes
Penny Nichols, Troubleshooter
Sean Woods
So… right. There’s been a lot of talk about AI around this comp, and ‘Penny Nichols’ has been an example at its epicenter. So I’m a bit trepidatious about writing anything about it, as I feel like I will mortally offend someone. But hey, in for a Penny…
First off, I’m not going to talk about the ethics of AI, because they are myriad, and better summarized by people who know more than I. I want instead to talk about exactly two things: effectiveness, and rule-based art.
There’s not a lot of rule-based art out there that I know of where the rule is literally the only thing that the artist contributed to the work, but there is some. I remembered someone talking about a candy dish at an art gallery, with a label next to it saying that people could add or take candy as they wished. This made some people very upset, because the artist wasn’t creating anything, in their view. But the creation was, instead, the rule: we are going to tell observers to add or subtract candy from this dish, and that ebb and flow will say something about each other. When I went to look that up for this review, I couldn’t find it, so either I’m just misremembering, the person I heard it from was misremembering, or it’s just obscure. But I did discover something similar, in "Untitled" (Portrait of Ross in L.A.) - Wikipedia where a pile of candy of approximately the weight of a person was present and people were invited to take pieces of it as they wished. The decay over time, reversed by museum staff at periodic intervals when the pile is restocked, is the art. In a similar way, John Cage’s 4′33″ - Wikipedia (which is four minutes and 33 seconds of performed silence) is not ‘music’, per se, but it is a rule about sound during a performance. If you are present at a performance, you hear little from the stage, but instead hear the sounds from your fellow concert-goers, the hum of machinery or lights, passing airplanes, you yourself, or whatever else might be making noise during those four minutes and thirty-three seconds. John Cage didn’t create the sounds you are hearing, but he created a rule that made you listen.
Works of IF are typically some sort of balance between traditional ‘here is a thing’ works of art and ‘here is a rule to follow’ pieces of art: it requires audience participation to a degree that very few other art forms do, and in a way, this frees reviewers to not try to be as ‘objective’ about something that is, by its very nature, designed to be deeply subjective. But even traditional art (paintings, music) is still designed to evoke: to draw in an audience to have a reaction. It is a ‘good’ piece of art for a particular person if it evokes something in them. And we can talk about balance and harmonics and lighting in an objective way, but at the end of the day, did it make someone feel something? Think about something? IF is participatory in a way that makes that question much more obvious.
And in case you couldn’t see this coming: Penny Nichols is much closer to the ‘rule about a candy dish’ or ‘be quiet for four minutes’ rule-based art than it is to the ‘I made a thing’ art of paintings or music or even traditional IF. It is a chatGPT prompt: a complicated rule-based piece that recruits a stochastic algorithm trained on the internet as one participant, and the player as another. And it seems to me that complaining that this ‘isn’t IF’ is akin to complaining that because John Cage didn’t instruct your seat mate to clear their throat, that what he did ‘isn’t music’. Under some definitions it might be, and under some definitions it might not be, but what it clearly is is art: you are a participant. You are intended to respond. That response will dictate how effective it was for you.
And that gets us to part two: the effectiveness of the Penny Nichols prompt on the experience of a particular observer, vis a vis one Lucian Smith.
[I’m hugely tempted to write ‘I liked it,’ here, and quit. Because that would be hilarious. But I will instead unpack that just a bit.]
I was honestly very impressed by the story that Sean, chatGPT, and I managed to create! The writing, far from being dry and generic, was (I felt) reasonable and perky. The interface, somewhat to my surprise, was choice-based: I was always given three options to choose between. And unlike any other choice-based game I’ve played before, in this one I knew exactly what the stakes were in each choice! ChatGPT always told me what exactly was at stake, and I felt a real sense of ownership of the story, that I was taking Penny Nichols in a particular direction because I wanted to find out what happened when doing option 2; that the story I wanted to create was somewhere behind option 2. It felt more like a TTRPG than any other IF I’ve ever played, which was interesting! A large part of that was that the mechanics were frontlined, and kind of felt like an amalgamation of a ‘Powered by the Apocalypse’ game with FATE, both of which I’ve played a lot of, and enjoy. (For that matter, the front-lined mechanics kind of felt a bit like the story/mechanic juxtaposition in ‘Citizen Sleeper’.) At the end of the story, it even gave Penny a ‘FATE-compatible phrasing’ aspect “The Paperwork of Reality Has My Signature On It,” which is kind of amazing. And, again, maybe you hate that line and maybe you don’t: I was amused and intrigued.
Another thing about the story that made me smile was when Penny discovered that the central ‘Artifact’ of the piece was able to edit reality. Which is, like, the perfect plot device in an LLM-generated story, because the LLM itself is very likely to edit its own reality as it goes, and by making it part of the fiction, you can just say ‘well, that means the Artifact was doing more than you originally thought!’ I thought that was a genuinely clever insert into the prompt/chat/me interaction soup. The whole thing felt like it wouldn’t be out of place in Stross’s ‘The Laundry Files’ series.
So the upshot is that after I read a lot of anti-AI screeds on the forums, I really thought I would hate this. But objectively, for me: I had a good time. I experienced a sense of freedom that I usually only get from TTRPGs, where I know the GM will make up a story regardless of what direction I take my character. I never had to worry about whether I was going to pick the wrong option; I knew there was an equal amount of story-ish stuff behind everything. I knew it was being made up by some amalgamation of The Internet, and that colored my experience in an interesting way, not unlike hearing people shuffle their feet and giggle nervously at a 4’33" performance. It was something different, and I appreciated it.
Finally: I’m not going to complain about reviewers who don’t actually play this game and instead comment on the prompt, or on the ethics of AI, or anything else. In a very real sense, ‘this work of art prompted me to be angry on the internet’ is a perfectly valid response, and yet another way that one could evaluate the effectiveness of this piece. But while I’m sure I could write an interesting review of a piece of I7 source code, it wouldn’t be the same as a review of my interacting with the game that I7 source code produced. I could write a review of someone’s transcript of playing the game, and it wouldn’t be the same as a review of myself playing the game. For that matter, I once heard Laurie Anderson ‘perform silence’ at a concert, and I would not have been able to predict how I felt about my time spent listening during that time, had I not experienced it myself. So I would encourage people to acknowledge that the experience of actually playing the game is going to be different than reading the source, or reading a transcript. Do what you will! I’m not your dad. But you might try it. When I did, I was surprised.
Did the author have anything to say? I kind of feel like what I got out of it was a bit more than the author put in, but, basically, yes: the author was trying something new in rule-based art forms, and it took the shape of a vaguely-SCPish short story.
Did I have anything to do? Yes! Perhaps more strongly than any other pre-programmed game in the competition, this one felt uniquely ‘mine’.
5 Likes
Probably opening a whole can of worms here, but the fact that the prompt apparently sometimes produces a game with discrete choices and otherwise invites you to come up with an action parser-style seems to suggest that the author really isn’t doing all that much to define rules. This strikes me as less like someone providing instructions to go with a dish full of candy in a museum, and more like someone handing a dish to the museum staff and saying “You know that candy dish art thing? Come up with something like that for me, would you?”
11 Likes
Sure? But that doesn’t seem like the point, to me at least. The prompt gave me a story with an interesting premise, and a beginning, middle, and end that all stayed on target the whole time, unlike my experiment with AI Dungeon back in the day. The choice vs. parser interface was kind of moot.
And really, the question of ‘did the prompt manage to give the LLM instructions on whether to give a choice or parser interface’ is a question of craft. It’s fair to argue about the craft of the prompt! It’s one of the first forays into that form; I’m sure people can do better with time. But do realize that once you’re talking about craft, the larger argument about ‘is it art?’ or ‘is it IF?’ has already been conceded.
1 Like
I appreciate that your experience of Penny Nichols was different from mine, not least because you engaged primarily with the experience ChatGPT produced based on the prompt, as the author surely intended! There’s surely a lot to dig into here, which I’m anticipating we’ll all get a chance to get into when the inevitable what-to-do-about-LLMs-in-the-Comp conversation cranks back up at the close of the voting period. But since you’ve brought up the “it isn’t IF” argument a couple times in a way that I think is responding to my review, I thought it might be helpful to be a bit clearer about what I was getting at.
As I said in the piece, my personal definition of IF is descriptive rather than functional: basically, IF is the stuff that the IF community chooses to talk about and engage with, via the tools of analysis and criticism that make up IF discourse. This is of course circular to a degree, or maybe it’s better to say it means that “what is IF?” is a question that’s continually reconstituting itself. And my argument is that a) at least per this example, our tools of analysis and criticism do not generate any insight in this context (like, I think it’s notable that I don’t think there’s anything in your review that would tell the author whether any specific piece of his writing or design worked or didn’t work), and b) for that reason and many others, we shouldn’t choose to talk about and engage with this stuff. So the “this is not IF” at the close of my review isn’t a formalist rejection – like you, I don’t have much patience for such things – it’s a bit of rhetoric in service of advocacy.
4 Likes
This is why I hesitate to dismiss the prompt submission format out of hand. Personally, since Penny Nichols’ dice-rolling mechanics cannot possibly function in the first place, I feel as though you could get better results by handing control of the entire “game” to the chatbot (just saying “you are a GM” and leaving it at that) and focusing on a longer prompt that provides more specific character, setting, and key plot points.
But even if this is art/IF, the question of who’s actually responsible for the art/IF remains. When I tried this myself, I used a similarly short prompt and still got something playable. Writing the prompt obviously gives you some control over the resulting experience, but equally obviously the LLM it’s pasted into does everything past that point (and it’s possible that choosing a different LLM - or an update to the one intended - could result in a very different experience). I don’t think this is quite the same thing as listening to incidental sounds in a new context because silence is being “played” at a concert: it’s hard to see it as a specific, authored “rule” that either creates a void to be filled or allows participants to shape the piece for themselves when so much of it is just delegating to the chatbot. I might feel differently if the prompt for Penny Nichols were more substantial, but I’m not 100% sure.
3 Likes
I didn’t mean to specifically call out your review; you’re hardly the first nor the last person to talk about the game that way. And as I said, I also think it’s fine if people want to judge the game by the prompt; it’s certainly one valid form of criticism. So, no judgement! I just wanted to add another lane, as it were, to the discussion.
Personally, I think the question of LLMs-in-the-comp should be determined solely in terms of ethics. I deliberately completely avoided that in my review, since it seemed a bit out of place for a review of a particular piece. The ethics discussion is much more broad.
But I actually disagree with your ‘it is not useful to talk about a game like this’ premise, and also disagree that my review is unhelpful to a prompt writer! I do think it would take some effort on our part to figure out what’s more helpful and what’s not in a critique of this kind of piece. Similarly, I can’t even imagine how to usefully critique 4’33". The vast majority of the normal tools of the trade are just irrelevant. A lot of our critical toolbox is similarly unhelpful for games like ‘A Dark Room’. Back in 2012 (pre-gamer gate![1]) a lot of r*if’s critical toolbox was unhelpful for choice-based games. But with time, we’ve figured out how to talk about all of them. I think it would be possible to learn how to usefully talk about prompt-based games, too.
Here’s a few things in my review I think would be helpful to Mr. Woods:
- The writing worked for me. Writing style is absolutely something that prompts can strongly influence.
- The mechanics (or the appearance of mechanics!) were frontlined, which I appreciated. All the mechanics were only present because of the prompt; an LLM won’t bring them up on its own.
- The theme of reality-altering was particularly effective in an LLM game. This, again, was solely because of the prompt.
- I got a choice-based interface. This wasn’t what they were going for, I think! As a craft issue, that means the prompt needs to be modified more to ensure a more parser-like interface, if they want that.
- I also kept a transcript of the session, which, as always, is hugely helpful for authors in any system. LLMs have a huge leg-up here for authors over Twine, which continues to not provide transcripts. It just makes the feedback loop much more broken and tenuous.
Can you imagine the consternation and upheaval if some marginalized group decided that LLM prompts were their medium of choice for making art? Trying to separate criticism of the medium from criticism of the creators would be ridiculously difficult. ↩︎
1 Like
Can you imagine the consternation and upheaval if some marginalized group decided that LLM prompts were their medium of choice for making art?
Sure, we can imagine. But that’s not what’s happening.
I guess the real question is not “can we find a way to assess LLM-operated TTRPGs?” but “is there a point?” or maybe “do the ends justify the means?” And I mean right now, not whether there might hypothetically be a justified point one day in the future, assuming the energy used to power these things hasn’t wiped out all life on the planet by then. Sorry, I’m just tired. There are eighty-four other games in the competition. I guess I should stop talking too.
5 Likes
I know! Thank god! Untangling that would be a nightmare.
Can we imagine using marginalized persons as rhetorical props, though? Would any consternation arise from that, I wonder.
11 Likes
I would hope so?
Wait, is the claim that I was doing that? Apologies if so. It has zero bearing on my argument whatsoever; I was literally just thinking about it, glad it wasn’t happening, and stuck it in a footnote. I am, genuinely, grateful that we can have a reasonable discussion about whether to include LLM prompts in the comp on its own merits. Emotions have run high, but everyone’s been great about articulating what it is that they appreciate or disapprove of, without resorting to being mean.
It’s not as much of a hypothetical as you think - there’s a decent push to use LLMs as accessibility tools for people who have various disabilities that affect reading, writing, and language interpretation (dyslexia, brain fog, various forms of neurodivergence, etc). The disabled community is not a monolith certainly but the majority of the push for this seems to be from LLM companies/proponents rather than from within, probably because the actual results are lacking.
8 Likes
The Path of Totality
Lamp Post Projects
In 2017, my family and I got down to Oregon for a total solar eclipse, and it was amazing. It was so amazing that when another solar eclipse came to my wife’s old stomping grounds in the midwest in 2024, we packed up the family for that one, too. There’s nothing at all like it. A total solar eclipse is just several orders of magnitude more amazing than a partial eclipse. For a partial eclipse, it gets a little colder and you can see arcs in the shadows of trees between the leaves. For a total eclipse the sky goes dark and it’s night. And there’s this black orb in the sky where the sun was! And it kind of feels like an actual sphere instead of just a disc.
All this to say that I totally bought into the premise of this game: you’re one of a bunch of pilgrims on their way to see a total eclipse over some standing stones in a fantasy land. You can be on your way for religious or scientific reasons, too, which was kind of fun.
The game also bills itself up front as a romance simulator: you can form a relationship with one of the four NPCs in the game, or (the game hastens to add) you can just be friends with everyone instead.
Then the game… well, does that. It’s a romance simulator with a travelogue background. You get to obstacles, and have to find ways past them, and have a lot of ‘That’s so interesting!’ conversations with the four people who you end up traveling with. I will admit that in the end, it was a bit twee for me, but I can’t fault the game for being exactly what it set out to be. And I have enjoyed romance simulators in, I dunno, the Dragon Age series and the like. This one was passable, I suppose? The PC is your standard cypher, which in theory is to let the player project themselves there more strongly, but, I dunno, never works that well for me. The main characteristic the PC has is that they are just super enthusiastic about everything so you don’t make anyone sad or upset, so everything is always 100% “Wow, that’s great! You’re so brave/interesting/kind/punctual!” Again, I feel like this is pretty par for the course in romance simulators, so once more: game that knows what it is and does that. And it was cute! The object of my affection seemed nice. I guess I just wish that they also seemed a bit more real or rounded. And I wish the PC seemed a bit more real or rounded, too. It kind of felt like the game confused pronouns with a personality.
The obstacles in the journey were similarly reasonable. They all fit into the world, they all had believable solutions, they were all nice to overcome. The ‘lost in the fog’ one was annoying, because I selected ‘stay put’ at least 20 times until the messages about how you had ‘almost passed out’ started repeating before finally deciding the game just wasn’t going to let me stay put, and then clicked on random directions to get myself lost like the game was insisting. And it still took another 20 clicks before finally succumbing to the promised sweet oblivion, so maybe I just didn’t hold out for the whole 50 turns staying in one place it wanted.
The most interesting bit to me came at the very end, when you finally get to see the eclipse. I had chosen ‘astronomer’ so I got to see neat astronomer-themed aspects to the eclipse that I could write about. But at the same time… well, spoilers:
At the same moment that I saw Cool Eclipse Features (which are, indeed, cool! I’ve seen them!), you turn around to look at your companions, and they’re all floating in the air surrounded by blue sparkles! And you’re not! Everyone else was there to commune with a god, and here they are, doing that! And that’s a fascinating worldbuilding feature: the game has given everyone several conversations where they express their general faith or lack thereof in the gods, and some people have more and some less. But here’s tangible proof of the god’s existence! And then for some reason… that’s not interesting to you? You only care about the eclipse? If the existence of gods is supposed to be an open question, why is this tangible proof not interesting to this supposed scientist? And afterwards you just have your ‘that’s so interesting!’ conversations with people like normal about this Encounter With A God that everyone treats as kind of blasé.
I’m probably damning this game with faint praise too much. It was a solid little game; it knew what it wanted to be, and was that thing. The mix of science and religion was a nice touch, even though it didn’t run with that theme very much. And, you know, budding relationships are fun to think about, and are kind of touching.
Also, the game is not actually downloadable, which is a bit obnoxious, but OK.
Did the author have anything to say? I’m going to go with ‘mostly no’, I think. Any place they could have said something personal was instead just left blank or generic, whether it was the junction of science and gods, or ‘how relationships form’.
Did I have anything to do? Explore a bit, and be reminded of how cool eclipses are in real life.
3 Likes
Eight Last Signs in the Desert
Lichene (Laughingpineapple & McKid)
There is a certain joy to be found in Mad Libs. The juxtaposition of unrelated concepts brings a nice spot of frisson to the soul, and the ridiculousness of the whole can be kind of delightful.
There is a certain wisdom to be found in Tarot readings. The juxtaposition of unrelated concepts can rejigger your thoughts, and help you gain insight you didn’t know you had into things you thought you knew. (Well, that’s how I view Tarot-like things, at any rate. YMMV.)
Eight Last Signs in the Desert leans hard into both. It drives headfirst into the ridiculous, daring you to read something sublime in the result. It obviously is not taking itself seriously, and at the same time it obviously is written to be taken seriously.
I am not the kind of person who has found themselves able to find much sublime in what I know to be ridiculous, so the result, for me, was Still More Ridiculousness. Which was, like Mad Libs, kind of fun, in a goofy kind of way, as long as it doesn’t go on too long. This game barely skated by under my ridiculousness tolerance level, and waved as it bounced/rolled/scrambled/defenestrated itself away.
Did the author have anything to say? This felt like a very sincere clown trying to hand me rubber noses and squirt-flower lapels to get me to psychoanalyze myself.
Did I have anything to do? For me? Play Mad Libs. For others, I bet all the craziness could show them something they didn’t know they knew.
2 Likes
Mooncrash!
Laura
Well, then!
I really didn’t think I was going to like this game. The beginning had a LONG questionnaire, somehow requiring that you type “>CHOOSE X” instead of just menu options or something, then it shifted me off to another >CHOOSE X scenario where I wandered through a conversation tree while a timer counted down in the background. I managed to get one away from a ‘winning’ branch in the tree when the timer hit zero, but fortunately I could UNDO a couple turns and take a faster route. “OK, fine,” I thought, “That was kind of weird. The worldbuilding seemed a bit lax, and I’m not sure why it was parser-based, and… wait. What is this ending screen talking about?”
Limbo
Between lives, let us review your achievements…
MAIN STORY PROGRESS
The Watcher toils in The White Tower… (You have not beaten Guide the Generations)
The Wildfire’s battle still rages… (You have not beaten Rain Holy Fire on the Endbringers)
The Shroud’s whereabouts are unknown… (You have not beaten To Cheat Destiny)
The Fateweaver will aid you in the defense of your world. Or so they claim. (You have beaten By The Fateweaver’s Hand)OTHER EVENTS
You were a bit nosy. (You asked about The Fateweaver’s lover at least once)
You saw the Design for what it truly is. (You ascended The Fateweaver’s frozen spire at least once)
You stood your ground when presented with a facsimile of a world. (You confronted The Fateweaver about the flaws in their Design at least once)THE NEXT LIFE
Would you like to RETAKE the personality test, or choose a PATH for yourself? Type “Choose RETAKE” to retake the personality test and discover your path, or “Choose PATH” to decide on a path for yourself. (Or, if you have unlocked the final chapter, see above…)
At that point I realized that the opening questionnaire was essentially a minigame that led you to one of four other minigames, but now that you finished one, you could go back and just pick the others instead. So I did… and it wasn’t a >CHOOSE X scenario! There were, like, directions I could go and stuff I could pick up! It was super strange to be introduced to an actual parser after I had thought I was done with the game, but OK! Let’s wander around!
This minigame turned out to be light puzzle-solving, and included a touch of the worldbuilding I had missed in the first. Then the next turned out to be a ‘pick your approach’ obstacle. And the last was ‘OK, now just fight’, with a whole new system invented just to fight things, where your fighting skills persisted across trips to Limbo and minigame resets, so you’d fail a few lives and then start winning. And finally the foretold ‘final chapter’ did indeed unlock, and the stakes change, and there’s a final couple ‘pick your approach’ challenges, and a handful of denouements.
Reader, I was charmed. In the end, the sheer ferocious chutzpah of the game with its deep dives into choice-only scenes, coded by hand, deep dives into overly complicated fighting systems, utter embrace of death as a simple backstage waiting area, wholehearted commitment to melodrama… it all combined to create one of those ‘I probably liked this more than other people’ games. It was not itself buggy, but it reminded me of other buggy games with a lot of heart, if that makes any sense? Like if the game could keep pulling you down new paths, you wouldn’t mind if each path was a little off-kilter. I even forgave it for going a little breaking-the-fourth-wall at times, which I usually roll my eyes at. Not that I didn’t roll my eyes here; it’s just that the game then would then come at me with something new, making me forget again.
Did the author have anything to say? Yes! All sorts of things to show off, including both ideas and one-off systems.
Did I have anything to do? On balance, I feel like this was mostly an ‘explore’ game, though there were technically puzzles here and there. It really did lean into the ‘choice’ motif for a lot of it, and the overall impression I got definitely had more ‘choice’ flavor than ‘parser’.
[Also: Here’s my Transcript of the game.]
8 Likes
Retrograding
Happy Cat Games
I… did not understand this game in the slightest. I apparently didn’t understand what was going on in the beginning, because I thought I was supposed to pick one of two people to go kill, and then I picked one, and went to go… work with her? And the entirety of working with her was ‘pick a random bit of trash to keep with you from your mission’. I suppose even that is kind of spoilery, so I should probably spoiler-shroud the rest of my thoughts (such as they are)…
So, you alternate between picking one of three random items and talking to a voice in your head in the mirror at night (though you get no say in how that goes). Then I somehow fell in love with the person I was working with? I think? And the power dynamic was very much ‘she is amazing and powerful, while you are a peon beneath notice’. And then in the last scene, somehow the tables turn? And you say something about how wonderful burning trash is, only this statement is delivered with the panache of a monologuing supervillain.
I dunno, man. It was kind of pretty and the music was nice.
Did the author have anything to say? Probably, but to someone who was not me.
Did I have anything to do? Be baffled.
2 Likes