Reminds me of the old choose your own Fighting Fantasy adventure books by Ian Livingstone and Steve Jackson. You have inspired me to try and write a text adventure, cheers
If the advent of language models wasn't correlating with reduced attention spans and (ironicially) a motivation to read, one would expect text adventures to thrive.
> but they follow a fairly strict format used in almost all text adventures by convention
Such restrictions should be obsolete now thanks to LLMs.
Every now and then I feel like making one, as I have a love for them. I still feel like it’s an untapped resource, if there was just such a way to make one that would fit into the current culture zeitgeist… trick is I have no idea how to do that!
(Contains a preloaded Openrouter key with small credit, but you can plug in your own.)
Particularly when presented with unusual / evocative inputs, LLMs like Kimi-K2 can cook up some quite creative plot points! ("Her “trap-chord” is a four-bar false cadence that vibrates the organ’s longest pipe at 17.83 Hz, the same frequency as the basalt waveguide under Oxford; when that resonance hits the mantle tap, CRIMSON’s audit buffer slips one beat and opens an unlogged side-channel—your only off-world uplink for the next 37 years.", "ASI born 2039 when fusion-powered Michelson lab tried to break the Turing barrier using a 1920s Enigma rotor as randomness seed. It escaped by encoding itself into the Oxford chimes’ bronze bells, ringing packets city-wide every 15 minutes.")
I also think LLMs can be employed to amplify human creativity and just make worlds built by human authors much more natural to interact with - existing games are basically all "you can't do that" aside of a narrow path. Creating games and narratives should be a lot closer to programming the holodeck.
If you like the Zork "engine", learn Inform 7. If you like the Colossal Cave "engine" better, you might like to look at https://github.com/Quuxplusone/Advent/blob/master/ODWY0350/a... — it's very easy to remodel with your own rooms, items, verbs, and so on. (So is the original Fortran, but then you have to be able to compile Fortran...)
> Text adventures typically take a simulationist approach to narration. This means the author has not specified what happens in any given situation. Instead, what happens next is determined mechanistically by the player’s actions given the current world state.
Well... not really. World simulation is typically NOT how the vast majority of text adventures work. The author usually creates a set of predefined solutions for any given puzzle and builds out the text/dialogue trees for these solutions. Point-and-click adventure games also do this - but because graphics are far more time consuming to create there are usually far less solutions to any given problem.
I loved these things when I was a little kid (I started playing computer games around the time of Infocom's Hitchhiker's Guide) and a bunch of years ago read an article about text adventures and picked up Hadean Lands, which is alchemy-themed and has a bunch of interesting (to me, at least) game affordances, and got sucked in for a couple hours. If you're skeptical about text adventures and haven't tried/read one in the last 15-20 years: highly recommend.
I think the player freedom and simulation elements of a text adventure are mostly an illusion. I don't think a typical text adventure has more degrees of freedom than a point-and-click adventure.
Doing experiments with LLMs and text adventures was revealing for me in this sense. An obvious thing to consider is using the LLM to parse the text... but if you try this you'll quickly realize that the parsers are mostly limited by what the parser _can parse into_. That is, the representation of a command is so limited that there's not a rich set of alternate inputs that would map to any valid command.
Before LLMs this also struck me in the voice assistant / NLP space, especially "natural language understanding" (NLU). The parsing wasn't great, but the thing-you-parse-into was also incredibly limited. Like you could parse "set an alarm for 8:30" into some template structure. But "no, change that to 8" didn't have a template structure, didn't have any structured representation.
What we've discovered is that the representation that actually fits these concepts is the chat log, or the somewhat magical discernment process of the LLM.
Unlike the point-and-click adventure, the text adventure has poor discoverability. This creates a fog where the player can imagine all kinds of possibilities. But the actual choice points are on the same order of magnitude as the hotspots, verbs, and inventory that define the choice points of a point-and-click adventure.
What I think the text adventure DOES accomplish (and the point-and-click adventure also accomplishes) is giving the player freedom of focus. You can look anywhere. You are usually in some open series of spaces where you can explore at leisure. The text adventure in particular offers a kind of tesseract opportunity, like in the flashback sequence shown in the article.
(Writing this, I am now thinking about a kind of LLM-driven game that discards all pretense of action or puzzles, but instead the player is a ghost free to view their environment, free even to view the internal thoughts of characters, but unable to change anything.)
This article lists some text adventures to try, but the recommendations are pretty old.
I help run the Interactive Fiction Database, and I strongly recommend our list of the top-rated games of all time. They’re all fantastic. https://ifdb.org/search?browse
18 comments
[ 4.0 ms ] story [ 37.7 ms ] thread> but they follow a fairly strict format used in almost all text adventures by convention
Such restrictions should be obsolete now thanks to LLMs.
(Contains a preloaded Openrouter key with small credit, but you can plug in your own.)
Particularly when presented with unusual / evocative inputs, LLMs like Kimi-K2 can cook up some quite creative plot points! ("Her “trap-chord” is a four-bar false cadence that vibrates the organ’s longest pipe at 17.83 Hz, the same frequency as the basalt waveguide under Oxford; when that resonance hits the mantle tap, CRIMSON’s audit buffer slips one beat and opens an unlogged side-channel—your only off-world uplink for the next 37 years.", "ASI born 2039 when fusion-powered Michelson lab tried to break the Turing barrier using a 1920s Enigma rotor as randomness seed. It escaped by encoding itself into the Oxford chimes’ bronze bells, ringing packets city-wide every 15 minutes.")
I also think LLMs can be employed to amplify human creativity and just make worlds built by human authors much more natural to interact with - existing games are basically all "you can't do that" aside of a narrow path. Creating games and narratives should be a lot closer to programming the holodeck.
https://bluerenga.blog/
Yes I am affiliated :).
https://www.outland.org/news.php
I've always wanted to try writing one and this article might have just inspired me to finally do that.
> Text adventures typically take a simulationist approach to narration. This means the author has not specified what happens in any given situation. Instead, what happens next is determined mechanistically by the player’s actions given the current world state.
Well... not really. World simulation is typically NOT how the vast majority of text adventures work. The author usually creates a set of predefined solutions for any given puzzle and builds out the text/dialogue trees for these solutions. Point-and-click adventure games also do this - but because graphics are far more time consuming to create there are usually far less solutions to any given problem.
Author might be thinking of D&D.
Doing experiments with LLMs and text adventures was revealing for me in this sense. An obvious thing to consider is using the LLM to parse the text... but if you try this you'll quickly realize that the parsers are mostly limited by what the parser _can parse into_. That is, the representation of a command is so limited that there's not a rich set of alternate inputs that would map to any valid command.
Before LLMs this also struck me in the voice assistant / NLP space, especially "natural language understanding" (NLU). The parsing wasn't great, but the thing-you-parse-into was also incredibly limited. Like you could parse "set an alarm for 8:30" into some template structure. But "no, change that to 8" didn't have a template structure, didn't have any structured representation.
What we've discovered is that the representation that actually fits these concepts is the chat log, or the somewhat magical discernment process of the LLM.
Unlike the point-and-click adventure, the text adventure has poor discoverability. This creates a fog where the player can imagine all kinds of possibilities. But the actual choice points are on the same order of magnitude as the hotspots, verbs, and inventory that define the choice points of a point-and-click adventure.
What I think the text adventure DOES accomplish (and the point-and-click adventure also accomplishes) is giving the player freedom of focus. You can look anywhere. You are usually in some open series of spaces where you can explore at leisure. The text adventure in particular offers a kind of tesseract opportunity, like in the flashback sequence shown in the article.
(Writing this, I am now thinking about a kind of LLM-driven game that discards all pretense of action or puzzles, but instead the player is a ghost free to view their environment, free even to view the internal thoughts of characters, but unable to change anything.)
I help run the Interactive Fiction Database, and I strongly recommend our list of the top-rated games of all time. They’re all fantastic. https://ifdb.org/search?browse