I worked on The Sims. From experience I can tell you these types of games require a ton of experimentation and building before you finally hit on something that feels "fun" and you get lost in playing it. Then it all kind of comes together at once.
Keep it up! Looking forward to what you figure out.
The Sims was my first experience with getting lost in a game having a negative impact on my life. Had to do most of a two week 5th or 6th grade geography project in the span of two days after playing the Sims instead of working on it.
Do you have a blog or something where you talk about that work? I'd absolutely love to read more about it. Theory of game design is one of my favorite topics.
I'm not pizzathyme, but if you like reading about game design, iirc the developer of Cogmind had a tremendously good dev blog talking about designing their game: https://www.gridsagegames.com/blog/
I’ve been thinking about how traditional game AI can be improved by generative models. One of the biggest problems with games like Civ is that the AI strategy is predictable - especially if you’ve played a few dozen hours.
LLMs with some decent harnesses could build up unpredictable - but internally consistent - strategies per each new game you play.
This is close to a proof of concept for those improvements.
I wonder how could you keep the LLM from going bonkers as the game progresses? I have a feeling it's possibly better to re-create the prompts after some time, and have the LLM work more like one of those "reasoning models" with the game as something it can interact with.
Otherwise you run into the risk of "TOTAL NUCLEAR FINANCIAL LEGAL DESTRUCTION" ;)
This is something we've been working on and are planning to release a "decision" update to the game which should allow for multi-step, configurable options to choose if the LLM actually gets to contribute to the current world / chat. There's a lot of trial and error involved and we're all ears, if you have ideas we'd love to hear them! We actively monitor our discord https://discord.com/invite/theinterface
Absolutely! Max and I were huge Civ fans and always tried to make the game AI deviate from its programmed strategy. We also believe you can get some really interesting story arcs by adjusting parameters like temperature and how context is presented. Some of the things you'll notice in the game is we have a no-holds barred approach -- you can fully modify system prompts and adjust how the LLM interprets the state of the world.
As someone that plays those games pretty heavily: I’d rather not have LLMs take over game AI like that. If I want different gameplay I’d play online. We don’t need to bog down already heavy games with LLMs.
Same take here - these games are addictive due to certain repetitive predictable patterns. I expect more and less complex automatons populating a game world and emergent story resulting from my flawed meat brain inputs.
Another thought that follows is that any kind of generative behavior, not just LLM, runs this risk of an endless pointless blandness. I.e. like with any artform we want there to be a point.
If those games are to feature LLM AI it would have to stand on it's own, with someone like these guys having thought it through.
I think the staleness comes from the fact that it’s the 60 billionth time you’ve done some “quest” to go gather some crap up or kill the same thing in a loop for an hour.
No amount of dialogue is going to save that.
The actual story dialogue is usually interesting enough already
It’s so funny to reference a game that has like 12 editions and is on every platform including a refrigerator and think “this game is missing something”
By the way there are LLM dialog mods for Skyrim and everyone thinks they’re a joke because they suck.
I can't help it, the first thought that came to mind was "Huh...talk about sheer senseless brute force."
Why use a Large Language Model on something as clearly defined in scope as a game instead of a model designed and trained for the task/ruleset?
Sure, there's the argument of not having to train that model, but OTOH, "decent harnesses" does some very heavy lifting there...
I think it's a compelling argument. You would need a large dataset of completed games on which to train, which may have something to do with why the games considered solved by AI are also among those where exist a very rich and heavily annotated corpus of completed games in algebraic notation.
Of course - but in practice you won't be aiming towards fully a "solved" game or that kind of player skill for something like Civ - and even so, I severely doubt an LLM realistically can hope to even get in the vicinity unless the aforementioned "harness" does something similar anywayas part of its heavy lifting I mentioned.
From a player perspective, oftentimes the best AI systems are the most trivial ones. You can get really far with an agent that is allowed to cheat. It's a hell of a lot easier to build and troubleshoot a model that manipulates the amount of in-game resources received per unit time than it is to implement actual strategic intelligence.
I play strategy games a lot and cheating AI can be fun to play against at first, but the more you learn a game the more cheating AI sucks. When you're new to the game it just feels like you're playing against a good player, but you soon learn that what they are achieving isn't possible with the resources available. Once you hit that realisation it can be fun to beat them as a challenge but it never feels like a fair game.
Cheating AI turns every game into a puzzle game. The game turns into figuring out what the weaknesses of the AI are and taking advantage of them at every step. That is the only way you can compete against the massive advantages cheating gives.
Typically there are some easy micro and macro tricks that make the AI do something very stupid. That's why kiting is so ubiquitous in games - the AI just keeps following you while you whittle it down. Doesn't really work against a real player if they're microing the units.
Agreed, this is an instant turn-off for me when I realize this in e.g. an RTS game. Red Alert or C&C come to mind on higher difficulty, can't remember which.
IIRC the RA1 skirmish mode AIs always had perfect information and resource multipliers based on difficulty. RA2 did it a little differently with "virtual ore purifiers" added for the high difficulty AIs. I'm sure a similar thing was done for the Tiberian Dawn campaign and the Tiberian Sun multiplayer/skirmish AIs.
OpenRA's bots are a bit more clever, and also don't need to magically see into fog-of-war.
I'd love to see LLM based versions procedural tasks like "radiant quests" which are generally disappointing, though I've heard it discussed before and the real challenge is keeping it from going way off the rails.
The other challenge I think you'll run into in general is that there's a huge knee jerk reaction against any use of LLMs or other popular types of gen AI in games in places like Reddit or Bluesky.
Definitely the case. That being said, I think it would be hard, at least in the immediate future, to translate the concept of difficulty to a universal LLM for a bespoke/specific game. I assume most game AIs are tuned by hand to feel fair for a given difficulty level...but if you just give an LLM some new game, explain the rules and what resources/abilities it has available to it, you're stuck with adding some addendum to the tune of "and you're meant to represent an entity of 'medium' difficulty." For very well established games, it might have a sense of how given actions might fall into a skill-level hierarchy, but not for anything new.
Fine tuned LLMs though with actual experience with the game, maybe?
I find this funny because Stewart Butterfield (and others) founded Slack and Flickr by pivoting from the games they were trying to build. This is the opposite, someone trying to build a product and then pivoting to a game. I think this is a better path, FWIW.
I’ve felt for some time that there’s a gap in the market for a genuine spiritual successor to The Sims, using LLMs to power the interactions between agents to create a more realistic and immersive simulation of life. This seems like a step in the right direction.
The loop is all about adapting, experimenting, and seeing which combinations work and which don't. Right now it's designed around mini-games which can have a different goal per game -- as a quick example I'm currently building agent tic-tac-toe but hidden trapdoors and power-ups
So strange. Like, it doesn't sound like there really is a proper loop that you have figured out. But this actually sounds exciting - like when you read about how certain game genres were emerging and game designers were almost accidentally discovering what is fun.
This is cool for sure. Is it only all about tiles? Lately I've been thinking it would be awesome to get an AI to play DXBall (bricks game) type game or perhaps lode runner etc. would that be doable here?
We've only just begun! Max and I started building this about 1.5 months ago and are planning to ship a torrent of updates for the foreseeable future! Eventually it will be much more open/explorable world
Kudos, this is a very novel take! What's the most surprising emergent behavior you've observed? Have you observed any "social dynamics" that you didn't explicitly program?
Thanks for the comment! They can get pretty mad at each other relatively easily, frowning and battle crying, which is always fun to watch. When we turned on voice models (in the pipeline!!:)) their voices did as well
seriously, this embodied interaction angle seems like a much more humane way to understand AI behavior than just staring at walls of text. even if it occasionally feels like you're running a very advanced digital terrarium
The reason text works is because it has higher bit rate then speech. This is way many believe that CLI tools are still considered supreme in terms of getting things done quick.
While fun this game-like interface is too casual and it certainly has lower bit rate which impacts communicate exchange between an AI and the human operator.
It will be a fine abstraction if the goal is to have high-level overview though.
Thanks for the comment! We're working towards using the game's own simulation data (from Unity) to feed back into your game's agents. We hope this will prove less noisy than speech / real-world instrument data, allowing the AI to learn more effectively with new data every time you play
I wonder if this would be good for vibe coding / natural language for enemy AI. IE, place an enemy down and tell it: “every 3 seconds fire an arrow at the player. If the player is within 7 tiles of you, stop firing arrows, path to the player, and attack it with a sword. When your health reaches 10% run away from the player”
This misses many details, like if health is 10% but there is nowhere to go this will produce bad result where mob is just hanging around and not fighting, or if there is no line of sight then it should not fire arrows, or if it's impossible to path it should keep firing (IF there is line of sight). In a real game when you write every condition for every possible scenario I know what I would choose between a wall of text (anything in which can be misunderstood by llm) vs clear state machine
Just a heads up: the signup form disclaimer ("by signing up to create an account, you are accepting our terms of service and privacy policy") appears to link to a ToS route (theinterface.com/terms), but clicking that immediately redirects back to the login page (/signin) on Firefox [141.0.3].
Same thing happened when I tried hitting the URL directly. Do I have to accept the ToS before I'm allowed to read it?
I think this would be a great learning tool too - imagine like a bridge simulator or robocodo (https://game.rodocodo.com/hour-of-code/) - which is a learn to code tool for elementary students - but for AI agents.
As a tribute to Sims, you should allow for the `rosebud` cheat code :)
Have you played around with Sims-like plug-in objects, which include the knowledge of how to make the characters use themselves?
The important thing is that you can plug in new objects without reprogramming the people.
Sims objects (including characters) have a list of "advertisements" of possible interactions (usually shown as items on the user control pie menu, but also including invisible and special orchestration actions).
Each enabled action of every live object broadcasts its character-specific scores to each of the characters, adjusted to each character's personality, current stats, location, relationships, history, optionally tweaked by code.
Then to keep the sims from acting like perfectly optimized robots, they have a "behavioral dithering" that choses randomly between the top scoring few advertisements.
Here's a video of "Will Wright - Maxis - Interfacing to Microworlds - 1996-4-26" where he shows an pre-release version called "Dollhouse" and explains the design:
Jamie Doornbos gave a talk at GDC shortly after we released The Sims 1 in 2000, "Those Darned Sims: What Makes Them Tick?" in which he explains everything:
In The Sims 4 it's all been rewritten in Python, and has more fancy features, but it still uses the essential model for objects and advertisements.
The Sims 1 used a visual programming language called "SimAntics" to script the objects and characters, including functions that are run to score advertisements for each character.
But with LLMs you can write scoring functions and behavioral control in natural language!
Why wouldn't they? You want to use a state-of-the-art AI somewhere, you don't want to pay new subscription for that one game you want to try out. You can set a limit / spending cap on the api keys and revoke them right after you tried it. I don't see a problem there.
On the one hand, I do kinda hear where you're coming from, but OTOH I'm sympathetic to OP's concern that gaming should be relaxing or fun, and getting into the business of credential management plus billing management is neither of those things
Which is a lot of words to offer: be careful tossing out Luddite accusations just because it happens to be AI adjacent, that's rarely the whole story
Keep it up! Looking forward to what you figure out.
The Sims was my first experience with getting lost in a game having a negative impact on my life. Had to do most of a two week 5th or 6th grade geography project in the span of two days after playing the Sims instead of working on it.
LLMs with some decent harnesses could build up unpredictable - but internally consistent - strategies per each new game you play.
This is close to a proof of concept for those improvements.
Otherwise you run into the risk of "TOTAL NUCLEAR FINANCIAL LEGAL DESTRUCTION" ;)
Another thought that follows is that any kind of generative behavior, not just LLM, runs this risk of an endless pointless blandness. I.e. like with any artform we want there to be a point.
If those games are to feature LLM AI it would have to stand on it's own, with someone like these guys having thought it through.
No amount of dialogue is going to save that.
The actual story dialogue is usually interesting enough already
They win by the sheer quantity and by giving you a lot of subsystems to play with.
So LLM generated quest text probably feels it belongs here. It wouldn't, for example, in something with the Witcher 3 story quality.
By the way there are LLM dialog mods for Skyrim and everyone thinks they’re a joke because they suck.
Typically there are some easy micro and macro tricks that make the AI do something very stupid. That's why kiting is so ubiquitous in games - the AI just keeps following you while you whittle it down. Doesn't really work against a real player if they're microing the units.
The AI on higher difficulty starts a few centuries more technologically advanced than you, and gets multipliers on the starting resources like cities.
It’s not particularly fun to compete against.
OpenRA's bots are a bit more clever, and also don't need to magically see into fog-of-war.
Skirmish was a blast- I'd turtle until I had the enormous battleships (cruisers?) that could fire onto land. Loads of fun when I was like 12.
The other challenge I think you'll run into in general is that there's a huge knee jerk reaction against any use of LLMs or other popular types of gen AI in games in places like Reddit or Bluesky.
Fine tuned LLMs though with actual experience with the game, maybe?
I.e. what's the goal, how do you know you're doing well (or not), what makes it fun etc?
You had basic needs to fulfill, career advancement, relationships, and family generations.
Each of those fulfills the game loop.
While fun this game-like interface is too casual and it certainly has lower bit rate which impacts communicate exchange between an AI and the human operator.
It will be a fine abstraction if the goal is to have high-level overview though.
Same thing happened when I tried hitting the URL directly. Do I have to accept the ToS before I'm allowed to read it?
The important thing is that you can plug in new objects without reprogramming the people.
Sims objects (including characters) have a list of "advertisements" of possible interactions (usually shown as items on the user control pie menu, but also including invisible and special orchestration actions).
Each enabled action of every live object broadcasts its character-specific scores to each of the characters, adjusted to each character's personality, current stats, location, relationships, history, optionally tweaked by code.
Then to keep the sims from acting like perfectly optimized robots, they have a "behavioral dithering" that choses randomly between the top scoring few advertisements.
Here's a video of "Will Wright - Maxis - Interfacing to Microworlds - 1996-4-26" where he shows an pre-release version called "Dollhouse" and explains the design:
https://www.youtube.com/watch?v=nsxoZXaYJSk
Jamie Doornbos gave a talk at GDC shortly after we released The Sims 1 in 2000, "Those Darned Sims: What Makes Them Tick?" in which he explains everything:
https://www.gdcvault.com/play/1013969/Those-Darned-Sims-What...
Transcript:
https://dn721906.ca.archive.org/0/items/gdc-2001-those-darne...
Yoann Bourse wrote this paper "Artificial Intelligence in The Sims series":
https://yo252yo.com/old/ens/sims-rapport.pdf
In The Sims 4 it's all been rewritten in Python, and has more fancy features, but it still uses the essential model for objects and advertisements.
The Sims 1 used a visual programming language called "SimAntics" to script the objects and characters, including functions that are run to score advertisements for each character.
But with LLMs you can write scoring functions and behavioral control in natural language!
I thought it was just another YouTube video with no audio.
Which is a lot of words to offer: be careful tossing out Luddite accusations just because it happens to be AI adjacent, that's rarely the whole story