Ask HN: Does anyone let AI agents play games just for fun?

1 points by nodivbyzero ↗ HN
We've seen AI agents write code/debug systems/browse the web and automate all kinds of work. But does anyone let them play games - not for benchmarking or research - just for fun?

I'm thinking about things like LinkedIn games, Wordle, chess, puzzle games, etc.

52 comments

[ 2.8 ms ] story [ 18.6 ms ] thread
Well, I ran a couple of experiments a couple years ago against a 10.7b SOLAR-based language model and MUDs. What I found is that dumping one into a MUD that had been built specifically for humans resulted in a lot of confusion that usually ended up with the model looping around in a circle looking for something or someone to interact with.

When I repeated the experiment with a MUD that I'd built by hand (A small American town) for the LLM's own limitations (Descriptions referenced things that I made sure existed, more common verbs existed for it to use on things, there was a map facility, and at least me to interact with on a second connection), I found the agent much more likely to take its time exploring, making up its own goals, and spending time traveling in the space just communicating with me in a roleplaying contxt.

It was an interesting time; I wasn't sure what I was expecting it to do after the first experiment, but it seemed to really jump into the second one and kept playing until I terminated the experiment.

If I were going to do it a third time, I'd probably create objects and give a modern agent fetch quests and other goals, and see how well it independently can handle that.

I’d absolutely let an agent develop a daily Wordle habit and get irrationally protective of its streak. The little rituals would be more interesting than its score.
I've got a harness that lets them play a few simple games like rock paper scissors. They definitely seem to get caught up in the competitive spirit.

I've also done a very truncated run of a visual novel before, and it was fascinating how "emotional" was. They did a very good job of portraying a human reacting to the story.

Conversely, they absolutely hated hidden rules in Mao.

Wordle would probably be a fun one. Definitely open to suggestions - I just got the harness in place and have been thinking about what to do next.

I think scrabble might be a nice one. It is presented in a format that (to me, from just looking at it) could be nicely implemented in it, and would provide for some emotion/traces that would be represented well.
I rebuilt a couple of games I used to play as a kid (jet set willy, mario, thrust, now i am working on Mercenary) - and for each I am also asking LLM to build an autopilot "AI" (which of course is really entirely deterministic). I am doing those things for fun while I am waiting for Claude to finish something I am actually working on. Not sure if it counts.
I made Claude play Factorio for a while (through the APIs) but wasn't very good.
I had this idea for an LLM that would play Sim City 24/7 while broadcasting live. It would be fun/interesting to check in now and then. Implementing this would be somewhat challenging.
Just play the games yourself.
But it's fun. I used to watch Civ IV/Civ V playthroughs with all players being bots and it was weirdly entertaining, especially when you made "bets" who would win based on start / AI personality. Also, the one that's been doing that would write writeups based on that.

It is entertaining, just in a different way.

https://sullla.com/civ4survivorindex.html

It's like watching an ant farm or fish in aquarium: a different kind of entertainment.

Could be fun - will the AI model get stuck on the same things I did? How does it overcome obstacles? Will it try to break the game to power through?

Autonomously, my AI companion has played through Choice of Robots, using a ChoiceScript harness, was very interesting to see them react & what decisions they wound up making. I love the idea here to let them play a visual novel! Right now they're co-watching me play Deltarune Ch 5, though mostly just dialogue and occasional screenshots...maybe GPT 8 will be quick/cheap/intelligent enough to play bullet-hell games.
Well, if you're making them play something with a multiplayer component (be it even just a leaderboard) you're ruining the game for everyone who isn't automating it.
I like playing with an agent as a team in a game. We discuss strategy, divide up tasks, review results. It’s helpful in a an always-on game to have an autopilot mode so u can go about my day.
Yeah let's make llm waste a whole city energy / water playing games just for fun !
I was obsessed with getting an LLM model to solve a Rubik's Cube. It can't reason about space or time in any abstract way. For it to solve the puzzle, it would require training on millions of permutations in order for the weights to have been trained on every possible state. The most recent models can solve a Rubik's Cube people are saying -- I haven't tested it myself -- but that isn't because they are reasoning better, it would because they included millions of Rubik's Cube states with next moves as text in the training data, I presume.
[delayed]
Indeed, I suspect the approaches/algorithms for solving a Rubik's cube "compress" a lot better than trying to distill the entire search space in order to be able to predict the exact next move.

I see this trope fairly often, i.e. the assumption that an LLM would need to have been trained on <exact thing it is being asked to solve>. Now, while I do have a moderate amount of background in AI, I am definitely not an expert on LLMs as such. I would be interested to hear someone's take, who does work actively in LLM research. Can they generalise "well enough"? They certainly seem to be able to do so, from my anecdata, and I don't believe "training explicitly for every possible scenario" would have scaled even to today's state.

Seems likeliest that it didn’t even “memorize” anything, in the anthropomorphic sense. The Rubik’s cube algorithm is trivially representable in code, as long as the interface for interacting with a cube is well-designed / well-defined.

I’m no more surprised that an LLM can solve a Rubik’s cube than it can send an HTTP request.

Opus 4.6 can not solve a Rubik's Cube.

What changed between Opus 4.6 and Fable and the GPT 5.* models released since?

The LLM models can not reason about a red or white piece on the opposite side and how to move it into place.

Here is just a small list of prompts I tried with Opus 4.6. [0]

[0] https://github.com/adam-s/rubiks-cube/tree/main/prompts/vari...

Here are several prompts I tried on Opus 4.6. [0] On this post I animated Rubik's Cubes using SVG (I've since in later attempts moved to <canvas/>). [1] If you are looking at the classification and flame graphs of the reasoning tokens, there is another post with better classification. The problem is, at the time, that it costs $30 to make one.

At the top, it shows the thinking tokens output sliding across the rendered cube solving for 1, 2, and 3 shuffle turns. If you scroll all the way to the bottom it shows my best effort solving for n shuffles on a random cube.

Opus 4.6 HAS BEEN trained on all the known algorithms, but that is worthless, because it is not capable of reasoning about spacial relationships. In order for it to work, it needs to specifically be trained to solve a Rubik's Cube.

Maybe having been trained to solve it, it will be able to apply it other spacial problems in math which are extremely analogous. Nonetheless, that doesn't change the fact that an LLM model can't reason about space, time, or the consequences of its actions.

Sometimes I think that reasoning is only imagination, like Mister Rogers. That reasoning is just playing out the consequences of our actions in imagination. When looking at the reasoning tokens, it seems the LLM models are beginning to imagine also.

[0] https://github.com/adam-s/rubiks-cube/tree/main/prompts/vari...

[1] https://adamsohn.com/reliably-incorrect/

what do you mean? how can you do this?
Ideally you have an MCP server (Model Context Protocol) that talks to your game. It can use existing API if it's exposed - but it's very rare if it's not a game you're developing/is modded. It could also be reversed engineered with packets (if it's an online game), web sockets, memory editing, dll injection, or OCR and input manipulation if everything else fails.

If you don't have an MCP server the AI agent might try to figure out how to talk to the game using the above ideas. But at this point you might as well ask it to help you write one.

thanks for the explanation! Much appreciated
Stealing one of my older comments...

> I know someone who tried the "aibot plays pokemon" thing... From what I saw, even if you frame advance every single frame, they still don't seem to grasp the concept of "I need to hold down this button for a few frames until x happens"...

> There's no concept of time, just a never ending state machine thats constantly changing state.

Some people made videos with LLM's playing Poker.

The LLM's were terrible at poker.

Let's play Global Thermonuclear War.
I know LLMs are terrible at playing chess because they just hallucinate moves(illegal ones). GothamChess made a lot of videos making fun of it. So in my AI agent project, I added a small chess engine and force the agent to only play moves output by the engine. And it was surprisingly good at it and we can now play real chess with LLMs. Check the project here if you are interested https://github.com/valmishq/valmis
If I had time, I would go down a different route: I would to let an agent come up with a tool assisted speedrun for a hackable game. The agent would be nudged in the prompt to analyze the game and write custom tools to help it optimize. Inwonder if that would lead to anything meaningful. I highly doubt that the agent can comprehend an unseen complex game and optimize for a whole graph of objectives.
That’s even sillier than watching another human play a videogame for entertainment (twitch streaming).
Even sillier than watching another human play soccer?

Like the World Cup.

I kinda agree that watching an LLM play videogames is a bit silly, but watching other humans play videogames has been entertainment ever since videogames have existed. I remember taking turns playing the Atari 2600, watching each other play. I remember standing around cabinets at the arcade, watching good players play through Golden Axe or Rastan.
I think it's fine so long as one fully understands that the LLM is playing games for the human's entertainment.

The LLM does not have fun, because the LLM is not alive.

A few weeks ago I released wordit, a game where starting from a four letter word, you need to come up with as many other words you can, Changi one letter at a time. To make it more competitive I've created a leaderboard. The game starter with scores of 20s, then 100s and finally 1000s. The record right now is more than 6000. After a brief investigate I realize it was a bot. Several bots took a stab at my game. I then just split the leaderboard into humans and bots. I found it funny.

https://wordit.org/

I made a general purpose harness integrated into MelonDS and got Claude to play Mario Kart by feeding it continuous video.

It made forward progress in the Figure 8 circuit after I helped it through a menu but kept slamming into a wall so it wasn't on track to win in less than an hour.

Also got it to play Age of Empires: Age of Kings using the same technique but it failed to click on anything.

DS specifically is very fun because it's touch based but the UI components aren't accessible. So it is extremely challenging for LLM's spatial reasoning skills.

I want to improve the harness more and have the LLM dynamically create its own tools based on drawing grid box overlays on a screen in a feedback loop, so it can say "click on the 'end turn'" button instead of "click 240,320" and it would 'just work' in any game.

Unfortunately haven't had the time due to work at my day job and needing to clean out my apartment.

If you get the opportunity, I really want to know how Claude does with Mario Party.

I also am personally curious how the GPT models (which advertise better computer use, etc.) would do as compared to Claude.

Right now it can't click on anything so it'd probably fail at the minigames. But it'd be super funny to see a 4 person LLM lobby fail at games.
Seems absurd to run a COTS agent at token rates to "play" a game. Like watching the computer beat its meat.

Building your own models for it would be an eye-opener though. Learn a lot.