Hacker News Reader

Comment by mickeyp | Hacker News Reader

Watching AI play games is fun, but I can't wait for someone to crack having small LLMs in games to enrich them in useful and interesting ways.

View on HN

Not sure about LLMs but I yearn for someone to train neural nets to play the NPCs in a shooter game well but without having aimbots or seeing through walls - basically a “true” version of F.E.A.R. AI where they play simulated squads that overwhelm the player with various tactics and adapt to their playstyle. Or even just make increasingly complex state machines and goal oriented systems which might be more performant but still encode the main rules.

You want to create a human feeling using AI, but AI is not needed for the feelings you want to create. All you really want is a feeling of variety and surprise in the way computer opponents behave. A LLM-level of AI might just do the same thing all the time and become quite boring.

And yet, that argument I just made often bothers me, because I do think there is a place for strong AI in games, and many people just reflexively say "strong AI wouldn't be fun". I think they are partly right, but I also think a strong AI that is purposely dumbed down in fun ways could be really enjoyable.

When it comes to shooters though, those games aren't really about competition and decisions, they are about giving the human player an experience. I do think strong AI is overkill in such situations.

A strong AI for chess absolutely improves chess computer games.

A strong AI for a Magic: The Gathering style card battler would improve the game.

A strong AI for Civilization would improve the game, but I think we're starting to get into a realm where you can fake some things. I do think Civilization games are lacking because of poor AI, but I don't know that we need LLM-levels of AI in Civilization.

Again, in shooters, there's no decision, no illusion of fair play or equal competition based on making decisions, it's all about the human experience, and so LLM-level AIs aren't needed here at all.

Perseverance might be the drive missing from those bots.

I expect the area of fun between painfully easy to beat and infuriatingnly overpowered with superhuman skill level to be quite narrow.

I have been playing with this. Having tiny models provide more contextual information when an NPC speaks is cool but the "small enough" models tend to go off the rails.

It's fun if you have a handful of people playing, but it would quickly go towards negative reviews if you tried to sell it.

I don't think an LLM should ever be having a two-way conversation with the player. It would be more suited to 2 npcs having conversations amongst themselves. All dialogue delivered to the player should be human written.

At most it should be mildly tweaking existing dialogue, e.g updating a line to reflect game events. But as soon as you present a prompt window to the player the entire illusion that you are in a game world instantly dies.

> I don't think an LLM should ever be having a two-way conversation with the player. It would be more suited to 2 npcs having conversations amongst themselves. All dialogue delivered to the player should be human written.

I don't agree. All -important- dialogue yes. Background NPC chatter? Even when studios spend a lot of effort crafting multiple lines, pretty soon you have seen them all. They also don't change based on player actions.

i think it can work for procedurally generated situations, and the llm is acting more like a translator between data model and player rather than being a data storage layer.

thats close to tweaking existing dialog, but with a bit more freedom.

you still dont have to give the player arbitrary response

You absolutely cannot let the player get any whiff of AI slop so the integration needs to be extremely delicate. Watching a film and noticing a person has 6 fingers in a scene would instantly kill any engagement someone would have with it. For it to remain as art it must always be fundamentally about human-human communication, I want to have a conversation with the game developers through the game, not with a LLM. If developer intent ever becomes muddled the game has failed.

Most dialogue in games is crap. I would much rather have an LLM writing the dialogue so that it can include references to player choices over what currently exists.

Nothing great was ever built by someone who was sweating the reviews. Fuck 'em, do what feels right to you.

I built a factorio "sidecar" LLM harness that would look at production stats every few minutes and identify issues. Even sonnet was good enough at finding potential things to alert on. coal production dropped and then iron production dropped two minutes later --> top potential issue.

Not useful enough that I actually play with it, but it was capable enough to be useful even just with sonnet.

It's a funny situation at the moment because most attempts are just sort of shoehorning LLM chat into a game that doesn't really need it, and then gamers quite rightly mock the results which don't really add anything to the game.

I think there's a bit of a lack of knowledge of game history on both sides, because we already know it can be fun when a game is designed around it. It worked in Façade (2005!) and it worked in Starship Titanic (1998!!), and both of those could have worked even better with the technology we have today.

One type of game that seems like an obvious one to me, is something a bit like The Sims except that the characters' decisions and speech are LLM-controlled. Maybe one of the characters is player-controlled, so you can interact and talk with everyone else. Or to describe it from the other direction, it'd be a kind of interactive fiction like AI Dungeon invented, but grounded in a physically simulated setting.

I sometimes wonder if Façade actually used neural networks for text generation or just simple markov chains... I remember pewdiepie saying melon and getting kicked out lol. Also the Sims would be awesome but imagine an RPG with rich worldbulding, and LLM's interacting with the lore, like in an Elder Scrolls or Fallout title. Wouldn't it be mind blowing?

There has to be a AAAA studio building the new, greatest MMO, that also utilizes LLMs as part of the game loop, surely?

(Not sure how prohibitive it would be to do anything interesting with local LLMs of any kind in a big game, given how many platforms/setups you have to make the stuff run smoothly on, and the sheer amount of RAM they would bind)

https://proxiai.com/ Will Wright (The Sims, etc) is dabbling with AI in his latest game. No idea if it'll be any good, but there's def companies out there trying.

AAAA studios don't take risks. Financing a top game is way too expensive to gamble on. Only indies and small studios working on cheaper games can risk trying unproven ideas.

not taking risks is the biggest risk in the times when the most anticipated game drop (gta 6) is probably going to be decompiled within days if not hours of release

In "Where Winds Meet" you can talk to NPCs

With the aid of SOTA LLMs, you can encode specialist intelligence into computer programs without going through the roundabout route of training small LLMs. And even when you do go the route of ML, it doesn't necessarily have to be an LLM or even a Neural network.

I hope to experiment with using classic tabular reinforcement learning in game AI one day. Imagine a framework that makes setting up small Q-learning tables easy, and you can just sprinkle dozens of Q-learning algorithms throughout your traditional AI scripting.

These old basic algorithms often have mathematical proofs that they will converge to optimal behavior, and they do so fairly quickly. Turning the complex world of the game into a Q-table would require some creativity and judgement, but is not that hard I think.

In the end we would have a controllable AI that actually does adapt. That's my hope anyway.

I cannot wait for NPCs having a casual conversation. Doesn't seem too hard, as current chatbots are already role-playing.

I understand https://neal.fun/infinite-craft/ uses an LLM under the hood each time a novel combination happens.

This... Hopefully inference can become so cheap games can have simple LLM's integrated into them. Imagine an RPG with LLM's and with rich worldbuilding; like Fallout NV with LLM's. Honestly, scary!

Inference is already cheap enough. It's basically free for the quantities people actually want to read.

LLMs only become expensive if you let them generate 1000 lines of code every 2 minutes.

Fable seems to be a potential franchise to do all this; imagine a "living fantasy world" where everything truly unfolds based on your decisions, giving NPCs all the functions/actions that a player has.

i think hooking in an llm to dwarf fortress for turning the very mechanistic descriptions into prose. same with the adventure mode talk. instead of a constant "well weve got beasts" an llm could read a bunch of the language tree to make a unique introduction for each.

same with the df artworks. get kitfox to design some art styles for dwarves, elves, goblins, and humans, and then instead of just reading

"this is a fine engraving of a dwarf. it concerns the events of 102 when ulrist fastbuttons was elected mayor. ulrist fastbuttons is sobbing"

SOTA LLMs would probably be fun opponents in turn based 4x.

Will be interesting to find out what they can bring over simply having better SOTA LLM coded opponents (I am sure, we will).

LLMs could be interesting geopolitical roleplayers if instructed to faithfully assume the leadership traits or cultural biases of the their intended role, especially in games with expressive diplomacy systems like Stellaris or Europa Universalis. It could also make the diplomacy experience a lot richer because the LLM is able to interpret ligusitic tone and other verbal cues. The modern diplomacy experience is basically just a dialog tree with stats and dice rolls attached to it.

Unfortunately for it to be sustainable it would need to be an LLM with VRAM usage probably under a couple gigs max so it can run locally on the gaming PC. That said it may well be better than current AI in strategy games.

Yes, the insane hardware requirements make it totally impractical - that's why I say "turn based" because turn based games would at least allow for a practical gameplay experience even with remote LLM APIs.

I bet a small 1B model trained solely on how to play the strategy game would do better than current AI solutions for these games and that would probably be able to run even on integrated GPUs.

I would love the revival of the RTS genre with better computer enemies.

Such AIs have existed in Brood War and Dota 2 for a very long time now (pre-LLM), problem is they aren't very fun for humans to play against.

bah, such a pity we didn't have those in wc3 times. bots sucked badly.

I could play online, but my dial-up costed $1 per hour for 250ms latency and near guarantee of at least one disconnect per hour.

with better being more convincing, rather than just stronger

its the same for like a chess bot. i dont want a perfect stockfish that occasionally makes a serious blunder when a dice rolls a certain way. i want one that actually falls into traps

You could play against an AI right now on the Brood War ladder right now. Granted, you have to make it to S rank first because that's where the bots end up.

SS13 but the ship's AI is actually an AI instead of a player in disguise

impossible, unfortunately, because LLMs suck at creative writing. without handholding, they are incapable of producing anything but cliche-ridden purple prose for young adults. you can feed it the first half of a book with lots of sex, violence, and profanity, ask it to continue the story, and it will degrade into safe slop within two chapters.

both American and Chinese labs aim their models at coding and knowledge work, so they penalize adult themes in post-training.

There are already AI-based mods for games like Grand Theft Auto V that make NPCs talk back and follow your verbal commands: https://www.youtube.com/watch?v=3_Eng-F4akI&t=106s

A little unpolished because it was bolted on by fans and not the original game developers, but otherwise very impressive.

Reply on news.ycombinator.com