> The rocket launch was only achieved through significant ‘manual’ effort on Sol’s part as it papered over its late-game failures to connect certain production systems by repeatedly ferrying supplies between areas of the factory.
Hah I dunno, I think most human players wouldn't find that very fun and would instead take that as an opportunity to refactor the factory for optimized performance (i.e. the core gameplay loop)
When you just need a few parts to be made, it is sometimes very tempting to drop a chest, fill it up with parts, then go do something else, one of the 1000 tasks you have been neglecting.
When you start a new game in Factorio, if you never played it, you can’t know how much of each resource you’ll have to extract and neither how much of an item you’ll need to produce.
I know it’s easy to forgot after thousands hours of playtime, but the game starts with your character having to manually mine coal one by one.
Not true, every item shows its raw materials right there in its menu. The rest is just arithmetic.
Also the LLM probably seem dozens of factorio-related websites and setups.
And also, player (artificial or natural intelligence) don't need to setup everything upfront, it's actually the main gameplay feature is to run between production areas to expand/optimize them.
If you haven't been keeping an eye on LLMs playing games, GPT 6.1 Sol can new beat games 100% vision-only (without harnesses that give the agent extra information about the game). Pokemon is no longer a difficult challenge for it, even with additional difficulty restrictions such as hard mode and level caps. Recently, GPT 6.1 Sol beat Final Fantasy VII and its bonus bosses vision-only surprisingly quickly.
I suspect Opus 5.5 can do even better but there aren't as many people testing it.
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.
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 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
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.
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)
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.
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.
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.
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 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"
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 discrete dialog tree with stats and dice rolls attached.
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.
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.
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.
One limitation of these type of demos (and similar with Minecraft, etc) is that:
1. Putting into "peaceful" mode makes response time a non-factor
2. These models have read and have access to every guide and video ever published on the subject. Most humans, especially on first play, learn game systems and interactions from scratch, then develop novel strategies and plans and "hypothesis test" their way into beating the game.
This is still impressive compared to where we were a few years ago, and being able to synthesize all existing information about a game and then play it to completion in a semi-realistic setting is great.
At a quick glance someone might mistake this for "brilliant AI solves Factorio", when in reality, given a similar but different game where it wasn't gifted all the information or training data (e.g. massive videos and screenshots) it might fare terribly (as others have pointed out in comments).
I don't have the budget to do so, but if I do, it would be interesting to put different models in the same game with the same prompt to observe the behaviors that different models have when told to problem solve.
The biggest thing the bugs introduce from an LLMs perspective is a necessity for realtime response. Without them, the LLM can take as long as it wants to decide and do its next operation on the factory. With them, not responding fast enough results in death (particularly early game where a lot of biter combat is player driven). I presume it doesn't do great at this.
That was exactly my thought when reading the headline, before even looking into the article. Sure it's still somewhat complex to get to a rocket launch, but without the biters where is the challenge, other than complexity? It's like a big excel sheet with conveyor belts, pipes and electricity poles at that point only.
The more interesting challenge is the one you eliminated, complexity, in getting efficient logistics and arrangements. Dyson Sphere Program (until late in its existence) and Satisfactory don't have waves of enemies attacking your bases but remain challenging and a satisfying puzzle. I remember I used a constraint solver to help me optimize some late game recipes for Satisfactory.
Patience, for one. I reach a point where I'm looking at a mess of coal-fed sorters that need a complete rip-and-replace with electric ones, and that's usually the point where I close the game and go back to something easier like Kenshi.
I know it's a pipe dream, but these kinds of games (Factorio, Satisfactory, etc) would all benefit from sort of brain-game interface where I could just think about what I wanted to do.
The tedium and repetition is really what kills the enjoyment for me. The problem solving, planning, forethought, etc are super fun but it's killed by the former.
Factorio rewards the players progression through the game by taking away points of pain. The helper bots being a key transition. Without that pain there would be no reward. The fact that refactoring is expensive in time and complexity is what makes it realistic - it makes the earlier decisions important and consequential.
You can always enter creative mode and make whatever changes you want in the game instantly, but in my experience that does make it less fun.
It's basically programming with really bad ergonomics. I'm a programmer by profession and I was immediately frustrated by Factorio because I realized I'd rather just write code and build a project because it would scratch the exact same itch while actually being useful.
I tried really hard to like Factorio, and I bounce off of it each time after playing a save for 4-8 hours for exactly this reason. I know a lot of programmers who love it, so I almost wish I could get into it.
Now I'm tempted to try again, because it feels like the kind of game I should really enjoy.
The biters are just a line item in the spreadsheet. Early players might build a wall with an ammo belt, for example. There's lots of options that don't involve responding in real-time.
Though these days I just manually clear the pollution cloud with a pocket full of turrets. The forward deployed engineer model, if you will.
It doesn't seem to understand the core mechanic of Factorio, which is putting outputs onto belts and routing them to other inputs. It does this in a few places, but for the most part everything is put into chests and moved around by the player.
Handfeeding only has higher throughput for short trips/a single item type, definitely not for an entire factory. A high throughput factory looks like this:
If you don't have enemies, you have unlimited time. Doing it thru player is what it learned and wasn't pressured to do better (it has no fatigue as opposed to humans).
It's just easier to do manual stuff when the complexity of scaling the automation becomes cognitively demanding. A new player does mostly everything manually while experienced players have a natural intuition for planning and scaling the automation.
It is weird and almost alien how AI designs factories. When Claude announced Fable 5 they also did a Factorio demo that was the same kind of designs and a lot of hand feeding and direct insertion.
In some sense it proves to me that llms are not just copying and are actually creating a bespoke solution. But I honestly expected more from frontier models
To them the player character is just as good as a belt for transporting materials.
The player character actually does have an extremely high capacity so if you are constantly running between chests your factory will still be reasonably high throughput.
> In some sense it proves to me that llms are not just copying and are actually creating a bespoke solution.
Is the weirdness of the designs meaningfully beneficial in any way that the AI has figured out, or could it for example be that the training data the LLMs learned from are too different from the type of data it would need to learn from in order to interface properly and play the game better? And then because of that incompleteness, the LLM has to hallucinate a lot and it results in weird outcomes.
I think one of the problems is that AI has no sense of laziness or time. Unless it's in their training data or prompted they're not going to bother automating something
This is a much better benchmark as a concept (without knowing the details of how the problem is framed to the agent) than some of these ridiculous things the frontier labs are reporting (vending machine simulator? come on).
Can you report more in depth at stumbling blocks? It's a surprisingly linear curve.
I'm going to try it on a local run.
I wonder if there is room for "recommendorio" where it gives you AI advice on your current factory.
Is the model pre-trained with any knowledge of the game?
Robots: they're just like us!
Hah I dunno, I think most human players wouldn't find that very fun and would instead take that as an opportunity to refactor the factory for optimized performance (i.e. the core gameplay loop)
When you just need a few parts to be made, it is sometimes very tempting to drop a chest, fill it up with parts, then go do something else, one of the 1000 tasks you have been neglecting.
youre not really sure how to set up some of the logistics/how much of what you will need where ahead of time
When you start a new game in Factorio, if you never played it, you can’t know how much of each resource you’ll have to extract and neither how much of an item you’ll need to produce.
I know it’s easy to forgot after thousands hours of playtime, but the game starts with your character having to manually mine coal one by one.
Also the LLM probably seem dozens of factorio-related websites and setups.
And also, player (artificial or natural intelligence) don't need to setup everything upfront, it's actually the main gameplay feature is to run between production areas to expand/optimize them.
I suspect Opus 5.5 can do even better but there aren't as many people testing it.
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.
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.
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 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.
thats close to tweaking existing dialog, but with a bit more freedom.
you still dont have to give the player arbitrary response
(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)
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"
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.
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
both American and Chinese labs aim their models at coding and knowledge work, so they penalize adult themes in post-training.
A little unpolished because it was bolted on by fans and not the original game developers, but otherwise very impressive.
This is still impressive compared to where we were a few years ago, and being able to synthesize all existing information about a game and then play it to completion in a semi-realistic setting is great.
At a quick glance someone might mistake this for "brilliant AI solves Factorio", when in reality, given a similar but different game where it wasn't gifted all the information or training data (e.g. massive videos and screenshots) it might fare terribly (as others have pointed out in comments).
Patience, for one. I reach a point where I'm looking at a mess of coal-fed sorters that need a complete rip-and-replace with electric ones, and that's usually the point where I close the game and go back to something easier like Kenshi.
The tedium and repetition is really what kills the enjoyment for me. The problem solving, planning, forethought, etc are super fun but it's killed by the former.
You can always enter creative mode and make whatever changes you want in the game instantly, but in my experience that does make it less fun.
Now I'm tempted to try again, because it feels like the kind of game I should really enjoy.
Though these days I just manually clear the pollution cloud with a pocket full of turrets. The forward deployed engineer model, if you will.
Part of the reason players try to automate manual feeding is the annoyance factor, which AI doesn't care about.
https://youtu.be/RI4ocNajf_w?si=O1aRi22riKwiuh_p&t=7493
Stack size varies, but let's take 75 as an average, which means a player can carry 5600 items.
But carrying isn't throughput. If a round-trip takes 30s, player throughput will be 186 items/s. That's a measly 12 T1 belts.
The player character actually does have an extremely high capacity so if you are constantly running between chests your factory will still be reasonably high throughput.
Is the weirdness of the designs meaningfully beneficial in any way that the AI has figured out, or could it for example be that the training data the LLMs learned from are too different from the type of data it would need to learn from in order to interface properly and play the game better? And then because of that incompleteness, the LLM has to hallucinate a lot and it results in weird outcomes.
Can you report more in depth at stumbling blocks? It's a surprisingly linear curve.
I'm going to try it on a local run.
I wonder if there is room for "recommendorio" where it gives you AI advice on your current factory.
Is the model pre-trained with any knowledge of the game?
Boring though without something to attack you imo