In my experience as a game designer, the code that LLMs spit out is pretty shit. It won’t even compile half the time, and when it does, it won’t do what you want without significant changes.
I just want to dump some thoughts that might be interesting to someone:
I wasn’t really thinking of normal code, more like “story beats” - like that dog on the ledge. Output a language more like a “move here” / “activate this” language for characters and environments, simple stream of output token language that can’t “syntax error”.
A combination of e.g.dwarf fortress that generates story beats based on a simulation and outputs mostly text that is turned into visual elements. But that just as an example, basically any game / sim engine can be reduced to outputing text story beats. Then LLM turns the synthetic “A […] appears and wants to […]” statement into something of a story and then into graphics. Basically like a giant lever for storytelling. Something like that is going to be the new minecraft.
And it’s ALSO going to make streamers more redundant. Instead of some centralized dude talking in your ear you’d have a commentator a bit like in Portal or Bastion. Or you have a party of people / friends that comment or interact with the world.
Or e.g. a sim game based on the world of Pride and Prejudice where social interaction has important meaning and you manage your estate and dynasty and try to marry off your sons and daughters. All “filled out” with AI generated content based on relatively simple text output. Or more like a settler game.
Basically you could overcome the problem with “samieness” in procedurally generated games (e.g. Elite Dangerous) and fill it up with interesting things. But you would need some kind of basic narrative language that an LLM can generate without syntax errors and can be turned into logic facing the player. Like generating puzzles like in Myst.
Like your simulation spits out “you arrive at the temple of Ninkompub. The entrance is locked.” and the game engine generates visuals and puzzles and enemies or some weird guy demanding you answer 3 questions. You’d need to adapt and train special AI models and need training data but I can see it being possible in the near future.
Basically a game engine / game development system with a live “game master” to torture the player and try to keep him in a punishment simulation for as long as possible. And whenever the player wins or stops playing the AI is punished itself and mutated so it learns constantly and evolves under “theatrical selection”. So when AI finally becomes sentient and evolves into an AGI it already knows exactly what to do :D
It could also allow for easier multiplayer, because the story lines and game progress can be generated “on the fly” without both players having to be on the same linear / narrative path.
Now combine that with VR and something like an AI girlfriend in your party that is designed to manipulate the player and create emotional bonds like a tamagotchy and you have something really disruptive.
I don’t mean to be dismissive of your entire train of thought (I can’t follow a lot of it, probably because I’m not a dev and not familiar with a lot of the concepts you’re talking about) but all the things you’ve described that I can understand would require these tools to be a fuckload better, on an order we haven’t even begun to get close to yet, in order to not be super predictable.
It’s all wonderful in theory, but we’re not even close to what would be needed to even half-ass this stuff.
You should refine your thoughts more instead of dumping a stream of consciousness on people.
Essentially what this stream of consciousness boils down to is “Wouldn’t it be neat if AI generated all the content in the game you are playing on the fly?” Would it be neat? I guess so but I find that incredibly unappealing very similar to how AI art, stories and now video is unappealing. There’s no creativity involved. There’s no meaning to any of it. Sentient AI could probably have creativity but what people like you who get overly excited about this stuff don’t seem to understand is how fundamentally limited our AI actually is currently. LLMs are basically one of the most advanced AI things rn and yet all it does is predict text. It has no knowledge, no capacity for learning. It’s very advanced auto correct.
We’ve seen this kind of hype with Crypto with NFTs and with Metaverse bullshit. You should take a step back and understand what we currently have and how incredibly far away what has you excited actually is.
The correct usage of LLMs in coding imo is for a single use case at a time, building up to what you need from scratch. It requires skill both in talking to AI for it to give you what you want, knowing how to build up to it, reading the code it spits out so that you know when it goes south and the skill of actually knowing how to build the bigger picture software from little pieces but if you are an intermediate dev who is stuck on something it is a great help.
That or for rubber ducky debugging, it s also great in that
In my experience as a game designer, the code that LLMs spit out is pretty shit. It won’t even compile half the time, and when it does, it won’t do what you want without significant changes.
I just want to dump some thoughts that might be interesting to someone:
I wasn’t really thinking of normal code, more like “story beats” - like that dog on the ledge. Output a language more like a “move here” / “activate this” language for characters and environments, simple stream of output token language that can’t “syntax error”.
A combination of e.g.dwarf fortress that generates story beats based on a simulation and outputs mostly text that is turned into visual elements. But that just as an example, basically any game / sim engine can be reduced to outputing text story beats. Then LLM turns the synthetic “A […] appears and wants to […]” statement into something of a story and then into graphics. Basically like a giant lever for storytelling. Something like that is going to be the new minecraft.
And it’s ALSO going to make streamers more redundant. Instead of some centralized dude talking in your ear you’d have a commentator a bit like in Portal or Bastion. Or you have a party of people / friends that comment or interact with the world.
Or e.g. a sim game based on the world of Pride and Prejudice where social interaction has important meaning and you manage your estate and dynasty and try to marry off your sons and daughters. All “filled out” with AI generated content based on relatively simple text output. Or more like a settler game.
Basically you could overcome the problem with “samieness” in procedurally generated games (e.g. Elite Dangerous) and fill it up with interesting things. But you would need some kind of basic narrative language that an LLM can generate without syntax errors and can be turned into logic facing the player. Like generating puzzles like in Myst.
Like your simulation spits out “you arrive at the temple of Ninkompub. The entrance is locked.” and the game engine generates visuals and puzzles and enemies or some weird guy demanding you answer 3 questions. You’d need to adapt and train special AI models and need training data but I can see it being possible in the near future.
Basically a game engine / game development system with a live “game master” to torture the player and try to keep him in a punishment simulation for as long as possible. And whenever the player wins or stops playing the AI is punished itself and mutated so it learns constantly and evolves under “theatrical selection”. So when AI finally becomes sentient and evolves into an AGI it already knows exactly what to do :D
It could also allow for easier multiplayer, because the story lines and game progress can be generated “on the fly” without both players having to be on the same linear / narrative path.
Now combine that with VR and something like an AI girlfriend in your party that is designed to manipulate the player and create emotional bonds like a tamagotchy and you have something really disruptive.
@nucleative@lemmy.world, @HeavyDogFeet@lemmy.world, @Traister101@lemmy.today, @FatCrab@lemmy.one, @mods_are_assholes@lemmy.world, @NounsAndWords@lemmy.world, @sudoreboot@slrpnk.net
I don’t mean to be dismissive of your entire train of thought (I can’t follow a lot of it, probably because I’m not a dev and not familiar with a lot of the concepts you’re talking about) but all the things you’ve described that I can understand would require these tools to be a fuckload better, on an order we haven’t even begun to get close to yet, in order to not be super predictable.
It’s all wonderful in theory, but we’re not even close to what would be needed to even half-ass this stuff.
You should refine your thoughts more instead of dumping a stream of consciousness on people.
Essentially what this stream of consciousness boils down to is “Wouldn’t it be neat if AI generated all the content in the game you are playing on the fly?” Would it be neat? I guess so but I find that incredibly unappealing very similar to how AI art, stories and now video is unappealing. There’s no creativity involved. There’s no meaning to any of it. Sentient AI could probably have creativity but what people like you who get overly excited about this stuff don’t seem to understand is how fundamentally limited our AI actually is currently. LLMs are basically one of the most advanced AI things rn and yet all it does is predict text. It has no knowledge, no capacity for learning. It’s very advanced auto correct.
We’ve seen this kind of hype with Crypto with NFTs and with Metaverse bullshit. You should take a step back and understand what we currently have and how incredibly far away what has you excited actually is.
Chatgpt once insisted my JSON was actually YAML
Technically it is, but I agree that is imprecise and nobody would say so IRL. Unless they are being a pedantic nerd, like I am right now.
The correct usage of LLMs in coding imo is for a single use case at a time, building up to what you need from scratch. It requires skill both in talking to AI for it to give you what you want, knowing how to build up to it, reading the code it spits out so that you know when it goes south and the skill of actually knowing how to build the bigger picture software from little pieces but if you are an intermediate dev who is stuck on something it is a great help.
That or for rubber ducky debugging, it s also great in that
That sounds like more effort than just… writing the code.
It s situationally useful