180
AI Driving Game Players Are Burning Through So Many Tokens Its Developer Had to Take Out a Bank Loan to Keep It Running (www.ign.com)
submitted 1 week ago by to c/fuck_ai@lemmy.world
38 comments fedilink hide all child comments
 

cross-posted from: https://piefed.ca/c/games@lemmy.world/p/1004331/ai-driving-game-players-are-burning-through-so-many-tokens-its-developer-had-to-take-out-a

The developer behind Teach My Little Sister How to Drive, a driving instruction sim where you instruct a generative-AI-powered, dynamic NPC, has been spending $1,000 dollars a day to keep the demo working.

you are viewing a single comment's thread
view the rest of the comments
[–] -1 points 1 week ago (13 children)

NPCs in games may be one of the very few actual good uses for AI. But not using the cloud obviously.

  • permalink
  • fedilink
  • source
  • hideshow 13 child comments
    •
  • [–] 2 points 6 days ago*

    Theoretically.

    In practice, 95% of consumer GPUs are too busy and too small to run an LLM good enough for dialogue, or a decision model smart enough to be warranted over hardcoded logic.

    CPUs are too slow. NPUs aren’t universal and are, unfortunately, all but useless with modern software frameworks.

    I am very much for LLM NPCs; there are some fantastic mods for such things.

    But in shipping games? That can’t be jank?

    It is not time. And I would never buy a game with a hard cloud dependency for SP. We will have to get past the RAMpocalypse, at the very least.

  • permalink
  • fedilink
  • source
  • parent
  • [–] 17 points 1 week ago* (last edited 1 week ago) (8 children)

    NPCs in games may be one of the very few actual good uses for AI

    No. No, they're not.

    Hey, you. You’re finally awake. You were trying to cross the border, right? Walked right into that Imperial ambush, same as us, and that thief over there.

    Yes, exactly. You know, I really wanted to cross that border to reach my family. My mother is very sick and needs my help. I brought some things along to help her, too... Some Trama Root and a Thistle Branch. The only thing left I need for her is a python script that reads a six column csv file that contains the start and end dates of our business quarters, the locations, the spending and income of those branches and their name and builds a nice graph to visualise the development each of those branches went through over the last ten years.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 8 child comments
    •
  • [–] 6 points 1 week ago (6 children)

    I thought we all learned this with the Batman Arkham games. "Real" AI would have made those games unplayable Souls-like nightmares as the enemies learn every single trick or something. We only got the dark knight power fantasy because of the very basic state machines they used. Like I teach them being praised at the time.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 6 child comments
    •
  • [–] -3 points 1 week ago (5 children)

    I think that it would be highly dependant on how the AI model is trained and what genre of game it is implemented into. For example, games that more heavily rely on dialogue or social interaction.

    Take Danganronpa, which is a series if linear narrative games about students trapped in a school forced to kill each other (and get way with it) to escape. You as the player have to investigate each murder and arrive at who the culprit was to win the game. They are fun on your first playthrough, but on any subsequent playthrough you already know the solutions. Like playing Clue but the game only ever has one outcome: Mr. Green, in the Library, with the Lead Pipe. You wouldn't even have to open the game box and you already know. Absolutely zero replayability.

    With an AI in such a game, you could have a new gamemode, like a New Game+, which players can play after the hand-crafted story. This is where I think something like a very carefully trained LLM could be useful, because it could create these kinds of situations at random, making them unpredictable and therefore adding replay value. Now, I don't think AI is currently sufficiently advanced for this use, but if it continues to develop I could see that niche making good use of it.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 5 child comments
    •
  • [–] 10 points 1 week ago

    Gonna have to agree to disagree. Part of the joy of any game is the careful crafting of the people who made them. Generating lines on the fly is going to lose that. Period. Even if they're more advanced. Unless they're actually sentient and understand what they're creating I guarantee it won't be the same.

  • permalink
  • fedilink
  • source
  • parent
  • [–] 8 points 1 week ago (3 children)

    Nah

    First of all, video games, including detective adventure games with randomised riddles to solve and different endings, have been done for quite some time, long before llms. See e.g. https://en.wikipedia.org/wiki/Blade_Runner_(1997_video_game) . You don't need a language model for that. Just be passionate about the project.

    Second, the whole idea of a video game like that is to tell a story. The last thing I would want is randomised bullshit to click through, wondering all the time if what I'm reading is even still within the boundaries of what the developers wanted or if the llm was already going off the rails. And that's not even considering the shale feeling of consuming not some created, curated story but merely the condensed diarrhea of stolen work.

    No thanks.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] 0 points 1 week ago (1 child)

    Well as I said, this would be for an additional game mode for replayability, not the primary story mode.

    Randomized content for a game like Danganronpa would be significantly more complex than Blade Runner. If a human/humans were programming it, each NPC (of which there are at least 18) would need their own randomizable schedule of daily activities, each would need their own dialogue lines which would easily number into the tens of thousands at least, and each would need the capability to commit a crime which is not witnessed by the player. The necessary variety of crime events would need to be very large to avoid the standard problem of procedural generation always feeling samey/generic, written by a human or not. Without large variety, it would feel like dungeons in Daggerfall.

    I mean yeah, a human probably could make that, but by the time they actually finished the content, the hardware their game is designed to run on would be considered "historically significant." It would easily take so long to develop that that it would not be worth the time investment. Unless they made serious content cuts, which would result in a pretty repetitive feeling final product. I dont see a reason that a specifically trained AI could fill in for the immense level of variety needed for that.

    Not that it really matters anyway, because I don't think anything exists that could run that performantly on a local device. Also the amount of time and data required to train something like that is probably not worth the time or cost investment either. Like I said before, I don't think current AI is sufficiently advanced enough for implementation. It would need further development, and if it did get further developed I could see a potential for it. But not right now, as I previously stated. And honestly it would probably need so long to develop that businesses will probably either have given up on AI (hopefully) or I won't be alive anymore to see it happen.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 2 points 1 week ago

    You're making some crazy statements and assumptions about a lot of things here, none of which are generally true. We saw a lot of games with vast amounts of content created with big teams. Saying it's impossible without ai in a reasonable timeframe doesn't magically become true by claiming it. And that's not even addressing the point that shitty randomised llm content will probably be just as shallow and repetitive as your negative examples.

  • permalink
  • fedilink
  • source
  • parent
  • [–] -2 points 1 week ago

    A locally run model with quest components and npc dialog as its training data could be tuned in such a way to make radiant quests more unique and engaging. The scenario you’re putting forward assumes a frame around a general purpose LLM, which would be a very bad idea (as the OP article shows.)

  • permalink
  • fedilink
  • source
  • parent