Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

I replaced Animal Crossing's dialogue with a live LLM by hacking GameCube memory

joshfonseca.com

Post reply on HN
Print view View original

61–70 of 198 posts

Re: I replaced Animal Crossing's dialogue with a live LLM by hacking GameCube memory

#61

Post by anticrymactic »

heckelson wrote: post #56
beckthompson wrote: post #55

Earlier quoted context omitted.

I'm unsure how useful they'll really be honestly! In many situations its very useful to know when your "done" talking with an NPC when they start repeating lines etc... There could probably be cool uses but I don't think it will be a pure "upgrade" as the repeating dialog is kind of a feature honestly. We'll have to see how it pans out xD

I guess you could just introduce a Symbol that marks the NPC as "I said everything I have" or gray out their text, or some other visual marker

Or they could just tell you. Imagine talking to someone over and over again. They would tell u to get on with wherever you promised them.

Re: I replaced Animal Crossing's dialogue with a live LLM by hacking GameCube memory

#63

Post by heckelson »

anticrymactic wrote: post #61
heckelson wrote: post #56

Earlier quoted context omitted.

I guess you could just introduce a Symbol that marks the NPC as "I said everything I have" or gray out their text, or some other visual marker

Or they could just tell you. Imagine talking to someone over and over again. They would tell u to get on with wherever you promised them.

oh yeah that's much smarter than what I suggested

Re: I replaced Animal Crossing's dialogue with a live LLM by hacking GameCube memory

#64

Post by cubefox »

Arisaka1 wrote: post #57
malfist wrote: post #19

Earlier quoted context omitted.

I used to care about repeat scripted dialogue like you, until I took an arrow to the knee

Yep. I can see how relying on non-repetitive dialogue generated by LLM's will inevitably make us lose common ground scenarios that end up being memorable moments, like the arrow to the knee from Skyrim or "such devastation! this was not my intention" from Final Fantasy 14. A way to bypass this problem is to keep the important dialogue fixed, but this only one of the problems. Another issue would be emphasizing the me…

A better application of LLMs could be merely using them as a text parser in text adventures. NPCs could continue to have some hidden, hard coded information or abilities, here stored in the context window, and the LLM is used to provide that information to the the player he (or she) puts in roughly the right words. Rather than exactly the right words.

Re: I replaced Animal Crossing's dialogue with a live LLM by hacking GameCube memory

#65

Post by Terr_ »

heckelson wrote: post #56
beckthompson wrote: post #55

Earlier quoted context omitted.

I'm unsure how useful they'll really be honestly! In many situations its very useful to know when your "done" talking with an NPC when they start repeating lines etc... There could probably be cool uses but I don't think it will be a pure "upgrade" as the repeating dialog is kind of a feature honestly. We'll have to see how it pans out xD

I guess you could just introduce a Symbol that marks the NPC as "I said everything I have" or gray out their text, or some other visual marker

> a Symbol that marks the NPC

That seems a bit like deck-chairs on the Titanic. The hard part isn't icon design, the hard part is (A) ensuring a clear list exists of what the NPC is supposed to ensure the user knows and (B) determining whether those goals were received successfully.

For example, imagine a mystery/puzzle game where the NPC needs to inform the user of a clue for the next puzzle, but the LLM-layer botches it, either by generating dialogue that phrases it wrong, or by failing to fit it into the first response, so that the user must always do a few "extra" interactions anyway "just in case."

I suppose you could... Feed the output into another document of "Did this NPC answer correctly" and feed it to another LLM... but down that path lies [more] madness.

Re: I replaced Animal Crossing's dialogue with a live LLM by hacking GameCube memory

#66

Post by maplethorpe »

Arisaka1 wrote: post #57
malfist wrote: post #19

Earlier quoted context omitted.

I used to care about repeat scripted dialogue like you, until I took an arrow to the knee

Yep. I can see how relying on non-repetitive dialogue generated by LLM's will inevitably make us lose common ground scenarios that end up being memorable moments, like the arrow to the knee from Skyrim or "such devastation! this was not my intention" from Final Fantasy 14. A way to bypass this problem is to keep the important dialogue fixed, but this only one of the problems. Another issue would be emphasizing the me…

Instead of generating truly random dialogue, you could choose to only generate random seeds within a certain number range, so that the probability that two players seeing the same line of dialogue increases, while still having enough variety for it to feel "random" to a solo user.

Re: I replaced Animal Crossing's dialogue with a live LLM by hacking GameCube memory

#67

Post by ares623 »

Arisaka1 wrote: post #57
malfist wrote: post #19

Earlier quoted context omitted.

I used to care about repeat scripted dialogue like you, until I took an arrow to the knee

Yep. I can see how relying on non-repetitive dialogue generated by LLM's will inevitably make us lose common ground scenarios that end up being memorable moments, like the arrow to the knee from Skyrim or "such devastation! this was not my intention" from Final Fantasy 14. A way to bypass this problem is to keep the important dialogue fixed, but this only one of the problems. Another issue would be emphasizing the me…

I agree. An important aspect would be lost. It would be as interesting and relatable as sharing that dream you had last night with your friends. “Uh huh. Cool story bro.”

Re: I replaced Animal Crossing's dialogue with a live LLM by hacking GameCube memory

#70

Post by madaxe_again »

Terr_ wrote: post #65
heckelson wrote: post #56

Earlier quoted context omitted.

I guess you could just introduce a Symbol that marks the NPC as "I said everything I have" or gray out their text, or some other visual marker

> a Symbol that marks the NPC That seems a bit like deck-chairs on the Titanic. The hard part isn't icon design, the hard part is (A) ensuring a clear list exists of what the NPC is supposed to ensure the user knows and (B) determining whether those goals were received successfully. For example, imagine a mystery/puzzle game where the NPC needs to inform the user of a clue for the next puzzle, but the LLM-layer botch…

No, this is the Einstein/student model that has been proposed for improving LLM output quality.

Basically you have your big clever LLM generating the outputs, and then you have your small dumb LLM reading them and going “did I understand that? Did it make sense?” - basically emulating the user before the response actually gets to the user. If it’s good, on it goes to the user, if not, the student queries Einstein with feedback to have another crack.

https://openai.com/index/prover-verifier-games-improve-legib...

Post reply on HN