Live data from Hacker News

My kids and I just played D&D with ChatGPT4 as the DM

obie.medium.com

231–240 of 256 posts

Re: My kids and I just played D&D with ChatGPT4 as the DM

#231
post #82

Out of curiosity, I asked GPT3 to write a Friends episode one time. On the one hand, I was kind of surprised how well it knew the show - it knew the characters and how they were related, it knew about the coffee shop and Rachel & Monica's apartment. It wasn't very _creative_ though, it just kind of seemed to talk in circles, and nothing ever actually happened (it also gave Chandler & Joey iphones). This actually seem…

Are you aware of Nothing, Forever ? https://www.theguardian.com/tv-and-radio/2023/feb/04/ai-sein... >

Nope, but you definitely filled up at least the next 30 minutes of my schedule!

Re: My kids and I just played D&D with ChatGPT4 as the DM

#232
post #56

Earlier quoted context omitted.

At some point I asked it for a transcript from the trial of the Hamburgular and that triggered the guardrails, It lectured me about misinformation and how the Hamburgular was just a character from a series of ads and that no such criminal proceedings had ever transpired. The nagging from OpenAI is getting out of hand. Maybe instead of "hello" it should just open the conversation with telling us how the thing we are a…

On the other hand, AI Dungeon seems to have been trained heavily on erotic fanfiction. I remember telling it something like "I request a potion from the goblin king" and it respond with something like "As the goblin king turns around to grab the potion, you check out his tight little ass." So I can see why other AIs might put on guardrails. Entertaining either way lol

at least it was funny

Re: My kids and I just played D&D with ChatGPT4 as the DM

#233

Ha, I had the same idea a few hours ago. One thing I noticed is that the DM will let you do basically anything without fighting you on it. For example I was ambushed by some zombies and skeletons and I attempted to pitch them on a multilevel marketing scheme. Here is the result (my characters are Femribubs and Natnari): In a surprising turn of events, Femribubs and Natnari decide to change their approach. Instead of…

Oh great! What have you unleashed!? Now we're going to have to deal with hoards of AI zombies selling us multilevel marketing schemes!

Re: My kids and I just played D&D with ChatGPT4 as the DM

#235
post #25

I’ve not been hugely impressed by ChatGPT as a DM itself, compared to the humans I’ve played with. But as a tool to generate quests, encounters, puzzles, dialogue, it’s been incredible for me. It takes all the stress out of planning a campaign and it’s actually surprisingly creative.

I'd like to hear an example of an actually good and novel puzzle that it's generated, everyone says that it's able to generate them but in my experience they're the equivalent of integrating a number slide puzzle in a video game, exceptionally mundane. This seems like a great tool to act as a copilot, but having read the medium article, that was an almost arid level of dry adventure with nothing standing out.

https://www.youtube.com/watch?v=MTqruZje3gA

In the video, Twitch streamer Veritas uses ChatGPT to alter some quests in the video game Escape from Tarkov to make them more fun. Results are really interesting.

Re: My kids and I just played D&D with ChatGPT4 as the DM

#236

Earlier quoted context omitted.

Another odd bit was, when they said they wanted to steal the lizard thing, instead of role-playing the scenario, it just gave them a list of the rolls they’d have to make. I haven’t played a ton of D&D, but at the tables I’ve been at, you’d typically discover the challenges as you go. It sort of read like the artificially constructed examples of play you’d get in the DM’s handbook though (since these are written with…

If I'm going to ask the player to do something multi stage I will usually flag it so they know that there's a few places where things might go wrong. "OK well sneaking up to it is going to be the first hurdle; after that you're going to need to figure out how to get the giant lizard to cooperate as you smuggle it out the back". That lets the party start thinking ahead a bit, so maybe the illusionist says "how about I…

I basically agree I think, or at least agree to within normal-group-variance. Which is to say, at least with my friends, is someone said “I want to roll stealth to steal the lizard,” the DM would probably say something along the lines of “it’ll be more complicated than a single roll, you’ll have to get in there, get the thing free, there might be unforeseen problems (it is a barely tamed lizard monster after all).” I think we would probably not need to be warned in the first place and would go right into planning the heist anyway, because who doesn’t love a heist?

I think the specific thing I didn’t like in this example was the presentation of a specific path to the player. Although taking a second look, it seems the autoDM just presented it as a potential path, so maybe it isn’t so bad.

Re: My kids and I just played D&D with ChatGPT4 as the DM

#237

Ha, I had the same idea a few hours ago. One thing I noticed is that the DM will let you do basically anything without fighting you on it. For example I was ambushed by some zombies and skeletons and I attempted to pitch them on a multilevel marketing scheme. Here is the result (my characters are Femribubs and Natnari): In a surprising turn of events, Femribubs and Natnari decide to change their approach. Instead of…

Oh great! What have you unleashed!? Now we're going to have to deal with hoards of AI zombies selling us multilevel marketing schemes!

[dead]

Re: My kids and I just played D&D with ChatGPT4 as the DM

#239

Earlier quoted context omitted.

I think this is a classic baby DM script tbh. Rolling to find mounts at a vendor is dumb. You should only roll for significant challenges with meaningful differences where failure is meaningfully likely and interesting in its own right. If failing makes a situation less interesting it shouldn’t be an option. Shopping is actually the bigger problem. It’s like asking players to watch other players navigate a menu one b…

When you see a bear dancing, you don't nitpick its choreographic decisions. Because the bear. is. dancing. Do you see? It's a bear. dancing. DM style is a matter of taste. From another perspective, this feels like quite tidy DMing in terms of letting the players explore the systems in an early non-narrative context, just to understand the mechanics. Like a tutorial. The exotic mounts were a secret you could find if y…

I’m not debating whether it’s impressive. Sure, it’s impressive. But it’s not good.

Re: My kids and I just played D&D with ChatGPT4 as the DM

#240
post #137

Earlier quoted context omitted.

FWIW, the last few paragraphs of the response is something that I found was extremely common output for "DM AI". Whatever the situation, if it's kind of open ended, it'll say you did the thing really well, and you came to learn a bunch of new stuff, discovered who your friends really were along the way and everything was sunshine and rainbows. You can get it to put out something a little more grimdark but you need to…

Yes, ChatGPT seems to strongly prefer happy endings. Ask it to write a scene where a Montague bumps into a Capulet, and accusations start flying. One of the characters will successfully de-escalate. Or the price of Verona will show up with guards and stop the brawl. He'll even throw in some foreshadowing about how this will all end in tragedy. I suspect that this relentless love of upbeat endings is a side effect of…

I was using koboldAI to work on a short story and it had two characters bump into each other, one of them pull a knife, viciously shank the other seven times leaving the other one to collapse into a pool of blood and die, and then apologize for the accident like it was a minor occurrence after which they both went on their own ways with a "slightly embarrassed smile and nod, wistfully wondering of what might have been".
Post reply on HN