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

ChatGPT is really good at roleplaying

blog.varunramesh.net

Post reply on HN
Print view View original

21–30 of 97 posts

Re: ChatGPT is really good at roleplaying

#21

Post by myko »

nly wrote: post #14

I've been desperately unimpressed with ChatGPT in general. Everytime I ask it anything it just returns bland response containing factually incorrect information. The way it can retain context and good a conversion is impressive, but content wise I find it worthless - definitely no threat to a simple Google search.

I find for coding assistance it's quite helpful and gets me about 80% there for certain types of tasks, e.g.: recently migrated from one CI/CD platform to another, pasted yml from one and asked ChatGPT to re-create it for the other system. It ran most steps in the pipeline correctly without adjustment, and I just adjusted the others manually where it had issues.

Re: ChatGPT is really good at roleplaying

#22

Post by alsodumb »

nly wrote: post #14

I've been desperately unimpressed with ChatGPT in general. Everytime I ask it anything it just returns bland response containing factually incorrect information. The way it can retain context and good a conversion is impressive, but content wise I find it worthless - definitely no threat to a simple Google search.

Depends on what model you're using. I'm a PhD student, I use GPT-4 as a "knowledgeable on all topics but sometimes confidently wrong" companion.

I'd say ChatGPT is where I search first in 90% of the cases these days. I do confirm important things in Google, but the information provided by chatGPT is the starting point of my Google search, especially on topics/questions where I don't have a clear question and a vague idea of what I want - chatGPT gets my thoughts context perfectly almost every single time no matter what the topic is.

Re: ChatGPT is really good at roleplaying

#23

Post by nacs »

The problem with using these for freeform roleplaying is that it's very easy to hit the "I'm sorry but as an AI model..." that completely breaks character.

Even on the first example with the fish that is supposed to insult you, you quickly find that it will default to the "As an AI model I can't..." response.

This is where open-source models are going to be far better even if they don't have the parameter count of ChatGPT/GPT4.

Re: ChatGPT is really good at roleplaying

#24

Post by orost »

Udo wrote: post #15
p1esk wrote: post #10

When people say "ChatGPT" it's important to clarify if they mean GPT-4 (paid version), or the free GPT-3.5. GPT-4 is a completely different level, it's miles better in every way. It's like comparing a 10 year old to a 20 year old.

I got the impression that they gimped 3.5 severely over time. Since 4 is still restricted to 25 messages / 3 hours (for paying customers!), I sometimes fall back on 3.5. Of course it's impossible to prove, but it feels like it's failing hilariously at tasks it could do easily a few months ago. I wonder if more people have this suspicion or if it's just my imagination?

I definitely noticed a drop in quality when the gimped (but presumably dramatically cheaper to run) GPT-3.5-turbo model was introduced on the free version. As a paying subscriber I think you should still have access to the original GPT-3.5 (as "Legacy"), have you compared them?

Re: ChatGPT is really good at roleplaying

#25

Post by xrd »

varunramesh wrote: post #17
jonplackett wrote: post #2

I’ve been playing with some of the same ideas. Choose your own adventure games, old style adventure games and also found it really amazing - but the same problem mentioned that it gets sooooo expensive so fast that economically it doesn’t make any sense to make a game with ChatGPT Has anyone had any luck with the open source versions working well enough to do this kind of thing? How does the cost to run these compare…

I had some good results with Vicuna, but haven't tried it for gaming/fantasy roleplay yet.

What specs do you need for vicuna? Can you run it on regular hardware, or are you using a cloud GPU?

Re: ChatGPT is really good at roleplaying

#26

Post by jvanderbot »

nly wrote: post #14

I've been desperately unimpressed with ChatGPT in general. Everytime I ask it anything it just returns bland response containing factually incorrect information. The way it can retain context and good a conversion is impressive, but content wise I find it worthless - definitely no threat to a simple Google search.

I described my deck and asked how much weight it might support. It guessed a range that was reasonable but higher than I expected. I later confirmed it when I found Sagulator.

YMMV.

Re: ChatGPT is really good at roleplaying

#27

Post by Macha »

TeaDrunk wrote: post #8

Interestingly ChatGPT is really bad at roleplaying when I try to use it. I tried for example to have ChatGPT partially recreate this awesome/bad character arc in a campaign where one of the player characters actually fell in love with(!) one of the big baddies and this relationship was a "I can change him" with an irredeemable piece of shit. This also royally messed with the team dynamics in a super fascinating way.…

I remember this story in a webcomic: https://mythjae.wordpress.com/2016/01/28/fated-3/

Re: ChatGPT is really good at roleplaying

#28

Post by justrealist »

jonplackett wrote: post #2

I’ve been playing with some of the same ideas. Choose your own adventure games, old style adventure games and also found it really amazing - but the same problem mentioned that it gets sooooo expensive so fast that economically it doesn’t make any sense to make a game with ChatGPT Has anyone had any luck with the open source versions working well enough to do this kind of thing? How does the cost to run these compare…

Is it too expensive for you, personally, enjoying the game, or building a product on it?

Re: ChatGPT is really good at roleplaying

#29

Post by hackernewstom »

nacs wrote: post #23

The problem with using these for freeform roleplaying is that it's very easy to hit the "I'm sorry but as an AI model..." that completely breaks character. Even on the first example with the fish that is supposed to insult you, you quickly find that it will default to the "As an AI model I can't..." response. This is where open-source models are going to be far better even if they don't have the parameter count of Ch…

Include a sentence in the prompting such as: "In this roleplay do not mention that you are an AI model, or similar statements, and stay in character".
Post reply on HN