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

"A Young Lady's Illustrated Primer" Simulated by GPT-4

andrealyip.com

Post reply on HN
Print view View original

131–140 of 163 posts

Re: "A Young Lady's Illustrated Primer" Simulated by GPT-4

#131

Post by empiricus »

craigyk wrote: post #76
jjcm wrote: post #43

Earlier quoted context omitted.

Definitely a Stephenson issue. It's a meme for sure, but there is some truth in it. Nowadays when I re-read any of his old works, I tend to stop after getting about 80% through. That said, I do think that these incredibly complex narratives are extremely hard to wrap up neatly. GRRM has this issue as well. That's the problem with letting characters act how they would actually act, and make decisions that make sense f…

Gene Wolfe. Somehow resolves or explains every seemingly random occurrence throughout the story, even if the first person narrator is extremely unreliable and doesn’t make the connection

I remember some books from his "Latro in the mist" series, where the last pages have at least as much action as the rest of the book. Very strange pacing, but it is memorable and it works.

Re: "A Young Lady's Illustrated Primer" Simulated by GPT-4

#132

Post by Foreignborn »

LordDragonfang wrote: post #62

>With hesitation, Lin approached and said, "Grandmother, today I discovered something new about myself, something that makes me feel uncomfortable and out of place." >The elderly painter paused and looked deeply into Lin's eyes. "Ah, young Lin," she began softly, "Much like the Formosan bear, every individual goes through stages of growth and change. These moments might seem unfamiliar and daunting, but they're a tes…

What I found is that works, but is very expensive:

- Do chain prompting where each character gets a separate prompt for the dialogue. Here you'll give it all the character info needed to figure out what the character would do (emotions, background, the story context) and then have GPT4 take passes at what this character would think/do/say.

- Optionally, you can use some sort of reACT or chain of thought prompting to proof-read/critique the answer to minimize noise. I haven't fully explored this, but I think this agent needs to know the entire story outline/context.

- After you get this, you can use an instruction-tuned agent like text-davinci or text-davinci-edit to actually write the prose. It works really well with few-shot examples. It's much better at creative writing. I'm sad they consider it legacy because to me it's the _only_ way to get usable creative writing.

- Now run the rest of the story. You may need to keep some sort of state about the character so that it informs the next character interaction, but I haven't gotten that far in my research.

Re: "A Young Lady's Illustrated Primer" Simulated by GPT-4

#133

Post by lynx23 »

jjcm wrote: post #43
cl42 wrote: post #22

Earlier quoted context omitted.

Do you think that's a Neal Stephenson issue? Don't get me wrong, I love his writing, but I found this to be the case with Fall too. I can see it becoming tiring if one is writing 400+ page books.

Definitely a Stephenson issue. It's a meme for sure, but there is some truth in it. Nowadays when I re-read any of his old works, I tend to stop after getting about 80% through. That said, I do think that these incredibly complex narratives are extremely hard to wrap up neatly. GRRM has this issue as well. That's the problem with letting characters act how they would actually act, and make decisions that make sense f…

Seveneves is IMO the worst Stephenson story I read so far. No good sign if you stop reading midway. Although the first act was nice.

I recently found my love for Greg Egan again. Incandescence is currently way up on my list of hard scifi novels I absolutely loved. I'd love to read more of this kind...

Re: "A Young Lady's Illustrated Primer" Simulated by GPT-4

#135

Post by dopidopHN »

biofunsf wrote: post #45

To zoom in on one part of the post: > Emotion & Empathy > PRIMER: seems to show a level of empathy and emotional understanding. > ChatGPT: can recognize and respond to emotional cues in the text, but it doesn't "feel" emotions in the way humans do. Its responses are based on patterns in the data it was trained on. Isn't the fictional PRIMER also just responding to patterns in whatever data it was trained on? Sure it'…

IIRC

One of the book character is a human contracted out by the primer AI itself to act out a voice for him.

That character understand that something is up and that the AI client has a new owner that is a volatile situation. And act on it. Ultimately that other human shows empathy.

Re: "A Young Lady's Illustrated Primer" Simulated by GPT-4

#136

Post by cl42 »

Foreignborn wrote: post #132
LordDragonfang wrote: post #62

>With hesitation, Lin approached and said, "Grandmother, today I discovered something new about myself, something that makes me feel uncomfortable and out of place." >The elderly painter paused and looked deeply into Lin's eyes. "Ah, young Lin," she began softly, "Much like the Formosan bear, every individual goes through stages of growth and change. These moments might seem unfamiliar and daunting, but they're a tes…

What I found is that works, but is very expensive: - Do chain prompting where each character gets a separate prompt for the dialogue. Here you'll give it all the character info needed to figure out what the character would do (emotions, background, the story context) and then have GPT4 take passes at what this character would think/do/say. - Optionally, you can use some sort of reACT or chain of thought prompting to…

You raise a good point: while ChatGPT might have stylistic limitations that make it less-than-ideal for a "Primer"-type product, LLMs more broadly don't suffer from that stylistic limitation. Whether it's via prompting, fine-tuning, or something else, you can build multiple characters who address some of the original commenter's concerns.

Re: "A Young Lady's Illustrated Primer" Simulated by GPT-4

#137

Post by Sharlin »

jjcm wrote: post #43
cl42 wrote: post #22

Earlier quoted context omitted.

Do you think that's a Neal Stephenson issue? Don't get me wrong, I love his writing, but I found this to be the case with Fall too. I can see it becoming tiring if one is writing 400+ page books.

Definitely a Stephenson issue. It's a meme for sure, but there is some truth in it. Nowadays when I re-read any of his old works, I tend to stop after getting about 80% through. That said, I do think that these incredibly complex narratives are extremely hard to wrap up neatly. GRRM has this issue as well. That's the problem with letting characters act how they would actually act, and make decisions that make sense f…

Iain M. Banks. Many of his endings are tragic, or at least bittersweet, and have some sort of a twist in them, so slightly YMMV. But they do always bring closure, and "then almost everyone dies" is an entirely reasonable ending if the characters were on a suicide mission all along.

Re: "A Young Lady's Illustrated Primer" Simulated by GPT-4

#138

Post by cl42 »

angch wrote: post #134
art-not wrote: post #116

Earlier quoted context omitted.

Do you happen to know what book/story?

It's from Diamond Age, alluding to a character (YT) that was from Snow Crash.

Thanks for correcting me! That reference is even cooler than I expected.

Re: "A Young Lady's Illustrated Primer" Simulated by GPT-4

#139

Post by cyberpunk »

geophile wrote: post #19

I loved Diamond Age. My favorite Stephenson book, by far. But the ending, the resolution of all the action, was such a letdown. Such meh. I find that this is often a problem with cyberpunk.

Wow, I mean.. bit harsh? I’m working on it..

Re: "A Young Lady's Illustrated Primer" Simulated by GPT-4

#140

Post by drodgers »

cl42 wrote: post #22
geophile wrote: post #19

I loved Diamond Age. My favorite Stephenson book, by far. But the ending, the resolution of all the action, was such a letdown. Such meh. I find that this is often a problem with cyberpunk.

Do you think that's a Neal Stephenson issue? Don't get me wrong, I love his writing, but I found this to be the case with Fall too. I can see it becoming tiring if one is writing 400+ page books.

Definitely Stephenson. He has a tendency to write one-and-a-half books where the main drama resolves, then a lot of new context appears full of fresh challenges for our protagonists, and then the book ends.
Post reply on HN