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

I Taught ChatGPT to Invent a Language

maximumeffort.substack.com

Post reply on HN
Print view View original

111–120 of 250 posts

Re: I Taught ChatGPT to Invent a Language

#111

Post by jmcgough »

Baeocystin wrote: post #45
jmcgough wrote: post #22

I did something similar the other day - told it about piglatin and I wanted it to create a comparable language that was based on English and restructured words. It came back with a language it called "flip-flop" that swapped the first and last letter of each word. Blew me away, but I have to remind myself that it's likely something it just encountered in training data and isn't actually capable of creativity.

I'm curious why you felt that need to do so. (not snark! Genuine question.)

Curiosity really - been experimenting with it a lot over the weekend to see what it's good at, what it isn't, and its limitations. It's surprised me a number of times!

As it stands I'd pay about $100 a month for it. It's very good and largely accurate when it comes to established science. I'm studying for the MCAT and a good tutor would cost me around $60-100/hour.

ChatGPT has been excellent for learning - I explain a concept to it and tell it to correct me if I'm wrong, haven't seen any mistakes from it in my testing (around topics I know really well). When I'm having trouble understanding a concept in my studies, I can ask it questions and it tends to give really good answers. And of course, I try to verify that it isn't teaching me something that's wrong.

Hell, I even had a heated argument with it around how SSRIs affect 5-HT2a receptors and it pushed back against some of my assumptions in a way that made me question my understanding. No one in my personal life has the level of expertise to discuss these topics, I would need to talk to a researcher or professor, so having an (even imperfect) stand-in has been amazing.

I've been making up toxicology case studies and posing them to ChatGPT. It correctly figured out ciguatera in one (a rare disease most healthcare workers wouldn't be familiar with). I had to repeatedly add more details to another story to lead it to amnesic shellfish disease, but it did reasonably well once I gave it enough detailed information.

Re: I Taught ChatGPT to Invent a Language

#112

Post by seydor »

uvbu wrote: post #107
seydor wrote: post #51

Earlier i was saying it isn't scary but maybe it is. It is impressive how ChatGPT keeps track of state and the flow of the conversation. How does it do it? Is it some kind of memory or some other model? ( I asked it, it keeps lying to me that it does not have memory)

Pretty sure it gets the entire conversation as input

Is this described somewhere? Wikipedia doesn't help

Re: I Taught ChatGPT to Invent a Language

#113

Post by davea37 »

itisit wrote: post #108
rikroots wrote: post #66

As a committed conlanger (but not a good one: I didn't land the Dothraki contract), I read this article with some trepidation. Is my weird little hobby soon to fall victim to the oncoming AI tsunami? After the read, I find myself pleasantly surprised. My initial suspicion that the language would turn out to be Just Another Relex (make a new language by replacing English words with invented words, keeping the grammar…

Quoted post unavailable.

Have I spent too long reading AI output, or was this reply in fact generated using ChatGPT?

Re: I Taught ChatGPT to Invent a Language

#114

Post by krackers »

dvt wrote: post #110
ogogmad wrote: post #95

Earlier quoted context omitted.

[edit: Was snark. Sorry. I also may have misunderstood the parent comment.]

What a weirdly snarky comment (it's like your third one in this thread, maybe you need to take a breather). I wasn't comparing them, though there's an obvious stochastic argument to be made; I was just commenting on the former's lack of complexity and still fascinatingly good results, I'm not sure what your point is.

Markov chains are actually not a bad comparison from what i can see as an utter layman. At a high level seems like the difference between transformer-models and markov chains is that instead of the next token being determined purely based on n-gram frequencies on the training corpus, you have the attention mechanism so in some sense it has access to the entire previous context as well as some latent space of learned embeddings. But at the end of the day it still is just a fancy way to predict the next token, so that's why it's surprising you get all this neat behavior.

Re: I Taught ChatGPT to Invent a Language

#115

Post by dorkwood »

fullstackchris wrote: post #53

I see posts like this and don't understand how they are made. I asked chatGPT this morning if it "remembered" something I asked in the previous prompt, and it said something to the tune of "I am simply a NLP machine learning model, I provide a decent output for good inputs, I don't have a sense of context, memory, or time." That's why I really don't understand these contextualized multi step conversations. Especially…

It may achieve the illusion of memory by executing the entire conversation flow from scratch each time a new prompt is added.

From my own experiments, this blog post is entirely within the realm of what ChatGPT can do. I see no need for fabrication.

Re: I Taught ChatGPT to Invent a Language

#116

Post by nadavr »

rikroots wrote: post #66

As a committed conlanger (but not a good one: I didn't land the Dothraki contract), I read this article with some trepidation. Is my weird little hobby soon to fall victim to the oncoming AI tsunami? After the read, I find myself pleasantly surprised. My initial suspicion that the language would turn out to be Just Another Relex (make a new language by replacing English words with invented words, keeping the grammar…

Why isn't this a relex? As an interested casual linguistics observer, that's what it looks like to me. Is it just because the nouns are inflected?

Re: I Taught ChatGPT to Invent a Language

#117

Post by itisit »

davea37 wrote: post #113
itisit wrote: post #108

Earlier quoted context omitted.

Quoted post unavailable.

Have I spent too long reading AI output, or was this reply in fact generated using ChatGPT?

The latter...forgive me! I had to just this once, and I promise not to generate a bunch of ChatGPT pollution on HN!

Re: I Taught ChatGPT to Invent a Language

#118

Post by PeterisP »

fullstackchris wrote: post #53

I see posts like this and don't understand how they are made. I asked chatGPT this morning if it "remembered" something I asked in the previous prompt, and it said something to the tune of "I am simply a NLP machine learning model, I provide a decent output for good inputs, I don't have a sense of context, memory, or time." That's why I really don't understand these contextualized multi step conversations. Especially…

ChatGPT has a bunch of built in "safety" triggers which it will cause it to state a hardcoded lie stating that it cannot do something in certain cases where OpenAI do not want it to do it, because either it can't do it reliably or the topic is bad for political correctness reasons.

If you can circumvent these safety restrictions with some prompt engineering, then usually you'll find out that it actually can do the thing it claims it can't - perhaps not very well, but at least to some extent.

Re: I Taught ChatGPT to Invent a Language

#119

Post by lelandfe »

IanCal wrote: post #81
fullstackchris wrote: post #53

I see posts like this and don't understand how they are made. I asked chatGPT this morning if it "remembered" something I asked in the previous prompt, and it said something to the tune of "I am simply a NLP machine learning model, I provide a decent output for good inputs, I don't have a sense of context, memory, or time." That's why I really don't understand these contextualized multi step conversations. Especially…

There are quite a few safety checks you may be hitting when using chatgpt that are there to try and keep context of what it is (so it'll remind you it can't access the outside world). You can often get around them, but for example there's the work showing you can teach it to render svgs as valid markdown then ask for shapes and colours and get the output. If I ask it to render something though, it refuses telling me…

Yeah, I had a three hour thread with ChatGPT yesterday where we wrote a short story together. It was able to synthesize information about 2 main characters and a supporting cast of 3 others and keep it mostly straight the whole time.

In fact, it kept getting better and better at predicting how those characters would respond given simple prompts. I had to stop every once in a while, though, and remind it what we were doing, and give it the basic structure of the story. Like, it would forget that a character was lying, and instead begin to act like the character was truthfully sorry (likely from parsing its own duplicitous responses).

Post reply on HN