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

Margaret Atwood Reviews a Margaret Atwood Story by AI

thewalrus.ca

Post reply on HN
Print view View original

51–60 of 79 posts

Re: Margaret Atwood Reviews a Margaret Atwood Story by AI

#51

Post by ACow_Adonis »

michaelmrose wrote: post #11

I don't know anything about poetry but I do know you have to put some effort into a prompt and iterate to get anything useful out. Here are some interesting results from literally feeding this article and prompting chatgpt to improve the poem and take into account her critique. They seem at least to my eyes a little better. =========================================================== Upon these shores, where time seem…

It reads like a someone tried to commit poetry by rhyming dot points in a list. Except they got a discount bulk-rate on purchasing commas at the grammar store so they stitched it all together with 50 of them and called it done :P

Plus like allmost all of its output, it appears to have the amount of profundity, aesthetic pleasure, insight and interest usually found in corporate boardrooms and marketing brochures. That is to say, vanilla dross.

Re: Margaret Atwood Reviews a Margaret Atwood Story by AI

#52

Post by crooked-v »

darkerside wrote: post #23
maximusdrex wrote: post #13

Earlier quoted context omitted.

I find your reaction intriguing since I had nearly the opposite reaction. Especially when it comes to the second piece I had little to critique about it technically (maybe a little from a structural standpoint but not as much grammatically), but I found it incredibly dull due to its complete lack of any real point. With your musical analogy, it felt to me like a piece where all of the instruments were played perfectl…

Agree. This reminds me of a kid's school assignment, which sounds impressive, but really misses the point. A kid with no real writing talent or experience given this same prompt would also cargo cult some uncohesive drivel like this.

That lines up pretty well with how much ChatGPT responses ape the stereotypical school essay style of being 1/5th introduction, 3/5ths content, and 1/5ths conclusion that reiterates the content.

Re: Margaret Atwood Reviews a Margaret Atwood Story by AI

#53

Post by famouswaffles »

armchairhacker wrote: post #32

I think a while ago I commented something along the lines of "let me know when we see a successful book/article/speech which gets revealed to be largely AI-generated". And that hasn't happened yet: AI-generated content has always been noticeable and generally considered bad ("was X written by ChatGPT?" is an insult). But I know AI is already being used to assist human writers, not just with boring emails and speeches…

LLMs can hit human quality writing just fine (not professional yet). The top LLMs today are all deliberately trained to sound bland, robotic and uninspired with rlhf etc. It's just the default voice, not some weakness of LLMs and it's not very hard to make them not sound like that.

Re: Margaret Atwood Reviews a Margaret Atwood Story by AI

#54

Post by inciampati »

dools wrote: post #31

The Weeping Willows of Winnipeg is a shit story, but if you were working on a short story, and you got 5 suggested rewrites for a given paragraph, or you were looking for ideas for a plot point or something similar, then you could use ChatGPT to help you out. In exactly the same way, sometimes I give ChatGPT a complete coding task and it can't do the job. But while I'm working on code I can get it to do certain thing…

It's most powerful as a tool for transforming and projecting text with text. It also excels at lateral exploration, enumeration, comparative critique. But yeah, don't ask it to write you a story and expect a gem.

Re: Margaret Atwood Reviews a Margaret Atwood Story by AI

#55

Post by inciampati »

boplicity wrote: post #7

Reading this, I realize that the "prompt" is what's lacking, in terms of the output produced. A writer has a prompt too -- which is created through the entirety of their experience leading up to the moment they sit down to write, often heavily weighted towards the hours leading up to that moment. AI will catch up to actual human writers when it can spend a day's visual, auditory, and linguistic stimuli crafting the p…

Beautifully said. I can add only that the right pattern of prompts may drive the creation of legitimately useful and even beautiful work. Humans need to guide the machines, maybe precisely because they lack the evolutionary drive that gets us up in the morning and keeps us cleaning our babies' behinds.

Re: Margaret Atwood Reviews a Margaret Atwood Story by AI

#56

Post by inciampati »

FrustratedMonky wrote: post #14
garba_dlm wrote: post #9

Earlier quoted context omitted.

it's not a prompt, it's a dialog, it's many prompts.

Yes When I have a conversation with GPT4, it keeps the history, and the entire history is a prompt for the next prompt. I would say, not totally unlike a human memory. just GPT4 is limited to 8000 characters or something. Humans have a bit more, but with added 'forgetfulness', so who knows how many characters for real. Humans don't actually recall perfect text very well.

But we do manage to remember stuff that happened "quadrillions" of tokens (of sensory input) back. Sure, we are wrong a lot, but we have a context that goes way back. A LLM has a fruit fly's version of what we work with, although in some versions it has perfect recall in that context.

Re: Margaret Atwood Reviews a Margaret Atwood Story by AI

#57

Post by passion__desire »

dools wrote: post #31

The Weeping Willows of Winnipeg is a shit story, but if you were working on a short story, and you got 5 suggested rewrites for a given paragraph, or you were looking for ideas for a plot point or something similar, then you could use ChatGPT to help you out. In exactly the same way, sometimes I give ChatGPT a complete coding task and it can't do the job. But while I'm working on code I can get it to do certain thing…

There is a story of a author who used to cutout words from newspaper and randomly arrange them in bogosort manner and find good combination of words.

At one time he found, "son sues father" which came true as a headline few years later.

Re: Margaret Atwood Reviews a Margaret Atwood Story by AI

#58

Post by CSMastermind »

darkerside wrote: post #48
danenania wrote: post #34

I wonder which model was used for this? Based on the poem taking "10 seconds" to generate, I'd guess the free version of ChatGPT, meaning 3.5 turbo. While I wouldn't expect Atwood's conclusions to change too much by using GPT-4 instead, I think it's interesting that even the majority of educated people and journalists outside of tech don't seem to realize that the best model is at least 10x smarter than the free vers…

I hear this a lot. I didn't notice a huge difference in quality with GPT4. Completely anecdotal, and could have been a failure to effectively prompt for that model. But I don't think it's safe to assume the results are 10x improvement.

I have. I don't propose some kind of scientific measure but I do have two data points to contribute:

First, I've been using GPT to build an application for work for the past few months and anything but GPT-4 consistently produces less consistent and reliable output. Things like occasionally producing malformed JSON.

Second, I have a set of questions I use to evaluate models testing different capabilities and GPT-4 does much better than other models, particularly at coding tasks. There are some exceptions, for example, Bard has been able to do better on stating facts sometimes and Claude has done better at summarizing long text.

I'd love to have another model as good as GPT-4 to use but I haven't found one yet.

Post reply on HN