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

Why wordfreq will not be updated

github.com

Post reply on HN
Print view View original

51–60 of 542 posts

Re: Why wordfreq will not be updated

#51

Post by TacticalCoder »

aucisson_masque wrote: post #24

It could be used to spot LLM generated text. compare the frequency of words to those used in human natural writings and you spot the computer from the human.

> ... compare the frequency of words to those used in human natural writings and you spot the computer from the human.

But that's a losing endeavor: if you can do that, you can immediately ask your LLM to fix its output so that it passes that test (and many others). It can introduce typos, make small errors on purpose, and anything you can think of to make it look human.

Re: Why wordfreq will not be updated

#52

Post by creshal »

darby_nine wrote: post #43
kevindamm wrote: post #32

Earlier quoted context omitted.

Yes but not quite as far as you imply. The training data is weighted by a quality metric, articles written by journalists and wikipedia contributors are given more weight than Aunt May's brownie recipe and corpoblogspam.

Aunt may's brownie recipe (or at least her thoughts on it) are likely something you'd want if you want to reflect how humans use language. Both news-style and encyclopedia-style writing represent a pretty narrow slice.

That's why search engines rated them highly, and why a million spam sites cropped up that paid writers $1/essay to pretend to be Aunt May, and why today every recipe website has a gigantic useless fake essay in front of their copypasted made up recipes.

Re: Why wordfreq will not be updated

#53

Post by dsign »

Somehow related, paper books from before 2020 could be a valuable commodity in a in a decade or two, when the Internet will be full of slop and even contemporary paper books will be treated with suspicion. And there will be human talking heads posing as the authors of books written by very smart AIs. God, why are we doing this????

Re: Why wordfreq will not be updated

#54

Post by thfuran »

Lvl999Noob wrote: post #36
aucisson_masque wrote: post #24

It could be used to spot LLM generated text. compare the frequency of words to those used in human natural writings and you spot the computer from the human.

It could be used to differentiate LLM text from pre-LLM human text maybe. The thing, our AIs may not be very good at learning but our brains are. The more we use AI, the more we integrate LLMs and other tools into our life, the more their output will influence us. I believe there was a study (or a few anecdotes) where college papers checked for AI material were marked AI written even though they were written by human…

>our AIs may not be very good at learning but our brains are

Brains aren't nearly as good at slightly adjusting the statistical properties of a text corpus as computers are.

Re: Why wordfreq will not be updated

#55

Post by jsheard »

Freak_NL wrote: post #42
kevindamm wrote: post #32

Earlier quoted context omitted.

Yes but not quite as far as you imply. The training data is weighted by a quality metric, articles written by journalists and wikipedia contributors are given more weight than Aunt May's brownie recipe and corpoblogspam.

It certainly feels like the amount of regurgitated, nonsensical, generated content (nontent?) has risen spectacularly specifically in the past few years. 2021 sounds about right based on just my own experience, even though I can't point to any objective source backing that up.

SEO grifters have fully integrated AI at this point, there are dozens of turn-key "solutions" for mass-producing "content" with the absolute minimum effort possible. It's been refined to the point that scraping material from other sites, running it through the LLM blender to make it look original, and publishing it on a platform like Wordpress is fully automated end-to-end.

Re: Why wordfreq will not be updated

#56

Post by surfingdino »

aucisson_masque wrote: post #18

Did we (the humans) somehow managed to pollute the internet so much with AI that's it's now barely usable ? In my opinion the internet can be considered as the equivalent of a natural environment like the earth. it's a space where people share, meet, talk, etc. I find it astonishing that after polluting our natural environment we know polluted the internet.

Yes. Here are practical instructions on how to turn it into an even more of a cesspit https://www.youtube.com/watch?v=endHz0jo9Ck I think it's now a law of nature that any new tech leads to SEO amplification. AI has become the Degelman M34 Manure Spreader of the internet https://degelman.com/products/manure-spreaders

Re: Why wordfreq will not be updated

#57

Post by darby_nine »

creshal wrote: post #52
darby_nine wrote: post #43

Earlier quoted context omitted.

Aunt may's brownie recipe (or at least her thoughts on it) are likely something you'd want if you want to reflect how humans use language. Both news-style and encyclopedia-style writing represent a pretty narrow slice.

That's why search engines rated them highly, and why a million spam sites cropped up that paid writers $1/essay to pretend to be Aunt May, and why today every recipe website has a gigantic useless fake essay in front of their copypasted made up recipes.

Ok, but what i said is true regardless of SEO, and that SEO has also fed back into english before LLMs were a thing. If you only train on those subsets you'll also end up with a chatbot that doesn't speak in a way we'll identify as natural english.

Re: Why wordfreq will not be updated

#58

Post by hhh »

next_xibalba wrote: post #45

[flagged]

I don't care about the political part, but twitter used to have nice, high-enough quality expert discussion around topics, and now its just a shithole with very stupid takes polluting the few spots free in replies between engagement farmers and llm slop-spewers
Post reply on HN