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

Fear of AI just killed a useful tool

techdirt.com

Post reply on HN
Print view View original

41–50 of 309 posts

Re: Fear of AI just killed a useful tool

#41

Post by Nullabillity »

JoeAltmaier wrote: post #30
LikelyABurner wrote: post #3

I’m with the artists on this one. Our obsession with converting everything into input for an algorithm that spits out an ill-defined number (what the hell is “vividness”?) needs to stop. We already tried this with human communication and gave birth to the dystopian nightmare that is social media, why keep repeating our mistakes?

That's silly. Humans review books all the time, using very similar words. Where's the outrage over that? This is manufactured, stretched, overhyped objections. I believe it's all as the OP suggests, because the word AI is in there. Not because anything illegal or immoral is going on. In fact it's a terribly useful tool, and once the mob cools off it'll likely return.

They… quantify the number of adverbs and voices? I'm sorry, but have you ever read either a book or a review of one?

Re: Fear of AI just killed a useful tool

#42

Post by jprete »

runj__ wrote: post #15

I honestly don't understand how people can refuse the idea of that some parts most certainly can be reviewed with the help of a computer. There wasn't a "how good is this book" score because a computer might not be able to tell that yet, I don't understand the issue in looking at the number of adverbs in a book with the help of a computer.

Software people are intensely susceptible to the McNamara Fallacy:

“But when the McNamara discipline is applied too literally, the first step is to measure whatever can be easily measured. The second step is to disregard that which can't easily be measured or given a quantitative value. The third step is to presume that what can't be measured easily really isn't important. The fourth step is to say that what can't be easily measured really doesn't exist. This is suicide.”

— Daniel Yankelovich, "Interpreting the New Life Styles", Sales Management (1971)

https://en.m.wikipedia.org/wiki/McNamara_fallacy

This is often justified as an attempt to be objective, but I think the actual reason is that it’s hard to deal with unquantifiable properties.

Re: Fear of AI just killed a useful tool

#43

Post by robertlagrant »

zinekeller wrote: post #8
jprete wrote: post #4

This is an incredibly biased article, hinging entirely on the assumption that AI training is fair use.

Unfortunately it seems majority of HN clings on "AI is fair use!" the same way GIFs (of movie and TV snippets) are "perceived to be fair use" (but is simply not fair use in most cases).

"majority" and "clings on" seem entirely arbitrary and perjorative. Why not make your actual point instead of just beating up a straw man?

Re: Fear of AI just killed a useful tool

#44

Post by getabrainplz »

jprete wrote: post #25
pmarreck wrote: post #10

Earlier quoted context omitted.

There is no “AI” in the project. It literally just seems to do Bayesian sentiment analysis of books. It does NOT then mine that data to score OTHER books. Please actually read the article next time before commenting

I did in fact read the article. The author isn’t concerned about the book project. The author wants to use it as a springboard for complaining about AI fear, and his argument fundamentally hinges on the claim that AI training is fair use. He’s claiming the book project is comparable to AI training. I think the book project has a better claim to fair use than AI training does, although excerpting entire pages is maybe…

If I read the book, take notes about the number of times $x happened and later post them to my website, are you fine with that?

if I automate that process with a script that uses regular expressions, are you fine with that?

if I add a neural network to my script and it spits out the same numbers, are you still fine with that?

hopefully you see the madness in your argument.

Re: Fear of AI just killed a useful tool

#45

Post by jprete »

rgoulter wrote: post #35
jprete wrote: post #4

This is an incredibly biased article, hinging entirely on the assumption that AI training is fair use.

IANAL. But, what's your consideration such that you think this usage of the books wouldn't be "fair use"? e.g. in the US https://en.wikipedia.org/wiki/Fair_use factors of fair use are "purpose/character of the work", "nature of the copyrighted work", "amount/substantiality of the copyrighted work", "effect on the market for the copyrighted work". The website shows a few statistics computed from a book, and a few exce…

I think the site had a better claim by far than AI training does.

Re: Fear of AI just killed a useful tool

#46

Post by t14n »

Hrm. It seems like the authors are caught up in things like "vividness" score and the "sentiment analysis" of the text; I guess because it's loosely related to AI?

But it seems like a bulk of the stats collected are things that I would find really useful. I've probably asked myself, "how many words are in this book" on 10+ separate occasions, both as a reader and as a writer.

It also seems like there were also counts of things like adjectives, verbs, adverbs, passive verbs, etc -- stats that I might want to know about a novel.

The bulk of the service seems rather "boring" and non-AI. Unfortunate that the whole thing was taken down because of a few features. Hopefully it'll come back.

Re: Fear of AI just killed a useful tool

#47

Post by golemotron »

LikelyABurner wrote: post #3

I’m with the artists on this one. Our obsession with converting everything into input for an algorithm that spits out an ill-defined number (what the hell is “vividness”?) needs to stop. We already tried this with human communication and gave birth to the dystopian nightmare that is social media, why keep repeating our mistakes?

> I’m with the artists on this one. Our obsession with converting everything into input for an algorithm that spits out an ill-defined number (what the hell is “vividness”?) needs to stop.

Anyone who is with the artists should pass a law. Moral outrage is not law.

Re: Fear of AI just killed a useful tool

#48

Post by Nullabillity »

runj__ wrote: post #21
Nullabillity wrote: post #14

Earlier quoted context omitted.

Honestly, this is the really offensive part of the article. Who cares about whether or not it's legal, the idea that it's, in any way, shape, or form, useful is bafflingly laughable. Not everything can be meaningfully quantified. Not everything needs to be.

Certainly something interesting is bound to come out of quantifying things? "Hm, this three act structure thing seems to work, I wonder why." "Children doesn't seem to understand texts which include these words, I wonder why." Patterns rarely show themselves before we investigate.

> Certainly something interesting is bound to come out of quantifying things?

In science they call this trap P-hacking. Even data "scientists" know to be wary of overfitting. We're really good at finding patterns, but few of them actually mean anything.

Re: Fear of AI just killed a useful tool

#49

Post by belorn »

Relying on Authors Guild, Inc. v. Google, Inc in order determine fair use for AI models, there are a few key aspects to consider. The outcome should not supersede, supplant or become a replacement for the original works, nor should they sell portions of them. It should also preferable enhance the sale of the original work to the benefit of the copyright holder.

In this specific case regarding prosecraft, all those criteria might be fulfilled, and it might be so that under those specific conditions the use of any copyrighted work for the creation of AI models is fair (or at least under US law).

Re: Fear of AI just killed a useful tool

#50

Post by jprete »

getabrainplz wrote: post #44
jprete wrote: post #25

Earlier quoted context omitted.

I did in fact read the article. The author isn’t concerned about the book project. The author wants to use it as a springboard for complaining about AI fear, and his argument fundamentally hinges on the claim that AI training is fair use. He’s claiming the book project is comparable to AI training. I think the book project has a better claim to fair use than AI training does, although excerpting entire pages is maybe…

If I read the book, take notes about the number of times $x happened and later post them to my website, are you fine with that? if I automate that process with a script that uses regular expressions, are you fine with that? if I add a neural network to my script and it spits out the same numbers, are you still fine with that? hopefully you see the madness in your argument.

I’m thinking specifically of generative AI’s fair use claims. I thought the book site (and all of your extension examples) was clearly fair use.
Post reply on HN