Hacker News Reader

Comment by al_be_back | Hacker News Reader

I think it's wise to take the concerns of the creative community seriously - after all their "labor of love" [1] matters immensely, without it LLMs are useless.

matters not how much the coder "loved" the project, or did yoga, or that they've not made money for years, after all most book authors aren't exactly raking in the money either.

also like many thinks in life, some tools/projects/startups etc just stop being needed/used and new ones/competitors take over. there's nothing to say that since tool X is using A.I. therefore it has to be adopted by one and all; smiles all around.

google has 'right to be forgotten', also looking into 'machine unlearning' and it's common for platforms to honor user's request to remove their data / close their account.

[1] From OP: destroying what had been a clear labor of love and a useful project

View on HN

The thing with this project is that it had no conceivable way of threatening the original authors, financially or otherwise. The analyses it produced weren't replacements of the original works, and served as a writing tool, not as something that generated new content.

In this case, the shutdown seemed to hinge entirely on the irrational fear of anything with "AI" written on it and an unshakeable conviction that this clearly transformative use case wasn't fair use.

>> had no conceivable way of threatening the original authors, financially or otherwise

how so? what's inconceivable about it?

>> seemed to hinge entirely on the irrational fear

how are the authors "fearing without reason" or "illogically fearing" ?

> how so? what's inconceivable about it?

Authors make money through sales of their work. This tool was a writing aid that analyzed text and included some copyrighted works in its dataset. There was no way of retrieving these books in full, and the excerpts that were allegedly shown to the users were used in an analytical context, unlike the original works. So, this website couldn't replace ownership of the actual book for readers, and had no capacity to hurt their sales. Basically, the way the service used this data should be transformative enough as to not have any impact on the authors.

> how are the authors "fearing without reason" or "illogically fearing" ?

I called it "fear" because there was no strong argument on the authors' side as to why this tool is bad. I called it "illogical" because I think that it's no coincidence that this controversy only came up now, in 2023. Back in 2017 and onward, the existence of this tool didn't appear to generate pushback. My pet theory is that in 2023, now that we have good generative AI, the advancements have spawned an entire subset of people that view anything "AI" as inherently tainted and immoral. The original complaints seem to lack understanding of this tool and multiple people have conflated it with generative AI, despite it having nothing to do with that.

>> Basically, the way the service used this data

it's not about the past (2017...), the authors are concerned about how the dataset could be or is likely to be used from now on.

many tech projects these days have or are thinking about integrating third-party A.I. providers in their services, either to harness the power of their large datasets or their large user-base. I think it's great if authors/users opt-in to this, but likewise I agree with those that want out (opt-out).

>> I called it "fear" because there was no strong argument on the authors' side as to why this tool is bad

their argument doesn't have to be a peer-reviewed journal, it suffices to say "i don't want my books in your dataset"

>> I called it "illogical" because I think that it's no coincidence that this controversy only came up now, in 2023. Back in 2017 and onward, the existence of this tool didn't appear to generate pushback ...

6 years have passed since 2017, life moves on, it's natural for things to change e.g: the project's code, updating servers, partnerships, emergence of third-party tools/libs/services etc etc.

> the authors are concerned about how the dataset could be or is likely to be used from now on.

The service in question wasn't introducing anything new that'd appear to justify all the recent pushback. It feels like you continue generalizing your statements, while the discussion topic is about what made Prosecraft specifically so preposterous that it warranted the outrage.

> their argument doesn't have to be a peer-reviewed journal, it suffices to say "i don't want my books in your dataset"

It's kind of a blunt statement, but why should they have a say? For example, say I create a website where I publish technical analysis of famous literary works, including basic statistics about a book and a review. Should the authors be able to just take that down? This use is legally protected (as is creating a dataset), so allowing authors to restrict this use seems as arbitrary as allowing them to say that no person can ever bring their books into the country of Moldova or that no one over the age of 50 may read it.

> it's natural for things to change e.g: the project's code, updating servers, partnerships, emergence of third-party tools/libs/services etc

And yet, in this specific situation, all of this is conjecture. Nothing about the project changed in some significant way in 2023 that would warrant this. Further proving it is that the people that are against Prosecraft don't seem to bring up any specific changes or reasons for their stance, only that it is "AI".

as I said earlier "6 years passed" since 2017 and the arrival of A.I. is causing authors (to say the least) real concern. the industry/ecosystem around X is enough to change to affect those associated with X.

you offered a "pet theory" (as you say) further up, but not much by way of proof that their fear is irrational.

i think we've reached in impasse here, the points are re-cycling.

> the industry/ecosystem around X is enough to change to affect those associated with X.

I think that you keep trying to generalize this argument because there's nothing that's strongly damning about Prosecraft. If these authors have a concern in regards to newer AI tech, they should address these concerns to services that actually implement that tech, not a tool that has peacefully existed for years and is guilty of nothing but being vaguely associated with these newer developments. It's like if Ford released a new car that could go 1000km/h, and then I, in protest of its obvious lack of safety, go to a museum and smash up a Model T. If that's not irrational, I don't know what would be.

> you offered a "pet theory" (as you say) further up, but not much by way of proof that their fear is irrational.

There's no concrete proof because I didn't find anything that I could even try to disprove. If you read through the anti-Prosecraft posts, they lack substance in bringing up how exactly this service can financially harm authors. The only complaint seems to be that the authors don't want it - even though the use of their works appears legal and is perfectly consistent with what many other services have done in the past. Thus, my only conclusion was that at least some of the hate stemmed from Prosecraft just being vaguely "something with AI".

Reply on news.ycombinator.com