#prosecraft

[crafantale]

The website Prosecraft has been nabbing popular and indie creators work for generative ai

The website has been deleted

But not the data

It's still being used for shaxpir software where they still get money

Thread

[seeingteacupsindragons]

God, this Prosecraft thing.

Is it, buddy? Are you sure? How, exactly, am I supposed to use any of that in my day-to-day writing? Why is this useful?

(Not even addressing the quality of this information, which is fairly dubious)

[lilithsaintcrow]

“The real targets of the AI backlash that swept Prosecraft away are the generative AI companies that are currently the toast of Silicon Valley, as well as the corporations planning to use those generative AI tools to replace human creative work."

Why the Great AI Backlash Came for a Tiny Startup You’ve Probably Never Heard Of

A literary analytics project called Prosecraft has shuttered after backlash from the writing community. It's a harbinger of a bigger cultura

Personally I'm disgusted by the attempt to "both-sides" Prosecraft's theft--the article flatly states (and Smith has admitted) that the "database" was built from stolen ebooks, literally taken from ebook theft sites.

That said, the article's still useful, I suppose.

[sainamoonshine]

In case anyone hadn't seen the absolute shitshow happening on twitter yesterday

Also this morning the dude said on his blog that he "deleted all the databases from all his servers" which has like. Big deleting evidence of a crime vibe

Anyway included in all the books he pirated and used to train his AI were Nora Roberts book, so he's mega done for. Aint no coming back from that one buddy, La Nora will find you and she will personally explain the process to you

[dduane]

PSA: "Shaxpir" AI writing software: AVOID!

The tl;dr: A guy is selling subscriptions to an AI-based software tool to "help you write better novels." And to train it, he's used tens of thousands of novels from authors you know, without those authors giving him permission.

...Sometimes things seem to blow up with unusual speed. This particular shit seems to have hit the fan yesterday, primarily on Twitter, when various authors discovered the guy's website, prosecraft.io. This site featured "clippings" of writing from the authors he'd stolen from... and the revelation that he had scraped their entire books, not just excerpts, to train his AI. ("2,470,720,986 words," his website bragged, "from 27,668 books, by 15,622 authors." The only authors who were off limits, apparently, were people using [or paying for] his software.) Though the guy hastily took prosecraft.io down when the online explosions began, if you take a look at this Google search you can see the covers of just some of the books the entire contents of which he exploited for AI training.

This usage goes well beyond the "fair use" defense that he belatedly (and ineffectively) attempted to employ on Twitter. It's straightforward copyright infringement, on a massive scale: good old-fashioned theft.

Gizmodo has a goodish breakdown of the broad situation here. AV Club also has one here.

The only upside to this sorry situation is that, at the legal end of things, this guy is certainly about to get nuked from orbit… because all those authors’ full-text works will still be in the guts of the guy’s AI, which is being used by him for commercial purposes. (Among the authors he made the gross tactical error of stealing from: Stephen King, James Patterson, the Pratchett Estate, and Nora Roberts. This... is not going to go well for him.)

Leverage's John Rogers sums it up succinctly:

Meanwhile: the guy who created this whole mess is still selling subscriptions to his Shaxpir software (I'm not adding the URL here) that he trained using stolen goods. So—until someone stops him—you might like to reblog this info for the attention of others here who prefer their writing to stay human-made as well as -fueled, and not to support the seriously ethically-challenged.