Thread by @jason_kint

ok, I've now read the NYT response this week to attempts by OpenAI to dismiss NYT's landmark lawsuit against the high-flying AI company.
Put simply, NYT makes it brutally clear on page one how you can tell the difference between the two companies.
Oomph. /1

A few other observations from me. Like NYT's original complaint, it's smart and future-focused on fair value. Where OpenAI made frankly bizarre claims NYT was hacking the platform as it detected OpenAI had its content, NYT is right. OpenAI isn't and can't dispute it copied it. /2

Um, 2022 > 2020 = TRUE. Where OpenAI tried to inject a statute-of-limitations argument that OpenAI's lifting of content was "common knowledge" in 2020, NYT points out that ChatGPT and OpenAI didn't go viral until Nov 2022. /3

This can't be pointed out enough so I'm glad NYT reminds everyone not only how the OpenAI business was built but that it has shifted from a non-profit front to a full-blown for-profit. This shift also plays a major role in Elon Musk's lawsuit against OpenAI. /4

In fact, these numbers capture what is going on with the allegation that OpenAI has misappropriated almost a century of copyrighted journalism helping boost its market cap by a trillions dollars.
This may well describe one of the biggest violations of the law in history. /5

And how was OpenAI aware that it was happening? "because The Times told it so." /6

In fact, NYT wisely picks up on OpenAI's defense by blog and PR spin including a claim that OpenAI was "working hard" to address the issues in the NYT complaint. That's proof of knowledge of the infringements if I've ever seen it. /7

And like in the complaint, there is an important point that the violations happen in the training and in the output. In fact, NYT's copyright [management information] is allegedly stripped out of the content during the training and then in the output. /8

Yes, OpenAI's (bad) argument is that the notice is small and at the bottom of the page which common sense says is irrelevant to their machines ripping the content who simply don't seem to care about copyright (or journalism) because they've likely been told it doesn't matter. /9

Here is where it certainly matters and a super important point in the lawsuit. This failure to recognize copyright through licensing results in harm to publishers by substituting a fast-growing licensing market with one where content can be lifted for free (aka stolen). /10

and so I agree with the Conclusion here. OpenAI's isn't even disputing what NYT claimed they're doing in the complaint. Taking all of those facts in, the MtD should be denied, let's get on with it. Value is being ripped. /11

Here is a link to NYT's full response to the MtD. /12 storage.courtlistener.com/recap/gov.usco…

and a link to my original thread on the lawsuit filed over the holidays. ht to @dcnorg, please also give it a follow. cheers. /13

https://x.com/jason_kint/status/1740141400443035785?s=20