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

OpenAI departures: Why can’t former employees talk?

vox.com

Post reply on HN
Print view View original

521–530 of 1001 posts

Re: OpenAI departures: Why can’t former employees talk?

#521

Post by jeffreygoesto »

andyjohnson0 wrote: post #490
mwigdahl wrote: post #290

The best approach to circumventing the nondisclosure agreement is for the affected employees to get together, write out everything they want to say about OpenAI, train an LLM on that text, and then release it. Based on these companies' arguments that copyrighted material is not actually reproduced by these models, and that any seemingly-infringing use is the responsibility of the user of the model rather than those w…

Clever, but the law is not a machine or an algorithm. Intent matters. Training an LLM with the intent of contravening an NDA is just plain . Everyone would still get sued anyway.

But then training a commercial model is done with the intent to not pay the original authors, how is that different?

Re: OpenAI departures: Why can’t former employees talk?

#522

Post by throwaway2037 »

calvinmorrison wrote: post #510
throwaway2037 wrote: post #500

Earlier quoted context omitted.

> LLMs not being copyright laundromats This a brilliant phrase. You might as well put that into an Emacs paste macro now. It won't be the last time you will need it. And the OP is classic HN folly where programmer thinks laws and courts can be hacked with "this one weird trick".

But they can, just look at AirBnB, Uber, etc.

No, lots of jurisdictions outside the US fought back against those shady practices.

Re: OpenAI departures: Why can’t former employees talk?

#523

Post by surfingdino »

throwaway2037 wrote: post #505
surfingdino wrote: post #487

Earlier quoted context omitted.

MS Word does not actively collect and process all texts for all available sources and does not offer them in recombined form. MS Word is passive whereas the whole point of an LLM is to produce output using a model trained on ingested data. It is actively processing vast amounts of texts with intent to make them available for others to use and the T&C state that the user owns the copyright to the outputs based on work…

First, I agree with nearly everything that you wrote. Very thoughtful post! However, I have some issues with the last sentence. > Collateralised Copyright Liability Is this a real legal / finance term or did you make it up? Also, I do not follow you leap to compare LLMs to CDOs (collateralised debt obligations). And, do you specifically mean CDO or any kind of mortgage / commercial loan structured finance deal?

My analogy is based on the fact that nobody could see what was inside CDOs nor did they want to see, all they wanted to do was pass them on to the next sucker. It was all fun until it all blew up. LLM operators behave in the same way with copyrighted material. For context, read https://nymag.com/news/business/55687/

Re: OpenAI departures: Why can’t former employees talk?

#524

Post by Sharlin »

ben_w wrote: post #436
throwaway598 wrote: post #250

Earlier quoted context omitted.

That OpenAI are institutionally unethical. That such a young company can be become rotten so quickly can only be due to leadership instruction or leadership failure.

We already know there's been a leadership failure due to the mere existence of the board weirdness last year; if there has been any clarity to that, I've missed it for all the popcorn gossiping related to it. Everyone including the board's own chosen replacements for Altman siding with Altman seems to me to not be compatible with his current leadership being the root cause of the current discontent… so I'm blaming Mi…

Everyone? What about the board that fired him, and all of those who’ve left the company? It seems to me more like those people are leaving who are rightly concerned about the direction things are going, and those people are staying who think that getting rich outweighs ethical – and possibly existential – concerns. Plus maybe those who still believe they can effect a positive change within the company. With regard to the letter – it’s difficult to say how many of the undersigned simply signed because of social pressure.

Re: OpenAI departures: Why can’t former employees talk?

#525

Post by StackRanker3000 »

ergocoder wrote: post #412
lazyasciiart wrote: post #352

Earlier quoted context omitted.

I think they mean that you had to buy all the ones you wanted to keep.

That is tautological... You buy what you want to own???

The point being made is that it isn’t all or nothing, you can buy half the vested options and forfeit the rest, should you want to.

Re: OpenAI departures: Why can’t former employees talk?

#526

Post by KoolKat23 »

surfingdino wrote: post #487
KoolKat23 wrote: post #421

Earlier quoted context omitted.

You don't even need to go this far. The word-probabilities are transformative use, a form of fair use and aren't an issue. The specific output at each point in time is what would be judged to be fair use or copyright infringing. I'd argue the user would be responsible for ensuring they're not infringing by using the output in a copyright infringing manner i.e. for profit, as they've fed certain inputs into the model…

MS Word does not actively collect and process all texts for all available sources and does not offer them in recombined form. MS Word is passive whereas the whole point of an LLM is to produce output using a model trained on ingested data. It is actively processing vast amounts of texts with intent to make them available for others to use and the T&C state that the user owns the copyright to the outputs based on work…

Legally, copyright is only concerned with the specific end work. A unique or not so unique standalone object that is being scrutinized, if this analogy helps.

The process involved in obtaining that end work is completely irrelevant to any copyright case. It can be a claim against the models weights (not possible as it's fair use), or it's against the specific once off output end work (less clear), but it can't be looked at as a whole.

Re: OpenAI departures: Why can’t former employees talk?

#527

Post by kdnvk »

jeffreygoesto wrote: post #521
andyjohnson0 wrote: post #490

Earlier quoted context omitted.

Clever, but the law is not a machine or an algorithm. Intent matters. Training an LLM with the intent of contravening an NDA is just plain . Everyone would still get sued anyway.

But then training a commercial model is done with the intent to not pay the original authors, how is that different?

It’s not done with the intent to infringe copyright.

Re: OpenAI departures: Why can’t former employees talk?

#528

Post by b112 »

mwigdahl wrote: post #290

The best approach to circumventing the nondisclosure agreement is for the affected employees to get together, write out everything they want to say about OpenAI, train an LLM on that text, and then release it. Based on these companies' arguments that copyrighted material is not actually reproduced by these models, and that any seemingly-infringing use is the responsibility of the user of the model rather than those w…

Copyright != an NDA. Copyright is not an agreement between two entities, but a US federal law, with international obligations both ratified and not.

Copyright has fair uses clauses, endless court decisions limiting its use, carve outs for libraries, additional junk like the DMCA and more slapped on top. It's a patchwork of dozens of treaties and laws, spanning hundreds of years.

For example, you can read a book to a room full of kids, you can use copyright materials in comedic skits, you can quote snippets, the list goes on. And again, this is all legislated.

The point? It's complex, and specific usage of copyrighted works infringing or not, can be debatable without intent immediately being malign.

Meanwhile, an NDA covers far, far more than copyright. It may cover discussion and disclosure of everything or anything, including even client lists, trade secrets, work processes, and more. It is signed, and agreed to by both parties involved. Equating "copyright law" to "an NDA" is a non-starter. There's literally zero legal parallel or comparison here.

And as others have mentioned, the intent of the act would be malicious on top of all of this.

I know a lot of people dislike the whole data snag by OpenAI, and have moral or ethical objections to closed models, but thinking anyone would care about this argument if you breach an NDA is a bad idea. No judge would even remotely accept or listen to such chicanery.

Re: OpenAI departures: Why can’t former employees talk?

#529

Post by SuchAnonMuchWow »

throwaway2037 wrote: post #520
NewJazz wrote: post #344

Earlier quoted context omitted.

My US history teacher taught me something important. He said that if you are going to steal and don't want to get in trouble, steal a whole lot.

Very interesting post! Can you share more about your teacher's reasoning?

It likely comes from the saying similar to this one: "kill a few, you are a murderer. Kill millions, you are a conqueror".

More generally, we tend to view number of causalities in war as a large number, and not as the sum of every tragedies that it represent and that we perceive when fewer people die.

Post reply on HN