These people have weightier issues like the ethics of the hypothetical technological future on their capacious minds, and mustn't be dragged down to your mucky little level of discussing the ethics of the very real business present.
Seriously: don't bother them with these little questions of the rights of actually enumerable humans to whom they have contractual responsibilities! They're busy thinking in great depth about how their product might affect the rights of all the humans they haven't yet sold it to.
One does wonder how the accelerationist ethic conceives of the tradeoff between a hundred future happy fulfilled lives or two hundred permanently indentured serfs.
"My 'we never actually took away people's remuneration using the contracts they signed swearing them to silence' t-shirt has people asking a lot of questions already answered by my shirt"
It doesn't need to be the same article to be a duplicate submission, it just needs to have no significant new information compared to recent submissions that already got a lot of HN attention.
This article is literally just summarizing the Vox article with no new information or insight. And the HN submission for the Vox article had 1784 points and 544 comments.
> "We want to make sure you understand that if you don't sign, it could impact your equity," one rep told an outgoing employee, according to Vox.
The article is doing a lot of click-bait stuff here. The quote, without the source, is shown at the top of the article, near Sam Altman's face and a title referencing Sam Altman. However, the quote is not from Sam Altman, which makes its placement a bit disingenuous.
I am not invested in defending Altman, but this type of journalism is trash clickbait and a far bigger issue than a tech company threatening employees' vested equity unless they comply in some sort of non-NDA NDA scheme.
That's looking purely at the headline and a single quote that they chose for it. There is a lot more if you read the actual article.
They make a case that Altman signed documents where it's quite clear the intention is to be able to claw back vested equity. And that quote in the headline isn't meant to try and attribute it directly to Altman -- but to show just how well-known the policy was.
I read the whole article, that's where I found the quote attribution. Like I said, I'm not here to defend Altman, or attack him. It's just disingenuous journalism - or click-bait journalism.
They never were in sheep’s clothing. They took $10 billion from Microsoft. Anyone closely partnering with a top wolf is probably a wolf cub on a journey into the wolf pack.
I meant after the deal. They might have been OK up to that point.
I think the case study still raises questions about whether we should require specific licensing or contractual terms for what these organizations produce. For instance, that they dual-licensed all their models or (if safety is a worry) always shared them with research institutions. That by itself might have given us GPT3 or GPT4 with the Llama- and Falcon-type communities building on them.
I love when a website uses the word "credibility", and half the website is filled with flashing ads and it has too many trackers and ad scripts to count.
I must be missing some key detail here because this seems even more over blown than AI itself. They wanted to motivate people to not disclose trade secrets? Oh no.
The key details are how completely outside the norms of SV / tech this behavior is; how they, and Sam specifically, clearly lied about it - especially egregious when they want to be seen as the trustworthy steward of "AGI"; and as a business story, how this seems to be eroding trust of the company among their current employees and how damaging this will likely be in their efforts to recruit top AI talent. I don't think there will be quite so many heart emoji tweets for Sam in the future.
Clawing back vested equity it materially different that a general “incentive”. It’s way outside of standard practice and borderline abuse given the imbalance of negotiating power. Further, if OpenAI or Sam lied publicly by saying they weren’t aware it was going on and did this seems to be the very type of untrustworthy lack of candor that the board identified when they ousted him.
I was a vocal defender of Sam in the immediate aftermath but these facts are definitely concerning. OpenAI is moving fast with a critical and dangerous technology. If their culture isn’t profoundly ethical they shouldn’t be trusted with the privilege and responsibility of what they are up to.
You are making factual errors in your response to the existence of this article.
First, this was not primarily about NDAs and trade secrets, even though an NDA was mentioned. This was primarily about non-disparagement and the ability to criticize OpenAI, see the linked article https://www.vox.com/future-perfect/2024/5/17/24158478/openai...
Second, this is not even about the existence of the non-disparagement agreement itself, it's about Altman claiming ignorance about its existence (see: the article again)
Yeah I have signed those before but also rejected them other times, I agree that it’s not exactly warm and fuzzy to have someone want you to sign on a line about your ability to speak freely!
No defense of the act here, but how can someone take a publication seriously that states an opinion in the first sentence, and in the 2nd paragraph uses a clickbait modifier like "draconian"?
It does after all refer to a contract people were apparently pressured into signing when quitting (and not when signing up), that apparently offered no new consideration, and required them to never say anything bad about OpenAI, for life, on penalty of losing remuneration another signed contract entitled them to.
It's journalism. But even the lawyers are saying "egregious". Which is really really not good in this context.
"Draconian" fits, considering how it is normally used.
Egregious is a much better word that doesn't intentionally sound alarming.
The only new information here is that his signature was on these documents. This article took 18 paragraphs to say that.
Real journalism that wasn't influenced by SEO or ad views should probably be one sentence with the new information and a link to a previous article filling in the missing context, for those who happen not to have it.
Are you basing this off of the assumption that every single person reading this has the same relationship to specific words as you?
Egregious and draconian create the same “idea” in my head. There is literally no way to write that takes into account each persons relationship with words.
Mmm, though I think these are two angular perspectives on the same idea, more than they are the same idea.
I feel "egregious" makes more sense for the idea of these bullshit contracts in the general case -- the idea that a lawyer wrote them or signed off on them.
"Draconian" is a really good fit for the experience and outcome for the individual, and for the process by which an individual found themselves signing.
Either way, I don't think there's any reason to minimise what is happening for the employee, or what it means in terms of characterising the conduct and culture of the ultimate unicorn. If stuff like this is allowed to slide, it's not good.
Now consider the thing that is happening: someone is being essentially pressured to agree a contract forcing them to stay functionally silent about anything bad they experienced or saw at OpenAI, on pain of losing things they've already earned.
For that individual's experience, "draconian" -- harsh, unusually severe, repressive -- seems like it would be a good fit, right?
And as it is not in the title or the subhead, it's not really being deployed to be clickbait, is it? The title is even more damning
(As an aside, why should this article not intentionally sound alarming? This whole situation is a pretty damning indictment of one of the most influential startups on earth right now. IMO it is alarming.)
> The only new information here is that his signature was on these documents.
This is what's called "confirming the story". News organisations do this all the time -- corroborate their competitors' stories -- and it is absolutely news.
In this case it's a really big bit of news, to people outside the tech world: it means there will likely prove to be concrete, potentially-lawsuit-bound corroboration that Sam Altman is not "consistently candid".
No evidence in the article that Altman knew about this, and intentionally clickbaity and misleading title.
Quoting: "there are only two plausible reasons for Altman's alleged lack of knowledge: either he didn't fully read the employment contracts he was signing or he was lying." No evidence was offered that he read the contracts.
You are technically true. However, if you sign contracts you haven't read, you would have a hard time defending your innocence in court. This applies even more for higher positions like CEO.
Yes, legally responsible. There's no legal violation here, though. The article made a specific claim: that Altman was aware of something. It provided no evidence that he was.
If it accused him of some criminal misdeed, then it wouldn't matter if he was aware of it, as long as he signed the illegal order. That would be a different story, but clearly in this case it doesn't apply.
> No evidence was offered that he read the contracts
This excuse only works for people who do not get paid for doing exactly this: the power to sign contracts, exit papers, stuff like that on behalf of the company, and, therefore, be aware of their contents. "I didn't go through the contract before agreeing to its content by signing it" is a damning verdict for someone in this position. You don't have to prove he read it: he signed, therefore he's responsible.
I have zero respect for companies with such predatory behaviour. Not a knee jerk reaction, but threatening employees "diplomatically" with such draconian agreements and even sometimes non-competes should actually be grounds for mass resignation. I worked in a similar company and till date have traumatic memories from that era.
For this and many other reasons, I would never ever use a proprietary AI model in any of my AI projects. "People first" should be the guiding principle for developers.
62 comments
[ 3.1 ms ] story [ 144 ms ] threadOr maybe you're not, Hacker News has evolved away from its mission and audience and I'm in the wrong place.
These people have weightier issues like the ethics of the hypothetical technological future on their capacious minds, and mustn't be dragged down to your mucky little level of discussing the ethics of the very real business present.
Seriously: don't bother them with these little questions of the rights of actually enumerable humans to whom they have contractual responsibilities! They're busy thinking in great depth about how their product might affect the rights of all the humans they haven't yet sold it to.
This article is literally just summarizing the Vox article with no new information or insight. And the HN submission for the Vox article had 1784 points and 544 comments.
The article is doing a lot of click-bait stuff here. The quote, without the source, is shown at the top of the article, near Sam Altman's face and a title referencing Sam Altman. However, the quote is not from Sam Altman, which makes its placement a bit disingenuous.
I am not invested in defending Altman, but this type of journalism is trash clickbait and a far bigger issue than a tech company threatening employees' vested equity unless they comply in some sort of non-NDA NDA scheme.
They make a case that Altman signed documents where it's quite clear the intention is to be able to claw back vested equity. And that quote in the headline isn't meant to try and attribute it directly to Altman -- but to show just how well-known the policy was.
I think the case study still raises questions about whether we should require specific licensing or contractual terms for what these organizations produce. For instance, that they dual-licensed all their models or (if safety is a worry) always shared them with research institutions. That by itself might have given us GPT3 or GPT4 with the Llama- and Falcon-type communities building on them.
I was a vocal defender of Sam in the immediate aftermath but these facts are definitely concerning. OpenAI is moving fast with a critical and dangerous technology. If their culture isn’t profoundly ethical they shouldn’t be trusted with the privilege and responsibility of what they are up to.
First, this was not primarily about NDAs and trade secrets, even though an NDA was mentioned. This was primarily about non-disparagement and the ability to criticize OpenAI, see the linked article https://www.vox.com/future-perfect/2024/5/17/24158478/openai...
Second, this is not even about the existence of the non-disparagement agreement itself, it's about Altman claiming ignorance about its existence (see: the article again)
Onerous? Extreme? Oppressive? Bullying?
It does after all refer to a contract people were apparently pressured into signing when quitting (and not when signing up), that apparently offered no new consideration, and required them to never say anything bad about OpenAI, for life, on penalty of losing remuneration another signed contract entitled them to.
It's journalism. But even the lawyers are saying "egregious". Which is really really not good in this context.
"Draconian" fits, considering how it is normally used.
The only new information here is that his signature was on these documents. This article took 18 paragraphs to say that.
Real journalism that wasn't influenced by SEO or ad views should probably be one sentence with the new information and a link to a previous article filling in the missing context, for those who happen not to have it.
Egregious and draconian create the same “idea” in my head. There is literally no way to write that takes into account each persons relationship with words.
I feel "egregious" makes more sense for the idea of these bullshit contracts in the general case -- the idea that a lawyer wrote them or signed off on them.
"Draconian" is a really good fit for the experience and outcome for the individual, and for the process by which an individual found themselves signing.
Either way, I don't think there's any reason to minimise what is happening for the employee, or what it means in terms of characterising the conduct and culture of the ultimate unicorn. If stuff like this is allowed to slide, it's not good.
https://www.merriam-webster.com/dictionary/draconian
https://en.wikipedia.org/wiki/Draconian
https://www.vocabulary.com/dictionary/Draconian
Now consider the thing that is happening: someone is being essentially pressured to agree a contract forcing them to stay functionally silent about anything bad they experienced or saw at OpenAI, on pain of losing things they've already earned.
For that individual's experience, "draconian" -- harsh, unusually severe, repressive -- seems like it would be a good fit, right?
And as it is not in the title or the subhead, it's not really being deployed to be clickbait, is it? The title is even more damning
(As an aside, why should this article not intentionally sound alarming? This whole situation is a pretty damning indictment of one of the most influential startups on earth right now. IMO it is alarming.)
> The only new information here is that his signature was on these documents.
This is what's called "confirming the story". News organisations do this all the time -- corroborate their competitors' stories -- and it is absolutely news.
In this case it's a really big bit of news, to people outside the tech world: it means there will likely prove to be concrete, potentially-lawsuit-bound corroboration that Sam Altman is not "consistently candid".
Quoting: "there are only two plausible reasons for Altman's alleged lack of knowledge: either he didn't fully read the employment contracts he was signing or he was lying." No evidence was offered that he read the contracts.
If it accused him of some criminal misdeed, then it wouldn't matter if he was aware of it, as long as he signed the illegal order. That would be a different story, but clearly in this case it doesn't apply.
This excuse only works for people who do not get paid for doing exactly this: the power to sign contracts, exit papers, stuff like that on behalf of the company, and, therefore, be aware of their contents. "I didn't go through the contract before agreeing to its content by signing it" is a damning verdict for someone in this position. You don't have to prove he read it: he signed, therefore he's responsible.
For this and many other reasons, I would never ever use a proprietary AI model in any of my AI projects. "People first" should be the guiding principle for developers.
At some point if it true its true.