3 ms·

Not really. But Helen Toner has been tweeting a little "To be clear: our decision was about the board's ability to effectively supervise the company, which was

by sigmar 3y ago

Not really. But Helen Toner has been tweeting a little "To be clear: our decision was about the board's ability to effectively supervise the company, which was our role and responsibility. Though there has been speculation, we were not motivated by a desire to slow down OpenAI’s work." https://twitter.com/hlntnr/status/1730034020140912901 https://twitter.com/hlntnr/status/1730034020140912901

dmix 3y ago
> Though there has been speculation, we were not motivated by a desire to slow down OpenAI’s work. Strange when their choice of interim-CEO was someone who explicitly stated he wants to see the pace of AI development go from a 10 to a 2 [1] and she regularly speaks at EA conferences where that's a major theme. This is probably double speak for she want's to not "slow down OpenAI's work" on AI safety but probably would have kneecapped the "early" release of ChatGPT's (as she claim they should have waited much longer in her paper) and similar things. [1] https://twitter.com/eshear/status/1703178063306203397 https://twitter.com/eshear/status/1703178063306203397
- 93po 3y ago
  They didn’t want to slow work, just the work on the stuff they didn’t like
  - dmix 3y ago
    Yes exactly.
- 0xDEAFBEAD 3y ago
  My current guess is that Helen and Sam had disagreements, and that caused Sam to be less-than-candid about the state of OpenAI's tech, and that was the straw that broke the camel's back for Helen. A disagreement within the board is one thing, but if the CEO undermines the ability of the board to provide oversight, that sort of crosses a line. Alternatively, maybe it became clear to the board that Sam was trying to either bully them into becoming loyalists, or replace them with loyalists. As a board member, if the CEO is badgering me into becoming a yes-man and threatening to kick me off if I don't comply, I can't exactly say that I'm able to supervise the company effectively, can I? See https://www.lesswrong.com/posts/KXHMCH7wCxrvKsJyn/openai-facts-from-a-weekend?commentId=3cj6qhSRt4HoBLpC7 https://www.lesswrong.com/posts/KXHMCH7wCxrvKsJyn/openai-fac...
- vasco 3y ago
  The way EA does donations is "I'll take a premise I believe in and will stretch the argument until it makes no sense". This is how they end up thinking that a massage for an AI researcher is money better spent than on hungry Yemeni children for example. Once you view it like this, I wouldn't put it past them to blatantly lie. Looking at the facts as you say, they tried to replace a guy that is moving ahead with a guy that wants to slow it down to a crawl, that's pretty much all we need to know.
  - jacquesm 3y ago
    Essentially EA is a stretchable concept that allows adherents to act out their every fantasy with complete abandon whilst protecting their sensitive sense of self. It redefines their side to always be the good side, no matter what they get up to.
  - concordDance 3y ago
    This is your daily reminder that most EAs will just donate to the top GiveWell charity even though people will talk a lot about the interesting edge cases.
ethbr1 3y ago
From everything I've read, safety still feels like a red herring. It just doesn't fit with everyone's behavior -- that's something that would have been talked about loudly. Altman lying to the board, especially in pursuit of board control, fits more cleanly with everyone's actions (and reluctance to talk about what exactly precipitated this). - Altman tries to effect board control - Confidant tells other board members (Ilya?) - Board asks Altman if he tried - Altman lies - Board fires Altman Fits much more cleanly with the lack of information, and how no one (on any side!) seems overly eager to speak to specifics. Why jump to AGI as an explanation, when standard human drama will suffice?
- vikramkr 3y ago
  But then that doesn't square with board refuses to tell employees or Microsoft or the public that altman committed malfeasance or provide examples. That would be pretty cut and dry and msft wouldn't be willing to acquihire the entire company we with altman as CEO if there was a valid reason like that.
  - ethbr1 3y ago
    That's not typically the sort of dirty laundry that's aired to employees or the public, by professionals, which all of the former board were. And absolutely Microsoft would have behaved as it did -- it doesn't really care about OpenAI-the-company; it cares about OpenAI-the-tech. Also, see essentially the same said by Microsoft's president: https://www.bbc.com/news/technology-67578656 https://www.bbc.com/news/technology-67578656
adastra22 3y ago
> our decision was about the board's ability to effectively supervise the company Sounds like confirmation of the theory that it was Sam trying to get Toner off the board which precipitated this.

by sigmar 3y ago

dmix 3y ago

> Though there has been speculation, we were not motivated by a desire to slow down OpenAI’s work. Strange when their choice of interim-CEO was someone who explicitly stated he wants to see the pace of AI development go from a 10 to a 2 [1] and she regularly speaks at EA conferences where that's a major theme. This is probably double speak for she want's to not "slow down OpenAI's work" on AI safety but probably would have kneecapped the "early" release of ChatGPT's (as she claim they should have waited much longer in her paper) and similar things. [1] https://twitter.com/eshear/status/1703178063306203397 https://twitter.com/eshear/status/1703178063306203397

93po 3y ago
They didn’t want to slow work, just the work on the stuff they didn’t like
- dmix 3y ago
  Yes exactly.
0xDEAFBEAD 3y ago
My current guess is that Helen and Sam had disagreements, and that caused Sam to be less-than-candid about the state of OpenAI's tech, and that was the straw that broke the camel's back for Helen. A disagreement within the board is one thing, but if the CEO undermines the ability of the board to provide oversight, that sort of crosses a line. Alternatively, maybe it became clear to the board that Sam was trying to either bully them into becoming loyalists, or replace them with loyalists. As a board member, if the CEO is badgering me into becoming a yes-man and threatening to kick me off if I don't comply, I can't exactly say that I'm able to supervise the company effectively, can I? See https://www.lesswrong.com/posts/KXHMCH7wCxrvKsJyn/openai-facts-from-a-weekend?commentId=3cj6qhSRt4HoBLpC7 https://www.lesswrong.com/posts/KXHMCH7wCxrvKsJyn/openai-fac...
vasco 3y ago
The way EA does donations is "I'll take a premise I believe in and will stretch the argument until it makes no sense". This is how they end up thinking that a massage for an AI researcher is money better spent than on hungry Yemeni children for example. Once you view it like this, I wouldn't put it past them to blatantly lie. Looking at the facts as you say, they tried to replace a guy that is moving ahead with a guy that wants to slow it down to a crawl, that's pretty much all we need to know.
- jacquesm 3y ago
  Essentially EA is a stretchable concept that allows adherents to act out their every fantasy with complete abandon whilst protecting their sensitive sense of self. It redefines their side to always be the good side, no matter what they get up to.
- concordDance 3y ago
  This is your daily reminder that most EAs will just donate to the top GiveWell charity even though people will talk a lot about the interesting edge cases.

ethbr1 3y ago

From everything I've read, safety still feels like a red herring. It just doesn't fit with everyone's behavior -- that's something that would have been talked about loudly. Altman lying to the board, especially in pursuit of board control, fits more cleanly with everyone's actions (and reluctance to talk about what exactly precipitated this). - Altman tries to effect board control - Confidant tells other board members (Ilya?) - Board asks Altman if he tried - Altman lies - Board fires Altman Fits much more cleanly with the lack of information, and how no one (on any side!) seems overly eager to speak to specifics. Why jump to AGI as an explanation, when standard human drama will suffice?

vikramkr 3y ago
But then that doesn't square with board refuses to tell employees or Microsoft or the public that altman committed malfeasance or provide examples. That would be pretty cut and dry and msft wouldn't be willing to acquihire the entire company we with altman as CEO if there was a valid reason like that.
- ethbr1 3y ago
  That's not typically the sort of dirty laundry that's aired to employees or the public, by professionals, which all of the former board were. And absolutely Microsoft would have behaved as it did -- it doesn't really care about OpenAI-the-company; it cares about OpenAI-the-tech. Also, see essentially the same said by Microsoft's president: https://www.bbc.com/news/technology-67578656 https://www.bbc.com/news/technology-67578656

adastra22 3y ago

> our decision was about the board's ability to effectively supervise the company Sounds like confirmation of the theory that it was Sam trying to get Toner off the board which precipitated this.