I was confused about the counts, but I guess this makes sense if Helen cannot vote on her own removal. Then it’s Altman/Brockman/Sutskever v Tasha/D’Angelo.
Pretty interesting that Sutskever/Tasha/D’Angelo would be willing to fire Altman just to prevent Helen from going. They instead could have negotiated someone to replace her. Wouldn’t you just remove Altman from the Board, or maybe remove Brockman? Why would they be willing to decapitate the company in order to retain Helen?
They instead could have negotiated someone to replace her.
Why do they have to negotiate? They didn’t want her gone, he did. Why didn’t Altman negotiate a replacement for her, if he was so very upset about the damages she had supposedly done OA...?
“I understand we’ve struggled to agree on any replacement directors since I kicked Hoffman out, and you’d worry even more about safety remaining a priority if she resigns. I totally get it. So that’s not an obstacle, I’ll agree to let Toner nominate her own replacement—just so long as she leaves soon.”
When you understand why Altman would not negotiate that, you understand why the board could not negotiate that.
I was confused about the counts, but I guess this makes sense if Helen cannot vote on her own removal. Then it’s Altman/Brockman/Sutskever v Tasha/D’Angelo.
Recusal or not, Altman didn’t want to bring it to something as overt as a vote expelling her. Power wants to conceal itself and deny the coup. The point here of the CSET paper pretext is to gain leverage and break the tie any way possible so it doesn’t look bad or traceable to Altman: that’s why this leaking is bad for Altman, it shows him at his least fuzzy and PR-friendly. He could, obviously, have leaked the Toner paper at any time to a friendly journalist to manufacture a crisis and force the issue, but that was not—as far as he knew then—yet a tactic he needed to resort to. However, the clock was ticking, and the board surely knew that the issue could be forced at any time of Altman’s choosing.
If he had outright naked control of the board, he would scarcely need to remove her nor would they be deadlocked over the new directors; but by organizing a ‘consensus’ among the OA executives (like Jakub Pachocki?) about Toner committing an unforgivable sin that can be rectified only by stepping down, and by lobbying in the background and calling in favors, and arguing for her recusal, Altman sets the stage for wearing down Toner (note what they did to Ilya Sutskever & how the Altman faction continues to tout Sutskever’s flip without mentioning the how) and Toner either resigning voluntarily or, in the worst case, being fired. It doesn’t matter which tactic succeeds, a good startup CEO never neglects a trick, and Altman knows them all—it’s not for nothing that Paul Graham keeps describing Altman as the most brutally effective corporate fighter he’s ever known and describes with awe how eg he manipulated Graham into appointing him president of YC, and eventually Graham had to fire him from YC for reasons already being foreshadowed in 2016. (Note how thoroughly and misogynistically Toner has been vilified on social media by OAer proxies, who, despite leaking to the media like Niagara Falls, somehow never felt this part about Altman organizing her removal to be worth mentioning; every tactic has been employed in the fight so far: they even have law enforcement pals opening an ‘investigation’. Needless to say, there’s zero chance of it going anywhere, it’s just power struggles, similar to the earlier threats to sue the directors personally.) Note: if all this can go down in like 3 days with Altman outside the building and formally fired and much of the staff gone on vacation, imagine what he could have done with 3 months and CEO access/resources/credibility and all the OAers back?
The board was tolerating all this up to the point where firing Toner came up, because it seemed like Sam was just aw-shucks-being-Sam—being an overeager go-getter was the whole point of the CEO, wasn’t it? it wasn’t like he was trying to launch a coup or anything, surely not—but when he opened fire on Toner for such an incredibly flimsy pretext without, say, proposing to appoint a specific known safety person to replace Toner and maintain the status quo, suddenly, everything changed. (What do you think a treacherous turn looks like IRL? It looks like that.) The world in which Altman is just an overeager commercializer who otherwise agrees with the board and there’s just been a bunch of misunderstandings and ordinary conflicts is a different world from the world in which he doesn’t care about safety unless it’s convenient & regularly deceives and manipulates & has been maneuvering the entire time to irrevocably take over the board to remove his last check. And if you realize you have been living in the second world and that you have the slimmest possible majority, which will crack as soon as Altman realizes he’s overplayed his hand and moves overtly to deploy his full arsenal before he forces a vote...
So Altman appears to have made two key mistakes here, because he was so personally overstretched and 2023 has been such a year: first, taking Sutskever for granted. (WSJ: “Altman this weekend was furious with himself for not having ensured the board stayed loyal to him and regretted not spending more time managing its various factions, people familiar with his thinking said.”) Then second, making his move with such a flimsy pretext that it snapped the suspension of disbelief of the safety faction. Had he realized Sutskever was a swing vote, he would have worked on him much harder and waited for better opportunities to move against Toner or McCauley. Well, live and learn—he’s a smart guy; he won’t make the same mistakes twice with the next OA board.
(If you find any of this confusing or surprising, I strongly suggest you read up more on how corporate infighting works. You may not be interested in corporate governance or power politics, but they are now interested in you, and this literature is only going to get more relevant. Some LWer-friendly starting points here would be Bad Blood on narcissistElizabeth Holmes, Steve Jobs—Altman’s biggest hero—and his ouster, the D&D coup, the classic Barbarians at the Gate, the many contemporary instances covered in Matt Levine’s newsletter like the Papa Johns coup or most recently, Sculptor, The Gervais Principle, the second half of Breaking Bad, Zvi’s many relevant essays on moral mazes/simulacra levels/corporate dynamics from his perspective as a hedge fund guy, and especially the in-depth reporting on how Harvey Weinstein covered everything up for so long which pairs well with Bad Blood.)
I… still don’t understand why the board didn’t say anything? I really feel like a lot of things would have flipped if they had just talked openly to anyone, or taken advice from anyone. Like, I don’t think it would have made them global heroes, and a lot of people would have been angry with them, but every time any plausible story about what happened came out, there was IMO a visible shift in public opinion, including on HN, and the board confirming any story or giving any more detail would have been huge. Instead they apparently “cited legal reasons” for not talking, which seems crazy to me.
I can imagine it being the case that their ability to reveal this information is their main source of leverage (over e.g. who replaces them on the board).
My favorite low-probability theory is that he had blackmail material on one of the board members[1], who initially decided after much deliberation to go forwards despite the blackmail, and then when they realized they got outplayed by Sam not using the blackmail material, backpeddled and refused to dox themselves. And the other 2-3 didn’t know what to do afterwards, because their entire strategy was predicated on optics management around said blackmail + blackmail material.
It would be sheer insanity to have a rule that you can’t vote on your own removal, I would think, or else a tied board will definitely shrink right away.
Wait, simple majority is an insane place to put the threshold for removal in the first place. Majoritarian shrinking is still basically inevitable if the threshold for removal is 50%, it should be a higher than that, maybe 62%.
And generally, if 50% of a group thinks A and 50% thinks ¬A, that tells you that the group is not ready to make a decision about A.
I was confused about the counts, but I guess this makes sense if Helen cannot vote on her own removal. Then it’s Altman/Brockman/Sutskever v Tasha/D’Angelo.
Pretty interesting that Sutskever/Tasha/D’Angelo would be willing to fire Altman just to prevent Helen from going. They instead could have negotiated someone to replace her. Wouldn’t you just remove Altman from the Board, or maybe remove Brockman? Why would they be willing to decapitate the company in order to retain Helen?
Why do they have to negotiate? They didn’t want her gone, he did. Why didn’t Altman negotiate a replacement for her, if he was so very upset about the damages she had supposedly done OA...?
“I understand we’ve struggled to agree on any replacement directors since I kicked Hoffman out, and you’d worry even more about safety remaining a priority if she resigns. I totally get it. So that’s not an obstacle, I’ll agree to let Toner nominate her own replacement—just so long as she leaves soon.”
When you understand why Altman would not negotiate that, you understand why the board could not negotiate that.
Recusal or not, Altman didn’t want to bring it to something as overt as a vote expelling her. Power wants to conceal itself and deny the coup. The point here of the CSET paper pretext is to gain leverage and break the tie any way possible so it doesn’t look bad or traceable to Altman: that’s why this leaking is bad for Altman, it shows him at his least fuzzy and PR-friendly. He could, obviously, have leaked the Toner paper at any time to a friendly journalist to manufacture a crisis and force the issue, but that was not—as far as he knew then—yet a tactic he needed to resort to. However, the clock was ticking, and the board surely knew that the issue could be forced at any time of Altman’s choosing.
If he had outright naked control of the board, he would scarcely need to remove her nor would they be deadlocked over the new directors; but by organizing a ‘consensus’ among the OA executives (like Jakub Pachocki?) about Toner committing an unforgivable sin that can be rectified only by stepping down, and by lobbying in the background and calling in favors, and arguing for her recusal, Altman sets the stage for wearing down Toner (note what they did to Ilya Sutskever & how the Altman faction continues to tout Sutskever’s flip without mentioning the how) and Toner either resigning voluntarily or, in the worst case, being fired. It doesn’t matter which tactic succeeds, a good startup CEO never neglects a trick, and Altman knows them all—it’s not for nothing that Paul Graham keeps describing Altman as the most brutally effective corporate fighter he’s ever known and describes with awe how eg he manipulated Graham into appointing him president of YC, and eventually Graham had to fire him from YC for reasons already being foreshadowed in 2016. (Note how thoroughly and misogynistically Toner has been vilified on social media by OAer proxies, who, despite leaking to the media like Niagara Falls, somehow never felt this part about Altman organizing her removal to be worth mentioning; every tactic has been employed in the fight so far: they even have law enforcement pals opening an ‘investigation’. Needless to say, there’s zero chance of it going anywhere, it’s just power struggles, similar to the earlier threats to sue the directors personally.) Note: if all this can go down in like 3 days with Altman outside the building and formally fired and much of the staff gone on vacation, imagine what he could have done with 3 months and CEO access/resources/credibility and all the OAers back?
The board was tolerating all this up to the point where firing Toner came up, because it seemed like Sam was just aw-shucks-being-Sam—being an overeager go-getter was the whole point of the CEO, wasn’t it? it wasn’t like he was trying to launch a coup or anything, surely not—but when he opened fire on Toner for such an incredibly flimsy pretext without, say, proposing to appoint a specific known safety person to replace Toner and maintain the status quo, suddenly, everything changed. (What do you think a treacherous turn looks like IRL? It looks like that.) The world in which Altman is just an overeager commercializer who otherwise agrees with the board and there’s just been a bunch of misunderstandings and ordinary conflicts is a different world from the world in which he doesn’t care about safety unless it’s convenient & regularly deceives and manipulates & has been maneuvering the entire time to irrevocably take over the board to remove his last check. And if you realize you have been living in the second world and that you have the slimmest possible majority, which will crack as soon as Altman realizes he’s overplayed his hand and moves overtly to deploy his full arsenal before he forces a vote...
So Altman appears to have made two key mistakes here, because he was so personally overstretched and 2023 has been such a year: first, taking Sutskever for granted. (WSJ: “Altman this weekend was furious with himself for not having ensured the board stayed loyal to him and regretted not spending more time managing its various factions, people familiar with his thinking said.”) Then second, making his move with such a flimsy pretext that it snapped the suspension of disbelief of the safety faction. Had he realized Sutskever was a swing vote, he would have worked on him much harder and waited for better opportunities to move against Toner or McCauley. Well, live and learn—he’s a smart guy; he won’t make the same mistakes twice with the next OA board.
(If you find any of this confusing or surprising, I strongly suggest you read up more on how corporate infighting works. You may not be interested in corporate governance or power politics, but they are now interested in you, and this literature is only going to get more relevant. Some LWer-friendly starting points here would be Bad Blood on narcissist Elizabeth Holmes, Steve Jobs—Altman’s biggest hero—and his ouster, the D&D coup, the classic Barbarians at the Gate, the many contemporary instances covered in Matt Levine’s newsletter like the Papa Johns coup or most recently, Sculptor, The Gervais Principle, the second half of Breaking Bad, Zvi’s many relevant essays on moral mazes/simulacra levels/corporate dynamics from his perspective as a hedge fund guy, and especially the in-depth reporting on how Harvey Weinstein covered everything up for so long which pairs well with Bad Blood.)
I… still don’t understand why the board didn’t say anything? I really feel like a lot of things would have flipped if they had just talked openly to anyone, or taken advice from anyone. Like, I don’t think it would have made them global heroes, and a lot of people would have been angry with them, but every time any plausible story about what happened came out, there was IMO a visible shift in public opinion, including on HN, and the board confirming any story or giving any more detail would have been huge. Instead they apparently “cited legal reasons” for not talking, which seems crazy to me.
I can imagine it being the case that their ability to reveal this information is their main source of leverage (over e.g. who replaces them on the board).
My favorite low-probability theory is that he had blackmail material on one of the board members[1], who initially decided after much deliberation to go forwards despite the blackmail, and then when they realized they got outplayed by Sam not using the blackmail material, backpeddled and refused to dox themselves. And the other 2-3 didn’t know what to do afterwards, because their entire strategy was predicated on optics management around said blackmail + blackmail material.
Like something actually really bad.
It would be sheer insanity to have a rule that you can’t vote on your own removal, I would think, or else a tied board will definitely shrink right away.
Wait, simple majority is an insane place to put the threshold for removal in the first place. Majoritarian shrinking is still basically inevitable if the threshold for removal is 50%, it should be a higher than that, maybe 62%.
And generally, if 50% of a group thinks A and 50% thinks ¬A, that tells you that the group is not ready to make a decision about A.
It is not clear, in the non—profit structure of a board, that Helen cannot vote on her own removal.
The vote to remove Sam may have been some trickery around holding a quorum meeting without notifying Sam or Greg.
I think it was most likely unanimous among the remaining 4, otherwise one of the dissenters would’ve spoken out by now.