I’m struggling to follow your argument here. What you describe as the situation at MIRI is basically standard risk management approach—if two people create a risk to MIRI’s financial security processes by falling in love, you make sure that neither signs off on risk taken by the other.
But in this thread you are responding with strong disagreement to a comment which says “if this relationship [between SBF and Caroline] were known to be hidden from investors and other stakeholders, should this not have raised red flags?”. You said “who EAs are fucking is none of EA’s business”, amongst other comments of a similar tone.
I don’t understand what exactly you disagree with if you agree SBF and Caroline should have disclosed their relationship so that proper steps could be taken to de-risk their interactions (as would happen at MIRI). It seems that you do agree it matters who EAs are fucking in contexts like this? And therefore that it is relevant to know whether Will MacAskill knew about the undisclosed relationship?
You could plausibly claim it gets disclosed to Sequoia Capital, if SC has shown themselves worthy of being trusted with information like that and responding to it in a sensible fashion eg with more thorough audits. Disclosing to FTX Future Fund seems like a much weirder case, unless FTX Future Fund is auditing FTX’s books well enough that they’d have any hope of detecting fraud—otherwise, what is FTXFF supposed to do with that information?
EA generally thinking that it has a right to know who its celebrity donors are fucking strikes me as incredibly unhealthy.
I think we might be straying from the main point a bit; nobody is proposing a general right to peer into EA sex lives, and I agree that would be unhealthy.
There are some relatively straightforward financial risk management principles which msinstream orgs have been successfully using for decades. You seem to believe one of the pillars of these principles—surfacing risk due to romantic entanglements between parties—shouldn’t apply to EA, and instead some sort of ‘commonsense’ approach should prevail instead (inverted commas because I think the standard way is basically common sense too).
But I don’t understand where your confidence that you’re right here is coming from—EA leadership has just materially failed to protect EA membership from bad actor risk stemming at least in part from a hidden conflict of interest due to a romantic entanglement. EA leadership has been given an opportunity to run risk management their way, and the result is that EA is now associated with the biggest crypto fraud in history. Surely the Bayesian update here is that there are strong reasons to believe mainstream finance had it approximately right?
Rereading the above, I think I might just be unproductively repeating myself at this point, so I’ll duck out of the discussion. I appreciated the respectful back-and-forth, especially considering parts of what I was saying were (unavoidably) pretty close to personal attacks on you and the EA leadership more broadly. Hope you had a pleasant evening too!
My (possibly wrong) understanding of what Eliezer is saying:
FTX ought to have responded internally to the conflict of interest, but they had no obligation to disclose it externally (to Future Fund staff or wider EA community).
The failure in FTX was that they did not implement the right internal controls—not that the relationship was “hidden from investors and other stakeholders.”
If EA leadership and FTX investors made a mistake, it was failing to ensure that FTX had implemented the right internal controls—not failing to know about the relationship.
I couldn’t quite bottom out exactly what EY was saying, but I’m pretty sure it wasn’t that. On your interpretation, EY said, “who EAs are fucking is none of [wider] EA’s business [except people who are directly affected by the COI]”. But he goes on to clarify “There are very limited exceptions to this rule like ‘maybe don’t fuck your direct report’ ”. If that’s an exception to the rule of EAs fucking being only of interest to directly affected parties, then it mean EY thinks an EA having sex with a subordinate should be broadcast to the entire community. That’s a very strict standard (although I guess not crazy—just odd that EY was presenting it as a more relaxed / less prurient standard than conventional financial risk management).
It also doesn’t address my core objection, which is that EA leadership failed very badly to implement proper financial risk management processes. Generally my point was that EA leadership should be epistemically humble now and just implement the risk management processes that work for banks, rather than tinkering around and introducing their own version of these systems. Regardless of what EY meant, unless he meant ‘We should hire in PWC to implement the same financial controls as every Fortune company’ then he is making exactly the same mistake EA leadership made with FTX—assuming that they could create better risk management from first principles than the mainstream system could from actual experience
By the way, I disagree with the objective position here too. Every FTX investor needed to know about the COI and the management strategy FTX adopted in order to assess their risk exposure. This would be the standard at a conventional company (if the company knew about such a blatant COI from their CEO and didn’t tell investors at a conventional company then their risk officers would potentially be liable for the fraud too, iirc)
I’m struggling to follow your argument here. What you describe as the situation at MIRI is basically standard risk management approach—if two people create a risk to MIRI’s financial security processes by falling in love, you make sure that neither signs off on risk taken by the other.
But in this thread you are responding with strong disagreement to a comment which says “if this relationship [between SBF and Caroline] were known to be hidden from investors and other stakeholders, should this not have raised red flags?”. You said “who EAs are fucking is none of EA’s business”, amongst other comments of a similar tone.
I don’t understand what exactly you disagree with if you agree SBF and Caroline should have disclosed their relationship so that proper steps could be taken to de-risk their interactions (as would happen at MIRI). It seems that you do agree it matters who EAs are fucking in contexts like this? And therefore that it is relevant to know whether Will MacAskill knew about the undisclosed relationship?
You could plausibly claim it gets disclosed to Sequoia Capital, if SC has shown themselves worthy of being trusted with information like that and responding to it in a sensible fashion eg with more thorough audits. Disclosing to FTX Future Fund seems like a much weirder case, unless FTX Future Fund is auditing FTX’s books well enough that they’d have any hope of detecting fraud—otherwise, what is FTXFF supposed to do with that information?
EA generally thinking that it has a right to know who its celebrity donors are fucking strikes me as incredibly unhealthy.
I think we might be straying from the main point a bit; nobody is proposing a general right to peer into EA sex lives, and I agree that would be unhealthy.
There are some relatively straightforward financial risk management principles which msinstream orgs have been successfully using for decades. You seem to believe one of the pillars of these principles—surfacing risk due to romantic entanglements between parties—shouldn’t apply to EA, and instead some sort of ‘commonsense’ approach should prevail instead (inverted commas because I think the standard way is basically common sense too).
But I don’t understand where your confidence that you’re right here is coming from—EA leadership has just materially failed to protect EA membership from bad actor risk stemming at least in part from a hidden conflict of interest due to a romantic entanglement. EA leadership has been given an opportunity to run risk management their way, and the result is that EA is now associated with the biggest crypto fraud in history. Surely the Bayesian update here is that there are strong reasons to believe mainstream finance had it approximately right?
Rereading the above, I think I might just be unproductively repeating myself at this point, so I’ll duck out of the discussion. I appreciated the respectful back-and-forth, especially considering parts of what I was saying were (unavoidably) pretty close to personal attacks on you and the EA leadership more broadly. Hope you had a pleasant evening too!
My (possibly wrong) understanding of what Eliezer is saying:
FTX ought to have responded internally to the conflict of interest, but they had no obligation to disclose it externally (to Future Fund staff or wider EA community).
The failure in FTX was that they did not implement the right internal controls—not that the relationship was “hidden from investors and other stakeholders.”
If EA leadership and FTX investors made a mistake, it was failing to ensure that FTX had implemented the right internal controls—not failing to know about the relationship.
I couldn’t quite bottom out exactly what EY was saying, but I’m pretty sure it wasn’t that. On your interpretation, EY said, “who EAs are fucking is none of [wider] EA’s business [except people who are directly affected by the COI]”. But he goes on to clarify “There are very limited exceptions to this rule like ‘maybe don’t fuck your direct report’ ”. If that’s an exception to the rule of EAs fucking being only of interest to directly affected parties, then it mean EY thinks an EA having sex with a subordinate should be broadcast to the entire community. That’s a very strict standard (although I guess not crazy—just odd that EY was presenting it as a more relaxed / less prurient standard than conventional financial risk management).
It also doesn’t address my core objection, which is that EA leadership failed very badly to implement proper financial risk management processes. Generally my point was that EA leadership should be epistemically humble now and just implement the risk management processes that work for banks, rather than tinkering around and introducing their own version of these systems. Regardless of what EY meant, unless he meant ‘We should hire in PWC to implement the same financial controls as every Fortune company’ then he is making exactly the same mistake EA leadership made with FTX—assuming that they could create better risk management from first principles than the mainstream system could from actual experience
By the way, I disagree with the objective position here too. Every FTX investor needed to know about the COI and the management strategy FTX adopted in order to assess their risk exposure. This would be the standard at a conventional company (if the company knew about such a blatant COI from their CEO and didn’t tell investors at a conventional company then their risk officers would potentially be liable for the fraud too, iirc)