If a global health organization made a mistake in judgment that caused [its] effectiveness to permanently decline by (say) 30%, and it was no longer effective in comparison to alternatives we could counterfactually fund, I suspect very few of us would support continuing to fund it. I would find it potentially concerning, from a standpoint of impartiality, if we do not apply the same standard to leaders. After all, we didn’t protect the hypothetical global health organization’s beneficiaries merely out of a sense of fairness.
I see the argument that applying such a standard to leaders could discourage them from making EV-positive bets. However, experiencing an adverse outcome on most EV-positive bets won’t materially impact a leader’s long-term future effectiveness. Moreover, it could be difficult to evaluate leaders from a 100% ex ante perspective. There’s a risk of evaluating successful bets by their outcome (because outsiders may not understand that there was a significant bet + there is low incentive to evaluate the ex ante wisdom of taking a risk if all turned out well) but unsuccessful bets from an ex ante perspective. That would credit the leader with their winnings but not with most of their losses, and would overincentivize betting.
Right—I think a major crux between Nathan and Titotal’s comments involve assumptions or beliefs about the extent to which certain leaders’ long-term effectiveness has been impaired. My gut says there will ultimately be very significant impairment as applied to public-facing / high-visibility roles, less so for certain other roles.
If almost all current leaders would be better than any plausible replacement, even after a significant hit to long-term effectiveness, then I think that says something about the leadership development pipeline that is worth observing.
If almost all current leaders would be better than any plausible replacement, even after a significant hit to long-term effectiveness, then I think that says something about the leadership development pipeline that is worth observing.
I think it’s relatively obvious that there’s a dearth of competent leadership/management in EA. I think this is even more extreme for EA qua EA, since the personal costs : altruistic rewards tradeoff for EA qua EA work is arguably worse than e.g. setting up an AI governance initiative or leading a biosecurity project.
I don’t think we actually want to incentivise positive-EV bets as such? Some amount of risk aversion ought to be baked in. Going solely by EV only makes sense if you make many repeated uncorrelated bets, which isn’t really what Longtermists are doing.
Fair enough—my attempted point was to acknowledge concerns that being too quick to replace leaders when a bad outcome happened might incentivize them to be suboptimally conservative when it comes to risk.
If a global health organization made a mistake in judgment that caused [its] effectiveness to permanently decline by (say) 30%, and it was no longer effective in comparison to alternatives we could counterfactually fund, I suspect very few of us would support continuing to fund it. I would find it potentially concerning, from a standpoint of impartiality, if we do not apply the same standard to leaders. After all, we didn’t protect the hypothetical global health organization’s beneficiaries merely out of a sense of fairness.
I see the argument that applying such a standard to leaders could discourage them from making EV-positive bets. However, experiencing an adverse outcome on most EV-positive bets won’t materially impact a leader’s long-term future effectiveness. Moreover, it could be difficult to evaluate leaders from a 100% ex ante perspective. There’s a risk of evaluating successful bets by their outcome (because outsiders may not understand that there was a significant bet + there is low incentive to evaluate the ex ante wisdom of taking a risk if all turned out well) but unsuccessful bets from an ex ante perspective. That would credit the leader with their winnings but not with most of their losses, and would overincentivize betting.
I suspect a big part of the disagreement here is whether this aspect of the analogy holds?
Right—I think a major crux between Nathan and Titotal’s comments involve assumptions or beliefs about the extent to which certain leaders’ long-term effectiveness has been impaired. My gut says there will ultimately be very significant impairment as applied to public-facing / high-visibility roles, less so for certain other roles.
If almost all current leaders would be better than any plausible replacement, even after a significant hit to long-term effectiveness, then I think that says something about the leadership development pipeline that is worth observing.
I think it’s relatively obvious that there’s a dearth of competent leadership/management in EA. I think this is even more extreme for EA qua EA, since the personal costs : altruistic rewards tradeoff for EA qua EA work is arguably worse than e.g. setting up an AI governance initiative or leading a biosecurity project.
I don’t think we actually want to incentivise positive-EV bets as such? Some amount of risk aversion ought to be baked in. Going solely by EV only makes sense if you make many repeated uncorrelated bets, which isn’t really what Longtermists are doing.
Fair enough—my attempted point was to acknowledge concerns that being too quick to replace leaders when a bad outcome happened might incentivize them to be suboptimally conservative when it comes to risk.