If we also want to ensure that “crime doesn’t pay” and internalize the benefits of bad acting, i.e. bad acting doesn’t counterfactually benefit actors according to their views, so we can appropriately deter such acts in the future, we may want to
Pay out extra to account for the bad acting that isn’t caught,
Pay out as a community by cause to roughly approximate paying back each grant (since I don’t think grantees should be responsible for paying back). If we pay out uniformly (or proportionally with respect to how non-FTX funding is being spent) across causes, FTX and associates will have been able to shift overall EA funding towards longtermism, which is still to their benefit on their worldviews. To prevent this shift, we would need to pay out disproportionately from the longtermist budget. (This doesn’t necessarily mean the share of non-FTX EA funding towards longtermism should decrease overall, since there are other reasons to increase it, and I don’t know how things will balance out.)
Benefiting from fraud or other instrumental harm is plausibly wrong according to many moral views, and this could be the right response on them. Even on consequentialist views, if we’re perceived as benefiting from fraud and too accepting of instrumental harm, we may push away many people who might otherwise contribute to our community and projects, and could be looked at more skeptically when engaging politically and with institutions. We’ll lose some public trust, and perhaps rightfully so, since we’ll be less worthy of it. Negating the benefits fully means we won’t have benefited from fraud, and lets us honestly say so.
I think there’s a case for paying back all of what we benefited from FTX and associates (and perhaps more to account for fraud not caught) to internalize the damage of bad acting. See this post by tailcalled. (However, I think it makes more sense for this to become the responsibility of major funders, not grantees.)
If we also want to ensure that “crime doesn’t pay” and internalize the benefits of bad acting, i.e. bad acting doesn’t counterfactually benefit actors according to their views, so we can appropriately deter such acts in the future, we may want to
Pay out extra to account for the bad acting that isn’t caught,
Pay out as a community by cause to roughly approximate paying back each grant (since I don’t think grantees should be responsible for paying back). If we pay out uniformly (or proportionally with respect to how non-FTX funding is being spent) across causes, FTX and associates will have been able to shift overall EA funding towards longtermism, which is still to their benefit on their worldviews. To prevent this shift, we would need to pay out disproportionately from the longtermist budget. (This doesn’t necessarily mean the share of non-FTX EA funding towards longtermism should decrease overall, since there are other reasons to increase it, and I don’t know how things will balance out.)
Benefiting from fraud or other instrumental harm is plausibly wrong according to many moral views, and this could be the right response on them. Even on consequentialist views, if we’re perceived as benefiting from fraud and too accepting of instrumental harm, we may push away many people who might otherwise contribute to our community and projects, and could be looked at more skeptically when engaging politically and with institutions. We’ll lose some public trust, and perhaps rightfully so, since we’ll be less worthy of it. Negating the benefits fully means we won’t have benefited from fraud, and lets us honestly say so.