So, hypothetically speaking, do you think that paying people without measuring them would be harmless? Though I am not sure whether this question even makes sense in real world, because at least you need to make a decision how long you keep paying someone (unless you just committed to paying them forever, regardless of whether they do something or not), which implies some form of measure.
I don’t know how it would be possible to pay without measuring. At the extreme of “unmeasured payments”, I presume a national or global UBI wouldn’t significantly harm any given community or shared space that isn’t related to the UBI criteria. Narrower payment schemes require AT LEAST objective identification of who pays (or is paid), why, and some exclusion mechanism to prevent overpayment. It’s possible this could be minimally-invasive, but improbable.
And, to the extent it’s not distortionary, it’s also not motivating. Payment is (almost always) intended to have an effect on behavior. When it has no effect, there’s not much reason for payment (there may be reason for grants or gifts as rewards, thank-yous, or kindness, but if they’re regular enough they become expected and impactful on future motivation).
Relatedly, the legibility and universality of currency will cause any payment scheme to be considered in terms of scaling. How do we make these measured dimensions bigger (or keep them smaller, in some cases), regardless of impact on all the unmeasured values that people take from the interactions.
When I started this subthread, I didn’t think this was the same as Goodhart, but maybe it is: if people like doing something, and you measure part of it for monetary purposes, you’ve got an imperfect proxy for that value.
So, hypothetically speaking, do you think that paying people without measuring them would be harmless? Though I am not sure whether this question even makes sense in real world, because at least you need to make a decision how long you keep paying someone (unless you just committed to paying them forever, regardless of whether they do something or not), which implies some form of measure.
I don’t know how it would be possible to pay without measuring. At the extreme of “unmeasured payments”, I presume a national or global UBI wouldn’t significantly harm any given community or shared space that isn’t related to the UBI criteria. Narrower payment schemes require AT LEAST objective identification of who pays (or is paid), why, and some exclusion mechanism to prevent overpayment. It’s possible this could be minimally-invasive, but improbable.
And, to the extent it’s not distortionary, it’s also not motivating. Payment is (almost always) intended to have an effect on behavior. When it has no effect, there’s not much reason for payment (there may be reason for grants or gifts as rewards, thank-yous, or kindness, but if they’re regular enough they become expected and impactful on future motivation).
Relatedly, the legibility and universality of currency will cause any payment scheme to be considered in terms of scaling. How do we make these measured dimensions bigger (or keep them smaller, in some cases), regardless of impact on all the unmeasured values that people take from the interactions.
When I started this subthread, I didn’t think this was the same as Goodhart, but maybe it is: if people like doing something, and you measure part of it for monetary purposes, you’ve got an imperfect proxy for that value.