There is reason to expect that spite strategies, which involve someone paying to harm others, are collective, rather than individual. [...]
Therefore, collective values are more likely than individual values to encode conflicts in a way that makes them fundamentally irreconcilable.
IDK if I get this. This seems to conflate (1) values that only exist because of collective dynamics and aren’t grounded in anything anyone wanted, as discussed in the previous paragraph about “the collective good”, and (2) values that can only be *pursued* by collectives, even if they’re “in the individual”. There seems a big difference between “encoding conflicts” by creating dynamics that “overwrite” whatever the individuals wanted, vs. by enabling individuals to coordinate to acheive their values (or encoding non-conflict by breaking individuals ability to coordinate).
Spite strategies function collectively but are encoded in individual adaptations, e.g. green beard effect. While green beard behavior aren’t well-optimized on an individual basis, the teleology points at a collective strategy, i.e. you compress it better by assuming the gene is selfish than assuming the individuals are.
With FAI there’s an issue of incorporating many different values. My intuition is that if we are choosing one or the other, we should encode selfish-individual values and not selfish-gene values. One reason for this is that selfish-gene values are more likely to be “mean” on a way that makes them irreconcilable.
Now that I’ve looked at the Sarah Constantin link:
I’ve updated that spite strategies are more plausible than I’d thought.
It seems there’s evidence that spite strategies also happen between individuals, though I didn’t look at the actual studies. If that’s right, do you also think “mean”/irreconcilable values held by an individual should be thrown overboard, if some values have to be thrown overboard?
IDK if I get this. This seems to conflate (1) values that only exist because of collective dynamics and aren’t grounded in anything anyone wanted, as discussed in the previous paragraph about “the collective good”, and (2) values that can only be *pursued* by collectives, even if they’re “in the individual”. There seems a big difference between “encoding conflicts” by creating dynamics that “overwrite” whatever the individuals wanted, vs. by enabling individuals to coordinate to acheive their values (or encoding non-conflict by breaking individuals ability to coordinate).
Spite strategies function collectively but are encoded in individual adaptations, e.g. green beard effect. While green beard behavior aren’t well-optimized on an individual basis, the teleology points at a collective strategy, i.e. you compress it better by assuming the gene is selfish than assuming the individuals are.
With FAI there’s an issue of incorporating many different values. My intuition is that if we are choosing one or the other, we should encode selfish-individual values and not selfish-gene values. One reason for this is that selfish-gene values are more likely to be “mean” on a way that makes them irreconcilable.
Now that I’ve looked at the Sarah Constantin link:
I’ve updated that spite strategies are more plausible than I’d thought.
It seems there’s evidence that spite strategies also happen between individuals, though I didn’t look at the actual studies. If that’s right, do you also think “mean”/irreconcilable values held by an individual should be thrown overboard, if some values have to be thrown overboard?
See the other thread about values related to adversarial games.