I think you’ve got the right idea that CEV aims to find that fixed, ultimately-best-possible set of values.
If I understand correctly, CEV is mostly intended as a shortcut to arrive as close as possible to the same ethics we would have if all humans sat and thought and discussed and researched ethics for [insert arbitrarily large amount of time] until no more changes would occur in those ethics and the system would remain logically consistent and always the best choice for all circumstances and in all futures barring direct alteration of elementary human values.
There may be some conflation between CEV and particular implementations of it that were discussed previously, or with other CEV-like theories (e.g. Coherent Blended Volition). I may also be the one doing the conflating, though.
I think you’ve got the right idea that CEV aims to find that fixed, ultimately-best-possible set of values.
If I understand correctly, CEV is mostly intended as a shortcut to arrive as close as possible to the same ethics we would have if all humans sat and thought and discussed and researched ethics for [insert arbitrarily large amount of time] until no more changes would occur in those ethics and the system would remain logically consistent and always the best choice for all circumstances and in all futures barring direct alteration of elementary human values.
There may be some conflation between CEV and particular implementations of it that were discussed previously, or with other CEV-like theories (e.g. Coherent Blended Volition). I may also be the one doing the conflating, though.