I was really more interested in whether or not we can observe these models running independently in real, currently living humans (or chimps or rats, really). This way, we could gather some concrete evidence in favor of this three-model approach; and we could also directly measure how strongly the three models are weighted relative to each other.
If you could reduce the cognitive cost of the model-based system by designing a “decision-making app”, you could directly test if it was beneficial and actually (subjectively or otherwise) improved the subject’s lives. If it was successful, you’d have a good chance of beta-testing a real CEV.
I was really more interested in whether or not we can observe these models running independently in real, currently living humans (or chimps or rats, really). This way, we could gather some concrete evidence in favor of this three-model approach; and we could also directly measure how strongly the three models are weighted relative to each other.
If you could reduce the cognitive cost of the model-based system by designing a “decision-making app”, you could directly test if it was beneficial and actually (subjectively or otherwise) improved the subject’s lives. If it was successful, you’d have a good chance of beta-testing a real CEV.