I don’t believe humans are badly modeled as single agents. Rather, they are single agents that have communicative and performative aspects to their cognition and behavior. See: The Elephant In The Brain, Player vs Character.
If you have strong reason to think “single agent communicating and doing performances” is a bad model, that would be interesting.
In this case, “convincing yourself” is clearly motivated. It doesn’t make sense as a random interaction between two subagents (otherwise, why aren’t people just as likely to try to convince themselves they have bad qualities?); whatever interaction there is has been orchestrated by some agentic process. Look at the result, and ask who wanted it.
I don’t believe humans are badly modeled as single agents. Rather, they are single agents that have communicative and performative aspects to their cognition and behavior. See: The Elephant In The Brain, Player vs Character.
If you have strong reason to think “single agent communicating and doing performances” is a bad model, that would be interesting.
In this case, “convincing yourself” is clearly motivated. It doesn’t make sense as a random interaction between two subagents (otherwise, why aren’t people just as likely to try to convince themselves they have bad qualities?); whatever interaction there is has been orchestrated by some agentic process. Look at the result, and ask who wanted it.