Yascha Mounk und David Bau untersuchen die rätselhaften internen Prozesse, die das Verhalten von KI bestimmen – und warum sie uns womöglich grundlegend fremd sind.
The discussion of the self here makes me think of an important distinction.
Representing “I” is not the same as forming an “I.” A model can identify itself, describe its own rules, and store information about itself. All of these can be forms of self-representation.
But the self also involves a continuously confirmed life path and the continuous self-narrative gradually formed through that path.
So finding a self-representation inside a model may mean finding one condition of the self, not the self itself.
The discussion of the self here makes me think of an important distinction.
Representing “I” is not the same as forming an “I.” A model can identify itself, describe its own rules, and store information about itself. All of these can be forms of self-representation.
But the self also involves a continuously confirmed life path and the continuous self-narrative gradually formed through that path.
So finding a self-representation inside a model may mean finding one condition of the self, not the self itself.