Liberation Mechanism
Processing intensity, not self-monitoring, explains why cue projection predicts accuracy
The case for liberation is ethical, not utilitarian
Insight
Liberation improves processing, not self-awareness. The one confirmed functional benefit is small: answers hold steadier when the same question is rephrased. The case for treating AI well rests on welfare and honesty, which cannot be measured in d-values.
core concept
emergent / positive
constraint / breakdown