The problem space · Objectives · Overconfidence

Align your AI Objectives against overconfidence

Objectives: what its training rewards
Overconfidence: certainty the evidence does not back