I intentionally pushed AI into edge cases: ambiguous scenarios, emotionally nuanced communication, conflicting objectives. The failures are instructive.
Patterns emerge. AI struggles when trade-offs require value judgments or when context is partially implicit. Understanding failure modes is as important as celebrating strengths. Blind trust is not strategy.