Questioning GLS-Coherence — AI Alignment Forum