Uncertainty estimation/Anomaly and Novelty Detection is far from being solved for real-world problems, despite numeruous papers presenting impressive, but mostly cherry-picked results.