Many people have shared this observation. Usually our focus is in forecasting snow, and in that context I suspect that the pattern is a significant driver to the feeling of reduced accuracy. For 8 years give or take we have been in a sort of semi-permanent La Nina which tends to mean northern-stream dominant. And as we know those are inherently more difficult for NWP to handle. If we ever get a canonical Nino response again it's possible that guidance may go back to more familiar levels of stability.
There's also the fact that for those 8 years, models have been less stable with predicting snow because, well, the patterns have not been conducive to snow. i can remember several instances where the models stably predicted TORCH from 16 days out, and lo it came to pass (Jan 2020 is an example that comes to mind)