Forecasts 10-14 days out have always had poor predictability and are bound to change, even significantly. There are so many factors on a mesoscale level (which might as well be impossible to predict at such a long range) that can influence how large-scale patterns evolve and the resultant temperature/precipitation people observe. At best, predictions that far out are educated guesses based on past observations. But if weather is anything like history, it can rhyme but never repeats.
As far as short-term forecasts, also in recent history, we as average joes have also become way overexposed to a lot more weather forecasting tools than we were traditionally privy to, and that's without necessarily having a complete understanding of the science behind them. If you think back in the 90s or even the early 2000s, we weren't able to easily pull up radars, satellites, model outputs, skew-t soundings, indices, SSTs, MJO plots, etc. that are used for predictions like we are in 2022.
I guess my point being, I'm not so sure NOAA's accuracy is any worse or better today than it was in the past. I think it's mostly an issue of information overload for us average joes in combination with modern meteorologists' desire to be overly precise with their forecasts that it leaves them more vulnerable to critiques when things don't go as they predict
That's my take, but maybe I missed your angle...