I think a big part of the problem is that while computer power has improved dramatically, the number of ground and sea-based observation platforms has been equally dramatically decreased, so the number of points that get fed into the 3D model is very small.
The radio sondes go up only 0000Z and 1200Z, and only at a few dozen places in all of Europe. These provide by far the best data.
The rest comes from satellite observations etc.
One thing which many pilots would find helpful would be to watch the MSLP charts daily. Most look at the "weather" only before they plan to fly, but looking at it daily soon teaches you to take one look at an MSLP chart and get a pretty good feel for what is coming.