I'm talking about in the US. You're right on the first point - 95% of patent cases do have a jury. The part I felt was dysfunctional was that often times judges seem LESS aware of many modern topics (e.g. tech) than a cross-section of laypeople. Plus a huge number of patent cases happen in some rando district of East Texas which is problematic, but that's really beside your point.

I'd be okay with a "jury of judges" approach (provided it was at least 6, like a normal jury) to reduce undue influence, if legal system could hypothetically arrange that.