I started doing patent cases in September, 1989. At that time, I could argue anticipation from multiple unrelated references. Claim construction? Maybe leave it for the jury, maybe the judge, maybe never decide it. DOE was the wild west. The court has done a lot of good.
Nowadays the disputes to me are narrower, but important. Section 101 is an obvious (ha, ha) area where the make-up of the panel could be outcome determinative. But that would probably be the case no matter who was trying to understand what the Supreme Court believes the law ought to be.
Those key splits I mentioned up front were, in contrast, fundamental questions that had never been resolved among the circuits despite 30 years of work with the patent act. Patent law is a lot better than it was -- better -- and it is a lot more efficient than it was -- I know who is going to tell me what the claim means, and a lot of (albeit indeterminate) rules to help me figure that out. DOE is nowhere like it was. Damages -- you would not recognize the damaegs theories we used to be able to assert.
Could it be better? I have no doubt about that it can: it is a human institution. The role of prior panel decisions sometimes is not given enough weight (a pet peave I'll write about under Section 285, shortly).
Would replacing it with the "old" system make it better? I have no doubt whatsoever that it would not. I have no doubt it would make things worse.