There is a limit to the complexity a tabletop RPG can reach before it becomes a demonstration for why videogames are better at handling complexity of mechanics, and it's a rather low threshold.
The threshold is this: if you're spending more time working mechanics than engaging in the situation using natural language, then your game is too complex for the tabletop medium.
The operation of game mechanics turns players into pilots, and their characters into robots- mecha, if you will. You can see why I find this undesirable now that videogame RPGs are well past 30 years of age themselves, and do that sort of mechanical complexity so much better than tabletop RPGs that it is wise to outright let the videogame makers have it to themselves and instead refocus tabletop RPGs back to the simplicity of natural language that the original Dungeons & Dragons (and other games of that launch era) did so well.
I have World of Warcraft for the sort of complexity you find in Pathfinder, and that includes all of the culture of play that goes with it. It is not surprising that a game defined by the operation of mechanics and the performance in a team promotes a "Best-or-Benched" mentality with min-maxing, number-crunching, and blah-blah-blah that sucks the fun out of RPGs for me and many others. Tabletop RPGs start sucking when they go that route.
In addition to other vitally-necessary changes that tabletop RPG makers much do to survive, this reversion has to happen; when you bring back the original play paradigm, you recreate the original play culture along with it- a far more friendly, sensible, and casual-friendly culture of play that working adults with children cannot help but find attractive. Your design must focus on using natural language, foster trust in the Game Master to issue rulings as necessary (because you CANNOT cover it all with rules), and get away from keywords, jargon, interactions between X and Y rules, etc.
We don't need more mech factories. We need games about playing characters.