Oddly, I'd say that all of those are complicated problems, but they aren't necessarily hard problems (most of them are "solved" insofar as there are pretty standard expectations of behavior for them, even if the implementation might be a pain in the ass). In my experience, the hardest parts of making a game are:

• The process of iterating over concepts to construct a pre-prod-ready game concept. Doing this at a pro level requires the ability to swallow your ego, sacrifice your sacred cows, and do your damndest to be objective about fundamentally subjective questions. IMO, that's part of what makes AAA so repetitive: finding a new game is hard. Tweaking an old product is easy.
• Organizing yourselves to actually make the game in a coherent order that avoids major blocks on productivity or gluts of tasks that need handling all at once.
• Realizing when you've made a mistake and need to take a step back rather than continue down a high-risk path out of sunk-cost fallacies.