@Cardena
This is what I was responding to:
It's pathetic that Anyone would complain about the AI. If foe wanted to they could make it so that you wouldn't be able to win or barely win any fights unless you had boosts 3 time more than the opposing forces.
I'm making the generous assumption that the majority of those prone to complaining about the AI being too weak, are capable of reliably beating it.
Outdoing a world master of chess was more than just a major project; it was the prestige project of the worlds biggest electronics company, IBM. The science behind it had been decades in the making, before Deep Blue finally accomplished the task. The team behind Deep Blue were the foremost experts on AI in the world, and they had the entire IBM Research organization behind them.
Obviously, there's no need for anything close to a grand master level AI to beat the vast majority of FoE players. There is, however, a huge distance between "relatively easy" and "creating a grand master AI". What took IBM Research eight years, and the hiring of top experts with previous years of experience, wouldn't be difficult for Inno; it would be entirely impossible. It's completely out of their league, and there's no shame in that. You can be a pretty darn good toy airplane manufacturer, even if you can't match NASA in technological expertise.
The current AI relies on wearing players down, through a large number of battles; essentially, when players who regularly fight are brought to a halt, it's not because they've lost a lot of battles, but because of a cumulative erosion of units. The hard battles in GE IV are double; the player takes on twice the number of his own troops, and will generally be at a disadvantage with regard to boost. The AI fights with a huge advantage.
I'd say that an AI that could reliably defeat the top half of active fighters, under equal conditions (same boosts, same number of troops), would require a huge improvement. Assuming that only the top 10% are prone to complaining about the AI, reliably defeating this group is significantly harder still. Furthermore, according to the comment I was answering, they would have to be completely outclassed, to the point where they'd need a triple boost advantage.
Even a fairly average human brain, once it has mastered a skill (which the players we're talking about has), and provided that it's a complex skill (which this is), is incredibly difficult to outperform with an AI. What's inside our skulls are still the worlds most powerful supercomputers. Exactly how long that state of affairs will last, is another matter. The year 2025 has been suggested for about twenty years now, an estimate that has changed disturbingly little. However, even once we're outdone in raw computing power, it will still remain to outdo us on software. We're much further away from managing that.