The moderate AI is capped on the number of villagers it can make artificially in order to make the difficulty work. That slow trickle is the result of having a limited economy after that initial buildup. It simply can’t sustain that kind of push for long.
Do you do anything other than taking gratuitous shots at the devs?
What real motivation would they have to design an extremely precise, competent AI for the top 5% of players in AOE when they’ve already given out all the resources to do exactly that to the community to mod it? Every company needs to work off a cost evaluation and the return for making a Barbarian AI is nonexistent.
And on the subject of Custom AI’s. The ResonanceBot in HD was a generally very good AI which was a challenge even for decently skilled (a friend and I would 2v3 them, whereas we really need a 5v2 to be really in trouble against the extreme AI’s) and even those had problems: They’d play Mongols like a toddler. We practically had a free win anytime the resbot would roll Mongols on random. They were that bad. Mind you, this was HD, but still.
The Extreme AI is compared to 800-900 Elo or something like that. If you start learning the basic, you will 1v1 the Extreme AI pretty easily.
At some point it is just time to just leave playing against the AI and start playing multiplayer. Much more fun in my opinion. In ranked 1v1s you get mostly evenly matched games. TGs have some issues, so those games are pretty much onesided for that reason.
In the end getting a good AI requires lots of resources. The Resbot that someone mentions is just a AI that cheats. If you are looking for a good AI, then you have to pick Barbarian. Even some pros would have had troubles against that AI. I have seen Viper and some other pros losing 3v5 against that AI. If they wanted a good AI, they had to implement that AI into the base game.