Bullshit. Both Fi and Ti are logical systems. If you simply re-labeled them you would clear the confusion:
Ti: Logic of simplicity.
Fi: Logic of complexity.
They work more or less this way:
Imagine that reality is like a tree. There is a common trunk and there are branches, and sub-branches and so on.
Ti works best when information is isolated from everything else. The ideal statement for Ti is one where two homogeneous bits are evaluated, like "Are 747 planes? Yes", "Are women men? No". For this reason Ti tries to go up the tree, because each hierachy level isolates it from others bits of information.
Fi does the opposite: it feels at home when it comes to balance several bits of information at once. So it tries to go down the tree and find the common root for all branches involved. That's why ethics, diplomacy, etc are the most widely known -but certainly not the only ones- manifestations of it.
Let's put a practical example:
It's war time and Joe, Marie and Anna are fighting for a piece of bread. What's one going to do?
Ti answer: 1/3 = 0.333(...) for each one. (ONE FACTOR INVOLVED)
Fi answer: The piece of bread will be cut into pieces. The size of each piece will be determned by looking at the level of undernourishment of each one. (TWO FACTORS INVOLVED)
Joe, Marie and Anna are persons which possess many individual characteristics. For example, they have an age, gender, body constitution, family background, etc. Ti cannot process so much information efficiently, so it tries to reduce the problem to its simplest form. Fi, on the contrary, processes large amounts of information efficiently and thus tries get as many details as possible.
This takes us to the socionics pairings of functions:
Fe pairs with Ti because Fe is an information reducing function: it's selective. This is most obvious on Fe dominants because they care deeply about their tastes and rarely care about anything that is outside them. This is comfortable for Ti types because the excess of information is removed, leaving place for simplified analysis.
Te pairs with Fi because it's an information increasing function. Te dominants have an ever increasing repository of information and that's comfortable for Fi types because they need details in order to make system-wide fine-tuned decisions.
In evolutionary terms, it seems to me that Ti is a shortcut to Fi. Its view and understanding of the world is artificial and unrealistic, but it is highly efficient when it comes to solve problems, so it's been rewarded by nature over time. However, an optimal solution made for a two factor problem is not automatically the optimal solution for a problem which has hundreds, thoushands or even more factors involved. And that's where the "shortcut" nature of Ti becomes evident: over time one realizes that short term solutions are at expense of long term ones.
Want examples? Well, remember the Y2K "issue". Some random programmer decided that it was ok to cut the date's year to two digits, in order to save memory. Perfectly valid logical statement, as the only factors involved were the need for memory and the observation that the first two digits for the date didn't change often. But what happened when other factors came in?