Those who are looking to become decent chess players seem to need to learn a repertoire of openings, tactical skills, and endgame patterns.
With standard chess, that seems to involve a lot of memorization, especially insofar as openings are concerned. They have largely been analyzed with computers, and people know what the strongest lines and responses are (though there are novelties in top-level play).
Fischer Random Chess is a bit encouraging in that it reduces the usefulness of memorization and forces a bit of creativity into every game. By randomizing the position of the back-row pieces like knights and bishops, it creates a wide variety of different starting positions. It is not feasible for human beings to memorize ideal lines in all of them, though computers will surely be able to do so eventually.
For the moment, however, during games between human beings, Fischer Random Chess seems to have good potential as a way of make the game more about realizing the implications of positions that are new to you, and less about remembering ideal responses calculated elsewhere.
Take rainbow tables out of chess! (at least some of the time)
P.S. The Chronos GX chess clock can produce randomized positions for ‘Shuffle Chess’, but it does not follow the Fischer Random Chess rule about having the king between the rooks. I wish it did, since I think Fischer Random Chess is likely to produce more balanced results than completely random shuffle chess, in terms of reducing the number of positions in which white has an overwhelming first-mover advantage and making castling reasonably fair and simple.
P.P.S. The Fischer Random rules are also probably better than shuffle chess insofar as they produce a game more similar in character to chess. It’s like chess, but with less focus on memorization. It’s not a totally different game.