For a harmonic oscillator, there is a simple duality that exchanges position and momentum. However, it does not imply that the harmonic oscillator and the dual harmonic oscillator are just special cases of some more general unifying theory. The electro-magnetic duality relates electric and magnetic monopoles. However, it does not imply that electric and magnetic monopoles must be only special cases of some more general unifying theory. In 2 dimensions, there is a remarkable duality between the sine-Gordon and the Thirring model. However, it does not imply that these two models are just special cases of some more general unifying theory. Given these well understood examples, then why, for the god sake, the string dualities do imply (or at least suggest) that all these string theories should be just special cases of a unifying but mysterious M (or F) theory? Is it just because there are many, and not only one, duality? Or just because otherwise string theory would not be cool enough? Or is there a more rational argument? What is wrong if we just say: OK, there are 5 string theories mutually related by dualities to each other, at low energies there is also duality to a yet another theory - 11 dimensional supergravity, and that's it?