I am surprised by your unqualified use of the term “intelligent machines.” To apply the descriptor “intelligent” to machines assumes that intelligence can be measured across types (e.g. humans and computers) with no relationship to the specific embodiment of the processes collectively referred to as “intelligence.” Similar ideational permissiveness produces at least one other interpretive indiscretion. That is, how can it be that “[i]n the heart of the computer . . . the distinctions between character, writer, and reader blur into strings of ones and zeroes . . .” when the specific context and embodiment of such information (voltage and magnetic differentials contained in the discrete registers and electrons “flowing” through the busses of a computer architecture) is separate from the imagined scene of “ones and zeroes” flowing, undifferentiated, through a computer? This is precisely the kind of substitution Norbert Wiener makes when he defines information as an abstraction separable from the context in which it occurs. Strings of “ones and zeroes” never blur one into the other except under circumstances of electronic or ontological crisis.