The 'digital divide' of the future won't be between those who do and don't have simple access to the internet--the meaningful difference will be between those who have the ability to *use* the staggering quantities of information (already) available to them, and those who are simply 'surfing' the net.
I would argue that this has always been the case, from the beginning; today's easy access to IM and email admittedly makes available new habits and lifestyles, but the simply communicative applications of the present--meaningful as they are--are already becoming a social commodity. The next step will be to allow the individual to make the world (the world's data) work for them in ways that are comfortable: social, intuitive, responsive ways. At the heart of this notion is the idea that 'smart' or 'intelligent' or 'agent' systems of the future (at least the mid-term future) will not be able to deliver their much-speculated benefits entirely on their own. Problems concerning human agency, context, and classification may not ever be entirely automated. Thus the person who understands the conceptual background and organization of the world's information will always have an (increasingly larger?) advantage--socially, economically--over the person who is at the mercy of such systems' recommendation.
The question to be considered: can incisive (social and technical) understanding combine with good design to create power-user tools for everyone? To what degree can Don Norman's "knowledge in the world" compensate for the expert's understanding, of both subject background and technical mastery?