Twoh!

" twoh --" as if the person who is listening to this algorithmicly created dialogical amalgamation of everything that is being said on twitter, searched on google, and talked about in relation to the top performing stocks on the Dow Jones -- had just started to say "Twitter" in recognition, but then had started to say "now" in correction, but then had lost all linguistic capability except an "oh!" of appreciation, because they were listening to the synthesized voice of the Internet. The emotional tone could raise and lower with the vicissitudes of the stocks. The sentences could be a composed of the most popular key words of the moment strung together by logical operators that create algorithmicly determined meaning.

The actual voice could be created by an automatic system that quickly samples Internet based audio recordings of the celebrities most mentioned at the moment and creates a synthesized voice based on an average of all the voices.

I guess this would be like an audio and temporal version of those face-average visualizations where they put a bunch of faces together.