One problem with conceptualizing the future is the difficulty of plotting multivariable scenarios. We imagine big changes in one area -- say, super highspeed internets -- without understanding what else might be different and what impact these other changes will have on innovation (previously on this topic). 

So you can't just think about fatter pipes, you have to think about things like the devices we'll be using to access and manipulate the data torrent. Yeah, the iPad ads look pretty neat -- but that's just the beginning. When nanophotonic avalanche photodetectors go mainstream, we'll want to have the bandwidth to use the new tools to the fullest extent possible:

Comments

[Fred Gregory]

This looks familiar

[Jim Caserta]

IBM's integration of optical components is significant for the same reason having fiber to your home is a bonus over coax-cable - optical media have bandwidths far in excess of electrical wiring methods. For many systems - IBM is targeting server-server communications - the bottleneck is no longer processor speed but processor-processor communication speeds. Lots of very smart people have been working on improving bandwidth capability of wired networks, integrating wireless components (I can point you to the initial work on integrating antennas onto chips), and the 'holy grail' of chip comms was integrating optical components. It seems like IBM has done it. They have pioneered a lot of manufacturing trends: Silicon-Germanium transistors, and making copper interconnects yield at high rates.