As I am every year, I was again somewhat shocked by how many people come to this meeting. The attendance just keeps going up, and unfortunately that usually correlates with a decrease in the utility of the average talk. For PIs, the meeting is more about interacting with each other, vendors, potential hires (postdoc or faculty) than it is about actually learning things from talks.
I did actually learn a few things in talks yesterday, though, even without going to sessions on the super hot topics (graphene, topological insulators).
I went to the first half of an interesting session about presenting science to the public, something that I think is very important. The first talk was about a theater group collaborating with MIT, with a history of developing plays based on physics (including Einstein's Dreams by Alan Lightman). Here, free of charge, are my suggestions for other play possibilities: The birth of Silicon Valley, emphasizing that in many ways it stems from the fact that Shockley was such a micromanager that no one could work with him; the genesis of the hydrogen bomb, emphasizing that Teller's difficult personality delayed the development of his much-loved "super" by at least two years, because very few people could work with him. (Sense a theme?). The second talk was by Odd Todd Rosenberg, an animator who works with Robert Krulwich of ABC and NPR science fame. The cartoons were funny and informative, though the discussion raised the recurring issue of oversimplification and inaccuracy in broadly popular marketing of science.
I also went to a very strong session all about vanadium dioxide, which was quite informative. Sounds like many people in the strong correlations game are starting to look very hard at ionic liquids for gating. Here the major challenge is whether people are inadvertently (or advertently) doing chemistry on their structures.
There was also a very fun talk by Hongkun Park of Harvard, who discussed "quantum plasmonics" - trying to couple individual emitters and absorbers to plasmonic waveguides. For example, you can have a GaAs nanowire touching a silver nanowire. When biased appropriately the GaAs emits photons which couple strongly into guided plasmon modes of the Ag wire, propagate as plasmons, and are then generate photocurrent at a junction with a Ge nanowire at the other end of the silver wire. This is a "dark plasmonic" circuit, with light being generated, propagated, and detected on the subwavelength scale. Cool stuff.