I have seen far from all the talks here, but from what I've seen and the buzz that I've heard the winner of the "coolest presentation award" was Manolis Kelaidis. He showed a paper e-book device that he's been prototyping. By means of conductive ink traces, a person touching a button on the page can trigger an action by an embedded processor. He had a book where pads on the page triggered actions on his laptop: going to web pages, playing songs on iTunes, and so forth.
It was a hand-bound Bluetooth book!
There's clearly a huge expense still involved in platform building and so on, but everything he did is compatible with contemporary printing technology, using inks that are commercially available (not experimental). Other developments in printable circuitry play into this as well: printable batteries, printable electronic components, printable speakers.
Of course this is a technology, not a solution, and there's a huge chain of associated requirements -- protocols for books to notify other devices; security regimes to define so that your new books won't hack your computer. Based on experience in other media, conventions for authoring and use of the new technology are likely to be the most dificult adaptation.
But the great thing about paper is that you can write on it!
I immediately started thinking that this would be a great addition to a notebook. I'd buy a book with 10 generic buttons to a page, and any of them could be mapped to a function on my machine easily. So I could take notes on something I was recording as audio or video and whack a button on the page to make a link. Or to record the page that I'm browsing right at this moment. 100 pages is 1,000 special buttons to memorize some data!
The other interesting thing is that you can make conductive traces by using a silver ink marking pen, so you could just draw buttons onto a page anywhere you want, connecting them to traces on the edge.
If the notebook shell had a slot for a memory and the processor, and a unique ID build in, you might be able to bring the hardware costs way down because you'd just have to clip the brain onto the book, and the brain would know what book it was connected to, so you could have fewer chunks of electronics.
Links to follow soon