Check out Photosketch, a new way of finding images - you draw and annotate objects, and algorithms, combined with image search, can compose the perfect photograph for you! Cool! Google image search can show you just line art, or images by color tone. Reminds me of a pioneering project called QBIC (from IBM Almaden) that saw little widespread use. Full circle, I say!
At IBM, where I started my career, I'd heard people say "We did this ten years back ('twas called something else)". This was actually true! Technology cycles! There is, however, a reason why cynicism around having tried something before (and failed) in technology is misplaced - the environment has changed. The second time is not just luckier - it is often bigger and better (and sunnier :-))
So, when the Network Computer was introduced in the 90's and it did not fly, it was not because it was a bad idea - in fact, it was brilliant - it was because of other conditions, e.g., the networks were not fast enough. Sun employees run the Network computer as an X display over DSL, and that's exactly the issue. ("Credit" to Sun for sticking with it, killing productivity in the process). Today, the network computer (Net-scape?) is now viable as Cloud computing, and Google's the dot in dot.cloud, the big honking mainframe!
When we built a scalable video server at IBM in 1995, with real-time support, striped file systems, high-bandwidth networks, and a million dollar SP2 m/c, the initial pilots worked great, but the business model floundered - karaoke-on-demand wasn't an established market, and the cost per video stream was too high :-). Today, multiple streams can be served by a Linux desktop, and the idea is viable - it's called YouTube! (Note the fact that today's user does not pay for the stream - the environment has changed! Free works quite well!)
Heard the same thing when trying to push forward on a project at Yahoo! - "We tried this back at Altavista and it did not work!" Famous last words...(actually he's still going strong... :-))
Memory-based computers were prototyped many years earlier, and are only now becoming viable with Flash memory getting to the right price point. Paper-like displays, seen in the 90's, are still not in vogue, but HD televisions are everywhere, and paper-like displays will come (Kindle 4?). Tablets failed, but tablets will come (Courier? iPad?) ! Windows Vista ebbed, Windows 7 may flow...
Don't ever listen to someone who says, "We tried it 10 years back, but it failed!" Remember that technology comes full circle, but the second circle can be brighter, and someone else is usually holding the flashlight!