Update 5/31/13 - Derek's also got a great review going at Pipeline.
Remember 2009, when you gasped for a moment at the beautiful IBM structure of pentacene?
|Credit: BBC News | IBM | Science|
|Credit: BBC | IBM|
|Credit: Science | LBNL | UC-Berkeley | Wired|
As if this couldn't sound any more amazing, the researchers were able to predict and visualize several products previously predicted by theory, but never directly observed (stabilized diradicals, anyone?).
So, will this be a standard technique for the practicing chemist? I'm guessing not for quite a few years, since the hardware involved still isn't commonplace, and the technique probably works best at prohibitively high dilutions with flat molecules. Med chem? Sure, you could watch a Suzuki coupling occur, or watch a Cope rearrangement, but for "3D" molecules (read: alkaloids, vitamins, sugars, etc...) I think NMR and X-ray crystallography will still be your best bets.
But to paraphrase the futurists, predictions ironically suffer from poor foresight - after all, just over a month ago, a Japanese group disclosed how to take on-demand crystal structures of just about anything. So, I'm sure someone will invent a "rugged" surface capable of guiding the AFM tip around points and curves to monitor, say, real-time Pictet-Spengler reactions. Can't wait!