Reading DNA without taking it out of a cell sounds like such an awesome idea. All you need is bunch of proteins to do that. All those proteins are just DNA sequences. possibly even Biobricks. So all that translates to "Just introduce a plasmid into a cell and the necessary woodoo happens inside the cell... which sets off some kind of cascade which can be read out by say a CCD or voltage detector." So the startup is all about just designing DNA sequences for protein. And these days even novices have begun to design proteins. This is just brilliant!!! Instead of 1000$ per genome how about 99 cents?? Of course I have no clue how long it should take to reach there beyond extrapolating on a log graph...
Someone here stated the moore's law for sequencing. Which while true needs disruptive startups for each transition. Fortunately or unfortunately these startups dont work like an intel doing similar drudgery year after year to continue doubling... each major leap happens by a radical switch of strategy. As he states the case of Ion Torrent, each step centers around an awesome breaktrhough. But it also raises the question... Will the cost of computation become a bottleneck for a 99 cent genome. Sequencing and synthesis are a lot more useful than just medical diagnosis. Let us not undermine the value of a 99 cent genome or a 1 cent genome sequencing/synthesis capability.