| As somebody who did tech work for traders who dealt with a number of exchanges, I am hugely skeptical that an exchange can be fully automated. A real exchange has an enormous team of well-paid people whose job it is to make sure things run smoothly. They also have incredible power: they can freeze the exchange, limit price movements, undo transactions, kick people out of the market, and who knows what else. They use this power rarely, but when they do market participants are generally glad they did. I'm sure somebody can automate 90% of the cases, or 99%, or even 99.9%. Most trading is routine. But will this work when there's a war? A financial panic? A major fraud or theft? An exploitation of a previously unsuspected bug? The New York Stock Exchange has been working this stuff out since 1817 (or 1792 depending on how you count). And we still have things like Black Monday [1] and the 2010 Flash Crash [2]. This will be an interesting experiment, but I will not be surprised at all if it's something most investors avoid for decades because it's a) novel and therefore risky, and b) lacking a lot of the regulatory protections of a real exchange. [1] http://en.wikipedia.org/wiki/Black_Monday_%281987%29 [2] http://en.wikipedia.org/wiki/2010_Flash_Crash |