You may have heard Peter Molyneux’s “virtual boy” Milo. He’s the uncanny valley’s answer to the Tamagotchi, and the latest project to take advantage of the Xbox’s “Project Natal.”
You can see Milo in the video below:
Here’s what I find particularly interesting about Milo. Milo learns.
According to James Orry at VideoGamer.com, no two Milos will be the same, because they are customized to the environment and the players on each specific Xbox. But not only will they learn from their players, but also from each other, connected via the Internet to communicate with other Milos in the cloud.
This brings up broader questions about cloud computing and cloud based data. While there are many obvious advantages to cloud computing (being able to access your data from anywhere with an Internet connection), and obvious disadvantages to the technology (being unable to access your data if your Internet connection goes down), one aspect of cloud computing that is, quite frankly, breathtaking, is the ability to gather data on the performance of the applications as a whole, and to roll out updates to all users simultaneously.
In the case of “Milo and Kate,” this takes the form of taking the data from thousands of instances of Milo, processing it, and then re-uploading it from the central server back to Milo, to be improved on again. Iterative, automated improvement.
This is not only useful for video gaming. One client I had operates a cloud-computing based monitoring tool for their software, and, through being able to see the aggregate problems of the entire user base, they can make a much more informed decision on how to improve their products and offerings to help the most people at the same time – then seamlessly roll out the improvements to the customers in one step.
This is partially why Google is so dominant in technology at the current time. Search, of course, only works “in the cloud,” and they are able to take the data from billions of its users and sort through that information in order to figure out what people are searching for, when, where, and maybe even why. They’re able to then take that data and improve their algorithms serving ads alongside search results.
Amazon already has adopted a framework that allows people to do datamining on the already existing Amazon EC2 and Amazon S3 cloud computing services. Some more technical views on data mining for the cloud (at least for SQL Server based clouds) can be found at DM(X).