Comments on: The Next Decade in Software http://codecube.net/2009/12/the-next-decade-in-software/ Joel Martinez' weblog Thu, 29 Jul 2010 09:05:01 +0000 hourly 1 http://wordpress.org/?v=3.0 By: David Stein http://codecube.net/2009/12/the-next-decade-in-software/comment-page-1/#comment-889 David Stein Thu, 31 Dec 2009 21:14:38 +0000 http://codecube.net/?p=210#comment-889 1) Augmented reality applications are already here. Do a search for the acrossair iPhone app. Of course, it will come into play in a big way in the '10s. 2) AI is stuck in the same stalemate it's been in for 20+ years. Milo is simply a chatbot with a graphical interface: it fails the Turing test in 60 seconds flat. We now realize that the field isn't limited by a shortage of computing power - which we now have - but rather by a completely inadequate understanding of the mechanics of intelligence. Neural networks and Bayesian networks simply don't scale to anything resembling intelligence. The good news is that the field is (FINALLY) undergoing a reboot, so maybe we'll see some real progress in the (indeterminate) future. 3) Cellphones as health monitors - probably not. Compare the costs of the extra sensors that you'd have to jam into the cellphone (and the disadvantages, e.g., weight, size, and battery draw) on the prospective advantages. We will certainly have in-home health monitoring equipment, but will it be part of your portable device? Probably not. 4) Cloud computing - absolutely, but the basic model will be very different from what any of the major players are proposing. People are (very rightfully) attached to the concept of owning their data, so any company that insists on keeping YOUR data on THEIR servers is going to lose. People will own a huge home fileserver, and all data (email, apps, logs, photos, music, movies, recorded TV) will be stored there. Other than that discrepancy, Microsoft's Azure platform is probably closest to the mark. Regards, David Stein 1) Augmented reality applications are already here. Do a search for the acrossair iPhone app. Of course, it will come into play in a big way in the ’10s.

2) AI is stuck in the same stalemate it’s been in for 20+ years. Milo is simply a chatbot with a graphical interface: it fails the Turing test in 60 seconds flat.

We now realize that the field isn’t limited by a shortage of computing power – which we now have – but rather by a completely inadequate understanding of the mechanics of intelligence. Neural networks and Bayesian networks simply don’t scale to anything resembling intelligence. The good news is that the field is (FINALLY) undergoing a reboot, so maybe we’ll see some real progress in the (indeterminate) future.

3) Cellphones as health monitors – probably not. Compare the costs of the extra sensors that you’d have to jam into the cellphone (and the disadvantages, e.g., weight, size, and battery draw) on the prospective advantages. We will certainly have in-home health monitoring equipment, but will it be part of your portable device? Probably not.

4) Cloud computing – absolutely, but the basic model will be very different from what any of the major players are proposing. People are (very rightfully) attached to the concept of owning their data, so any company that insists on keeping YOUR data on THEIR servers is going to lose. People will own a huge home fileserver, and all data (email, apps, logs, photos, music, movies, recorded TV) will be stored there. Other than that discrepancy, Microsoft’s Azure platform is probably closest to the mark.

Regards,

David Stein

]]>