Jeremy while obsessing about the quality of the sound and focusing all of his brain cells not dedicated to filtering the Internet dating scene has tasked me with figuring out how to properly enclose and promote the revival of our college radio show in the form of a podcast. The initial test can be found on a blogger blog called Entropy and packaged as an RSS feed using Feedburner.

Among the interesting things that I found for promoting the podcast was the Podzinger site. It is a search engine that indexes podcasts into text and then allows you to search for the content and load the section of the podcast that raised your search results. At first glance this appears to be some incredible technology. What I found interesting about it given that I have been doing some research into the speech to text world is that Podzinger is a BBN technology. Now for those people who are unfamiliar with BBN, they were a big time government contractor building among many things the original infrastructure for the Arpanet, the government’s private predecessor to the public Internet. One speech to text expert had mentioned to me that BBN had a large contract in the 80s building speech to text for the government, probably the NSA, in order to filter international telephone calls. The system would convert International conversations to text and then identify conversations that might be worth listening deeper into for national security. If BBN has been working on this technology since the ‘80s for Uncle Sam then they probably are going to be getting to the point by now where they are good enough at it to really recognize what it happening when people are talking. They might even be better at it than the Microsoft Speech to Text engine.