One reason I love the internet is that it's historically text-based and therefore accessible to me. One fear I've had is that we're moving towards audio and video content that I can't hear or understand.

I suppose I shouldn't have worried. Three projects that prove why:

1. Sphinx, an open-source voice recognition engine developed at CMU. You can set this baby up yourself - in fact, DJ got an install running on his Olin laptop in the casual span of an evening. And it's on par for accuracy with commercial packages.

2. Informedia, a software suite (also by CMU) that I'd describe as "automatic closed-captioning for videos." Using the Sphinx engine, it automatically processes your movies and gives you a stream of text time-synced with the video itself. That means you can textually search movies and have it jump to specific spots in the recorded conversation.

3. Google video... with captions. The result of something like Informedia being applied to all the media content out there.

Hurrah for all the people behind these creations! I hunger for information - things like this open up a whole section of the world's data to me.