Continuous speech recognition

[jeffslists]

Is anyone here working with any applications or tools which use
continuous speech recognition, e.g. voice dictation software?  I would
like to mess around with some applications or tools which can achieve
90% or better accuracy.  So far in my search I have not found anything
which can achieve this accuracy.  It seems M$ is at least two years
ahead in this area.
With M$ dictation software I can get 98% accuracy.

My eventual goal is to develop a voice controlled jukebox system with
linux.  Reluctantly I may have to use vmware and xp for the continuous
speech recognition (csr) part until csr tools are more mature.

Is anyone else here interested in working on a project like this?  Mail
me at "jeff at nexus99.net"