We are standing today on the edge of a new era. Not an era which opens as a stage curtain with orchestral fanfare on opening night, but more like a majestic mountain landscape emerging from the early morning mist. Practical speech recognition, available from multiple vendors in multiple forms, has been successfully pioneered in numerous applications. The scale and scope of these applications has recently begun to grow appreciatively, in large part, due to the confluence of several key factors. The technology itself has improved sufficiently to provide for satisfactory operational performance, both in terms of accuracy and clear user benefits. Significant capabilities can now be more easily integrated into valuable widely used applications. With the recent availability of powerful inexpensive memory and processors (often packaged in personal computers), high performance recognition has become more affordable. The research tasks, while ever formidable and challenging, are visibly chiseling away old problems (though always revealing new ones!). Through a succession of contributions, by researchers worldwide, the technology base has become progressively stronger. Despite significant unknowns, the prospects of success are now undeniable. This presentation proposes to illustrate, by example, a sample of notable applications and research programs through which we catch glimpses of the enticing vistas ahead.