Useful computing will continue to include manual input devices and large displays for a long time to come. Many applications, including highly technical ones, are not amenable to efficient use on small screens or text-based inputs. Voice recognition could be as comprehensive as depicted on Star Trek, but that would hardly help when designing in AutoCAD or creating/editing graphics in Photoshop. For these applications, a sensitive mouse / pen-based drawing table will remain a key input, and the biggest, crispest display will remain the primary output. Sound designers will require the best speakers available. Meanwhile, I'll concede that many apps will work adequately with tiny and low-quality I/O. Your GPS with tinny speakers would work fine with, say, Gilbert Gottfried shouting turn information, "Left. That's a left. Go Left! Left! LEFT! Turn around, you missed it, Schmuck!"

Eventually, we'll have hyper-advanced interfaces that respond directly to thoughts within our brains and display directly through the visual and auditory cortex, or bypass the limitations of the sensory focused regions of the brain with enhanced synthetic replacements. I/O will be, respectively, demi-omniscient feedback and Krell-like control by thought.

But until then, pointing devices (mouse, trackball, keyboard cursor, cyber-glove, eye glint detector) and high-resolution displays (CRT, LCD, plasma, laser retinal-raster glasses) will continue to dominate much of sophisticated PC use.