I guess the central question is how important realistic conversation is in a story-based game. IMO it’s a big deal and LA Noire is likely to be a huge leap in that respect. Conversations have three basic components: text, voices and facial expressions. The first two have improved a lot over the years and reached an acceptable quality but the third is still deep at the bottom of the uncanny valley even in state of the art games like Heavy Rain. LA Noire doesn’t completely escape the uncanny valley but to me at least looks like a big leap on the way out. It’s not at all hard to believe that in a few years this technology perhaps combined with more powerful consoles will deliver largely natural conversations in video games. IMO that does represent a potential revolution.
As to how this is different from full-motion video, obviously those games were a dead end and they just didn’t deliver interesting gameplay for the most part. One could theoretically deliver gameplay like LA Noire using FMV but constantly cutting from video to the interactive game and back would be jarring and not a lot of fun. Another way of looking at is that LA Noire is pretty much doing what the FMV games attempted but that finally the technology exists to do it right and create a genuinely compelling and seamless gaming experience with realistic facial expressions. Much like 3-D films have existed for decades but only recently have we seen the technology to create truly compelling 3-D films like Avatar.