First, we're talking about a number of things involved. Currently, someone using only Poser could create a movie with a very realistic feel to it. It'd take time, organization, and a number of machines and people, but it could be done. The technolgy is already available, just look at some of the portraits in the gallery here. Second, the things that will make it really feasable are faster machines and better software. Software needs to be made to allow for direction, the same way a live director can instruct real actors. Blocking, mood, expression... Third, text-to-speach has come a long way, but it isn't there yet. It will be, and probably sooner than you think. The ones leading the charge aren't the movie companies, but the people who do call trees. The more realism and greater ease of change of message makes for a better and more competitive product for them. Movie people can still hire starving actors for voices for much less than the research into realistic text-to-speech. Well known actors often work at scale in animated movies too, just for the experience and free advertising they get. Finally, the best movie style to try this with is the old 1940's and 50's film noir. No special effects, no color, and dramatic but static lighting. Not so much an action picture, but more a mood and emotion picture (read "chick-flick" drama). "Sin City" was close, but I think something like "Dark Victory" or a good Sam Spade movie would be better.