Is a silent film a movie? Do we have to have Avengers Endgame level of quality before we can say text to movie is a thing? Because there will likely be “movies” of some level of quality that are about an hour or so but might be silent, inconsistent, and extremely boring. So by a very lenient definition of “movie” I would say 1-2 years. However, if you mean roughly Avengers Endgame level of quality, length, etc with only a text prompt, it will likely take a lot longer in the same way full self driving and text to image has not yet been perfected. Sure cars can drive fully autonomously without crashing most of the time in a lot of scenarios, but they fail in too many edge cases. Text to image generators fail to make perfect hands, put red boxes on top of green boxes, etc. Consistent text to movie on roughly the level of any multi-million dollar budget movie would mean some extremely complicated things were worked out, to the point where we’d likely have full self dive VR and have had transformative AI for several years if not already having AGI. So 10+ years is more reasonable to me if you mean text to movie in that sense.

Edit: typos