[lairv]

I wouldn't have thought we'd have these text-to-video models before the end of this year, so we might actually not be that far from minutes-long video generation