Text-to-4D dynamic scene generation

Find it here, via Ryan Watkins.  Further improvement is required, but the pace of current breakthroughs is remarkable.