Can’t compare it to other AI Video Generation models as it’s the only one I’ve used. The process is really interesting given that the software is still in its infancy. Story board your narrative with Claude or whatever frontier model one uses. Then you generate in Midjourney characters, props, settings, etc when precision is necessary. Then you basically pull the creative slot machine over at Dreamina (Seedance’s consumer facing platform). 15% of the results are exactly what you want. 45% is unusable and the other 40% you have to make compromises while editing to fit the square pegs into round holes when necessary. Bc it’s a Chinese model, the filters were a total bish considering I was making subversive content about techno-autocracy.
But in general I’m super impressed. Probably took me 15-20 hours bc I had no clue what I was doing. Think I could probably cut that down by two thirds. Outside my normal Claude subscription cost me just under $100 to make.
