Models of ImageWAM: Do World Action Models Really Need Video Generation, or Just Image Editing?