No, its just a fact of life...
In many cases you won't be able to start to apply one transformation until you have completed the first. Some of these activities are by their very nature multiple pass operations. For example, equalising audio levels or colour balance across an entire segment. You will need a pass to identify the maxima and minima before you can decide what scaling to apply.
Many image and audio processing processes are interactive - requiring manual assessment after each stage to work out what to do next. So in practice, sequential file processing its not going to happen for anything other than routine batch adjustments.
So take a big image, load into photoshop, make half a dozen filter operations on it and save it. Personally I would opt for having enough ram to hold the image, plus any other channels and history steps required to do it all in RAM every time, and not want to rely on the image temporary file paging.
If you can make it work equally fast off disk, then you are obviously wasted here, I am sure Adobe will pay good money for that capability.