i´m striving through the raw therapee forum.
they have really good attempts on integrating some kind of video (aimed at timelapse purposes) functionality.
really good ideas on gpu acceleration, open exr integration (oexr is a rock solid, free image format....whereas tiff is just a container format and can be everything, making it somewhat dangerous to work with) , debayering methods.... and so on.
what i understand in that forum, they really need people to code. just raw man power.
for me the raw recording function and dng extraction works stable and i would use it as a real world production workflow.
now that the first step is done, it would be a really great opportunity to keep magic lantern raw recording really open source and independent.
depending on black magic and hoping they implement and support the #1 pocket cinema killer..... well, it works somehow. but depending on the good will of a profit oriented firm could be the wrong idea.
the adobe way is not free and nobody knows what will happen next in the adobe cloud.
developing a true ML own workflow would be perfect. inventing the wheel from scratch, while there´s a great free tool.... it just feels wrong

but a cooperation with the raw therapee folks could be the best short cut to a true open source consistent and big name independent raw system.
because: raw means your image is finally created in post