Personally I would love to see an open source raw developer based on the technology. The results are outstanding not just for demosaicing, but for denoising, inpainting, deblurring, super-resolution and many other applications. There is a very significant gap in terms of image quality between these methods vs the usual single-image based approaches.
However the computational resources needed are so high it makes them impractical for an everyday use. It can take up to several hours -days for some algorithms- to process a single image.