Why is this interesting? It claims to be applicable to Fourier transforms, which is at the heart of LLR testing, and promises high speeds. Of course, this will be a case of we need to see how it works beyond the promises made at the moment, and software would need to be implemented to make use of it. Assuming it could be done, it could be orders of magnitude faster than any CPU/GPU we use right now. If the diagram in the video is representative, also expect the price to be much higher also. Still, it will be looked at as a value judgement and will either become successful or disappear into lost memories.