Those were basic research, not 'we have a bunch of cool tech, let's glue it all together and go do crazy stuff'. If Google would pump money into materials science that would be Bell Labs, if they worked on next level interaction models that should count too, in that sense I think their general AI/ML work; speech recognition; synthesis and image classification work is more in line with what Bell & Xerox were all about than Makani, which is most re-using existing technology in ways that do not make much sense.
