After having read Sapiens and Homo Deus, I was quite keen to see the melding of an historians viewpoint and this moment in history that we find ourselves in with the rise of AI.
Harari has pointed out in his earlier works that the capability that marks the rise of our species is the telling of stories. The capability to hold in the minds of many, the same narrative that explains your current circumstance and gives a guide on how to act.
He slightly reframes this observation in Nexus to cast the story as information and the institution that that curates it as an information network.
I like the fact that he points out that the study of history is not the study of what happened, but the study of change and therefore gives the historian a seat at the table with the futurists.
The book examines the struggles of these information network institutions, to maintain order through information flows. He makes clear the difference between information and truth. He examines the tension in how you organize information flows between distributed (democratic) information flows and centralized (authoritarian) information flows.
The book examines the changes in the information flows when some of the stories are created by non human intelligences, and warns that we have to be very deliberate as our culture is first aped and then maybe lead by an alien intelligence.
I found it a very valid contribution to the public discussion on AI. It does not attempt to explain technical details of AI. It does make lots of parallels to human organization problems down the ages and how information flows caused them or relieved them. I very much like the fact that Harari does not try and conflate consciousness and intelligence, and makes it clear that something can be trying to appear conscious without being conscious.
I think he fell short of examining human organization as an information system in and of itself. While we may find the stories and objectives that an alien intelligence adopts strange, I think we also find the myths, stories and objectives from one age to another are hard to fathom. For example, the tiny Micronesian island of Yap stores its wealth in immobile carved stones. We might find how they ascribe value hard to understand, and they would presumably find a defined benefit pension hard to understand.