Member Reviews
Harari's latest offering is an exploration of information networks and a cautionary warning against Artificial Intelligence. Overall, there were a lot of good nuggets of wisdom and thoughtfulness in this book, but it was overly long and repetitive, and some sections were a real slog to get through. I don't share Harari's sense of doom about AI, but I did appreciate his critiques. Honestly, I might have enjoyed the sections where he went off-topic more than the ones about the dangers of AI. Thank you to NetGalley and Random House Publishing Group for a digital review copy.
A super interesting premise. I enjoyed learning about how algorithmic our world is, and has been. The abstract idea of an algorithm has always confused and intrigued me. I didn’t realize how the idea of an algorithm can be applied to so many things. This book did feel denser and a bit less accessible than his other books.
This is a great history of information networks. I can’t say that Harari’s view of AI is positive and he points out a lot of potential pitfalls in terms of global politics. I love his writing style which I find highly conversational and fluid. The book was well-paced and I found it hard to put down. He never belabour points or get bogged down in minutiae. Importantly, Harari is very clear when he speculates. Thank you to Netgalley and Random House for the advance reader copy.
Yuval Noah Harari, the author of Sapiens: A Brief History of Humankind and Homo Deus: A Brief History of Tomorrow is back. This time, the focus is on the history of information networks and its impacts on humanity.
I really enjoyed Sapiens, but still need to read Homo Deus. As with Sapiens, the author covers different historical events as supporting examples of information networks at work, and this is still my favorite part. Yuval Noah Harari adds a different perspective to these historical events and I always learn something new about history.
Nexus is a fascinating look at the history of information networks, the wonderful and terrible ways it's been used, and how humanity is facing new challenges with AI. It's thought-provoking and definitely had me saying, "Huh. I hadn't thought about it like that" more than once.
"The only constant of history is change."
Thank you to Random House Publishing Group and NetGalley for a copy of the ARC in exchange for an honest review.
3.5 stars
I came into this reading experience somewhat biased, because Sapiens, one of the author's earlier books, was one of the first nonfiction books I actually read for fun. The accessible content, that's structured to educate without overloading the average reader, was something I appreciated at the time, and that still holds true with this latest book.
Harari is very talented at laying out his arguments within a broader narrative, which keeps the book moving at a pretty consistent and engaging pace. He also brings in a lot of historical contexts and ideas, which make the information more interpersonal and less just a long list of arguments.
However, the same issues that were present in Sapiens (even if I didn't pick up on them at the time) are present in Nexus. As much as I often found myself agreeing with the author's arguments, I had a hard time giving them too much weight because they so often lacked references. As much as Harari does still provide a massive bibliography, along with many in-text citations, they still somehow felt lacking at times. For example, he argued on multiple occasions about the belief systems of entire groups of people, without providing any examples or sources that this was what said group believed. I understand having to make generalizations, especially in a book of this scope, but some of them almost felt like fictionalized arguments used to prop up his own cases. Once again, I often actually agreed with his cases, but the fact that they were presented as fact, without adequate evidence, made me inclined to push back more against the author.
In terms of content, I feel that the subtitle is somewhat misleading. Out of 9 chapters, almost 6 focus exclusively on subjects related to AI, and so framing it as a broad history of information networks seems ambitious. The discussion of AI was informative and interesting, as he did not take a particular side beyond arguing about the need for more education on the subject. Often talk of AI will either fully defend it, or fully denounce it, and so I appreciated this nuanced discussion of both its potential and its faults. It also included a fair amount of history within the AI chapters, so there was still content on previous information networks, but I still wish more time had been dedicated to it.
Overall, I'd recommend this to anyone interested on the subject, even those (like me) who have a limited understanding of it. However, I would recommend coming into it with a fairly critical eye. As Harari argued, it is not truth that is power, but information, and as we continue this conversation on AI it is important to be able to separate the two.
Yuval Noah Harari's fourth book "Nexus" is an engaging and thought-provoking read that dives into the history of information networks and their impact on human civilization. The book covers a wide range of topics, from the early days of human communication to the complex information systems of today. As per his other books, Harari's writing style is both accessible and insightful, making this book suitable for readers familiar with his previous works and newcomers alike.
In "Nexus," Harari explores how information shapes our reality, emphasizing that most information isn't merely an attempt to represent reality—it actively creates new realities by connecting disparate elements. Errors, lies, fantasies, and fictions are all forms of information. Institutions must have self-correcting mechanisms to navigate this complex landscape. Harari also delves into the delicate balance between truth and order in government information networks, highlighting the challenges posed by modern technology and the rise of artificial intelligence.
The heart of the book lies in Harari's warning: As carbon-based life forms, we must grapple with silicon-based networks (AI) that can make decisions and generate ideas independently. Unlike previous inventions like the printing press or radio, AI possesses the ability to gather vast amounts of information and engage in total surveillance. Whether this development will be beneficial or detrimental remains uncertain, but it will undoubtedly be alien and fallible.
In summary, "Nexus" is a wake-up call, a warning, and a glimmer of hope all rolled into one. Harari's exploration of information networks challenges us to consider the future of our interconnected world.
Harari takes an insightful look at the technologies of communication, from language to printing to AI, and how they affect civilization. His examples include Copernicus, Napoleon, witch trials, Nazi Germany, Donald Trump, the current war between Gaza and Israel, major events in robotics, and much, much more. Along the way, he clearly explains some important and thought-provoking ideas about truth, democracy, and other big questions.
The book is about the history of information networks, but it has important lessons for the future of AI as well. It's beautifully written and very accessible. A great choice for book clubs!
I already reviewed this book before. An extremely important book to be read by everyone. I enjoyed it immensely and recommended it highly.
Yuval Noah Harari's 2014 book Sapiens: A Brief History of Humankind presents a theory of how humans went being just another species of apes to dominating the planet in terms of Cognitive, Industrial, and Scientific revolutions. One of Harari's key observations is that the religions, mythologies, lore, and national founding stories that human tell ourselves enable us to organize in increasingly greater collections and have been a major component to our dominance as a species.
Harari's follow up book in 2016, Homo Deus: A Brief History of Tomorrow, takes a similarly sweeping approach to how humankind will continue to evolve into the future. Among the many visions for the future that Harari develops is the theory of dataism, the valuation of information and data flow as ends in themselves rather than means to ends. Harari predicts that dataism will ultimately lead to humans giving algorithms control of their personal and societal lives, which will lead to a potentially different and even inhuman future.
Harari's current book looks at how the intersubjective reality of stories combines with the alien power that dataism gives to algorithms in our modern world. Nexus: A Brief History of Information Networks from the Stone Age to AI seeks to understand the unique opportunities and threats that artificial (or "alien", as Harari prefers to think of them) intelligence offers our current age. Harari counters a "naïve" view that information inevitably leads to truth by recognizing that information had the dual ends of truth and power. Harari illustrates these dual roles in the first part of the book where he shows how increasingly powerful information technologies allowed humans to evolve into increasingly complex societies. Harari refers to information technologies such as stories, which create an intersubjective reality, books, which allow societies to interact with a codified intersubjective reality, and error correction, where constitutional amendments allow these reality to gradually evolve and learn from errors of the pat.
Nexus next considers what is different about information systems in the modern world: that for the first time we have created systems not merely able to be modified by humans, but for the first time capable of making decisions for and modifying themselves based on additional data. Harari paints yet another uncertain vision of the future, this time looking at possibilities where algorithms and alien intelligence can both strengthen and threaten totalitarian control. Similar to the pose he takes in Homo Deus, Harari is wise enough not to draw conclusions about whether these trends are good or bad; clearly these are tools that will be both beneficial and harmful in myriad ways, and all that is certain is that humankind will come out substantially different because of what they enable.
As a historian, Harari is most helpful to the reader in identifying threads that course through long histories to understand the causes of our current situation and the risks that follow. This value makes Sapiens, which is primarily historical, satisfying in a way that Homo Deus and Nexus are not. Sapiens gave the reader a simple model through which to understand how qualities from humans' past have determine our present. In moving the story forward into the future, neither Homo Deus nor Nexus are as satisfying or as easy to summarize. Nonetheless, Harari's historical eye identifies key issues as being critical to determine what type of future we will inherit. In Nexus, Harari points to how important information systems have been in our development, and what a singular moment we are at today, when we find ourselves at the doorstep of a major change in the power of these systems, and in their power over us. The book is as insightful, intelligent, and potentially prescient as readers have come to expect from Harari.
While Harari's "Nexus" explores the history of information networks from the Stone Age to AI, its core focus lies in illuminating the power and potential dangers of Artificial Intelligence. The historical overview, while interesting, sometimes overshadows the book's primary concern: AI's profound implications for our future.
Harari effectively conveys the urgency of understanding AI's risks, drawing a compelling analogy between AI and nuclear weapons. He argues that an AI arms race is inevitable, with nations vying for dominance in AI capabilities. This competition poses a significant threat, as it would be difficult to monitor and control the advancements made by different countries. "Nexus" serves as a wake-up call, highlighting the need for informed decision-making by our leaders.
Having read Harari's previous works, "Sapiens" and "21 Lessons for the 21st Century," I found "Nexus" to be more closely aligned with the latter. While some readers might find the overlap repetitive, it doesn't detract from the book's overall message. Nexus does explain the benefits of AI, however, "Nexus" presents a fairly dystopian outlook. This perspective might be necessary to fully grasp the potential consequences of unchecked AI development.
I recommend "Nexus" to readers who are prepared to engage with its central message about AI's potential helpfulness along with its dangers. While the book could have been more concise, it offers valuable insights into a critical issue facing our society.
Sapiens by Yuval Noah Harari should be required reading for everyone. I don’t agree with everything Harari asserts in the massive nonfiction book, but it’s an eye-opening exploration of the themes within human history. There’s even a three-part graphic novel if you prefer illustrations instead of nearly 500 pages of solid text.
What you’ll learn from reading the book (or any of Harari’s other work) is that it’s a deeper exploration of not what happened in history, but why it happened and what conclusions we can draw about these motivations in our modern age.
Nexus (his next book releasing in September 2024) is the next logical progression from Sapiens. Building from the ideas presented about the human cognitive revolution, Harari delves into the spread of information and how they shape human narratives and societies.
I was particularly interested in this book because it extends the notion of storytelling as a uniquely human trait. Unlike other species on Earth, we can invent intangible things like governments, laws, companies, and money through stories. I’ve written several books about storytelling so these ideas deepened my appreciation for the power of human narratives.
Like Sapiens, Nexus is a lengthy and challenging read. At over 500 pages, Nexus rambles on about the history of bureaucracies, the differences between democracy and totalitarianism, and the fallibility of religious texts like The Bible. Once again, I don’t fully agree with every conclusion Harari draws within the book, but I’m intrigued by his arguments and appreciate the facts used to support them.
Nexus is an especially important book for our current time because it addresses two major global trends: the rise of populist authoritarian figures and the rise of Artificial Intelligence. Harari seeks to use historical lessons to point to how we can deal with these two potential threats.
Remember, Sapiens was written (originally in Hebrew) in 2011 — before Donald Trump, COVID-19, or Chat-GPT. The world has changed astoundingly in the 13 years since and Nexus feels like a quick update to account for these changes.
If Sapiens was a look at the past, Harari’s two other books — 21 Lessons For the 21st Century and Homo Deus — looked into the present and future, respectively. I’ve read both of these books as well, but with more mixed feelings. I had the biggest issues with Homo Deus and its bleak, almost-nihilistic look at how technology will warp humanity and transform us into gods.
Perhaps the biggest knock on Deus was the author’s confident assertion that we would never again have a global pandemic. This was in 2018, less than two years before it would be proved painfully wrong. It was only asserted once in the text, but it was a microcosm to show Harari’s own fallibility.
Reading Homo Deus was a disturbingly surreal experience that extended outside of Harari’s strength, which is history. Nexus is a more grounded version of this future-facing perspective. It does a better job of staying founded on historical examples and doesn’t reach as far into the unknown.
No one knows what will really happen with artificial intelligence (or alien intelligence, as Harari argues we should say). Many people are willing to guess with levels of accuracy and confidence. I appreciate this look at our collective future and how it weaves together disparate information to weave an interconnected look at the network of human information.
Nexus presents an interesting perspective on the history of human networks as a means of guessing what might happen with AI. It’s pure Harari — both the good and the bad of what that means. It’s worth reading and taking with a few grains of salt.
Note: I received a free advance copy of the book from NetGalley in exchange for a fair review, which you hopefully just read.
Fantastic look at the history of information networks. Cannot recommend enough, especially given the author's past titles.
Thanks to NetGalley and the publisher for the ARC of this fantastic book.
Yuval Harari is well-known by now, perhaps the best known of public historian/intellectuals. He is as much of a philosopher as a historian, and his books are interpretative more than descriptive. But the concepts that he uses to frame and interpret historical events and processes are fascinating, and require a lot of reflection as they as full of nuance.
The central contention here is that history, in the sense of actions taken by individuals that are reflected as actions taken by collections of individuals, are deeply shaped, if not directed, by information flows. There is a complex relationship between information, truth (defined as a representation of reality), and order. The two insights that I found most impactful are that more information does not necessarily mean more truth, and that those who are the mediators of information flow (the nexuses) have tremendous power.
The relevance of this to our current age has to do with the emergence of algorithms and AI. For the first time, the nexuses are becoming non-human, and may be guided by considerations that are alien. The actions of AI, and Harari’s contention is that AI does make decisions and does take action, are not necessarily aligned with human goals, and do not necessarily wind up serving the greater good of humanity. In fact, Harari seems quite pessimistic that they will do so unless carefully shaped and directed here in their infancy. The future may be utopia, but there are many reasons to be concerned about dystopia, instead.
I find the arguments quite compelling. And one is not in the least be reassured knowing about things like the Chinese total surveillance state employed against the Xighur and the Iranian monitoring of the hijab law (which I was unaware of). Of course, there are those who may argue that if one is doing nothing wrong, what is there to fear. But the description of social credit scores, where one is assigned points when one behaves in conformity with an external set of standards, means no-one gets to opt out. And since the algorithm is inscrutable, a situation like Kafka’s trial is a very possible result.
So the idea here is that we don’t even need a superintelligent AI overload to get into deep trouble. AI as it currently exists can become highly destructive and dangerous even without further technical evolution. Most people are not paying attention to this, and most voters will not read this book. But they should.
Yuval Noah Harari's "Nexus: A Brief History of Information Networks from the Stone Age to AI" attempts to trace the evolution of information systems through human history. While ambitious in scope, the book falls short in several key areas, revealing the limitations of Harari's approach to complex historical and technological topics.
Misleading Structure and Focus
Despite its subtitle promising a "Brief History of Information Networks from the Stone Age to AI," "Nexus" dedicates an disproportionate amount of space to artificial intelligence. Nearly six out of nine chapters focus exclusively on AI, leaving the actual historical analysis feeling rushed and superficial. This imbalance may disappoint readers expecting a more comprehensive historical account.
Lack of Depth and Rigour
Harari's trademark sweeping statements and broad generalisations, while engaging, often lack the necessary academic rigour. Many of his claims about historical belief systems or societal structures are presented without adequate evidence or citation. This approach, while making for easy reading, undermines the credibility of his arguments and may mislead readers unfamiliar with the topics.
Oversimplification of Complex Issues
In his attempt to make complex ideas accessible, Harari frequently oversimplifies nuanced topics. His discussion of AI, while highlighting important concerns, often lacks the technical depth necessary for a truly insightful analysis. This superficiality is particularly problematic given the book's heavy focus on AI and its potential impacts.
Repetition and Verbosity
At close to 450 pages, "Nexus" suffers from significant bloat. Harari often meanders around points, including tangential anecdotes and repetitive explanations that add little to the core argument. This verbosity not only tests the reader's patience but also dilutes the impact of his more salient points.
Lack of New Insights
For readers familiar with Harari's previous works, "Nexus" may feel disappointingly familiar. Many of the ideas and even specific examples appear recycled from "Sapiens" and "Homo Deus," raising questions about the book's originality and value for longtime Harari readers.
Pessimistic Bias
Harari's treatment of AI is notably pessimistic, focusing heavily on potential dangers while giving short shrift to potential benefits. This imbalance may stem from his stated intention to counterbalance optimistic narratives from AI proponents, but it results in a skewed perspective that may unduly alarm readers.
Neglect of Important Contexts
In his focus on information networks, Harari often neglects crucial political, economic, and social contexts that shape these systems. This oversight leads to an incomplete analysis, particularly when discussing contemporary issues like the erosion of democratic information networks.
Conclusion
While "Nexus" showcases Harari's talent for engaging writing and big-picture thinking, it ultimately disappoints as a rigorous historical analysis or a nuanced examination of AI's potential impacts. The book's structural imbalances, lack of academic rigour, and tendency towards oversimplification undermine its credibility and utility.
Readers seeking a comprehensive history of information networks or a balanced analysis of AI's potential future impacts may need to look elsewhere (e.g. “The Information: A History, a Theory, a Flood” by James Gleick). While "Nexus" might serve as a starting point for discussions about the role of information in society, it should be approached with a critical eye and supplemented with more specialised, rigorous sources.
In his last two books and also this book, the author proves that he only has a few genuine ideas that are being repeated. The rest of the book act as a filler to promote those few ideas. Overall, I think this author has gained more attraction than his ideas are worth studying.
I tried to like this book, but unfortunately, with the quick way things are shifting with AI, and with the glaring missing information and context about the true cost of things like algorithms and storage, this book was ultimately a DNF for me. 2.5/5
I am a fan of his prior books and generally, how his mind works. This book falls in line with the others - thoughtful and brilliant. It’s a deep dive into society, networks, and the human condition. It’s also about how we view information and believe the stories around it. Harari straddles the line perfectly between informational and conversational.
Definite recommendation.
I don’t envy anyone trying to write a book about AI right now. The field is so fast-moving and riddled with hype that sorting through it all (and not having your writing become instantly outdated) has to be an immense challenge. Hats off to Nexus for taking on the challenge.
I enjoyed Harari's historical approach to this book—situating AI within the context of other information networks and revolutions across time illuminated how AI extends from those traditions and may deviate from them. The writing is engaging, the research is thorough, and the examples are compelling. Harari makes a strong case for thinking about information in terms of networks and taking AI risks seriously.
That said, Nexus contained some surprising omissions. Harari spends quite a bit of time discussing theoretical data colonialism but never discusses the fact that many AI models are built on stolen materials. The hypothetical future of data extraction he worries about is already here for the writers, artists, creators, and everyday people whose content has been appropriated by AI companies for profit, without compensation. It seems irresponsible not to discuss these topics.
Harari also mentions that algorithms don’t take up much space. That may be true—but data centers do, and AI requires an extraordinary amount of energy. Amid Harari’s many discussions of how AI could lead to an apocalypse, it again seems strange to have drawn no connections between AI and energy use/climate change.
The discussions of AI and fallibility also don’t focus on AI hallucinations as much as a reader might expect (especially given Harari’s emphasis on understanding that more information doesn’t necessarily move us closer to the truth). Harari does a great job unpacking the risks and potential consequences of algorithmic bias but doesn’t address AI’s weirdness/lies/fiction in nearly as much detail. He cites an example of AI giving better medical advice than doctors in experimental conditions—what about everyday uses of AI encouraging people to put glue on pizza or making up citations to academic texts that don’t exist? Some of the current limitations of AI are hand-waved away, and it paints a more threatening picture than is perhaps warranted.
Overall, an interesting read but not always a convincing one.
I can't express how much I enjoyed this book. The history behind how we have reach the point in technology of AI was captivating. I couldn't put this book down. I am looking forward to exploring more work from this author.
I consider myself extremely fortunate to have discovered Yuval Noah Harari's books albeit quite by accident. When an acquaintance of mine texted a mutual friend and myself expressing concern about her college age son who was interested in reading his book Sapiens and not understanding why she was "concerned", I checked it out. So, after doing a bit of research, I knew Sapiens was a must read.
All.that to say, is that I have read (and loved) all.of his books and Nexus is no exception.
If you have read any or all of Harari's books Nexus will not disappoint...in fact it may very well be his best.
I admire how he tackles issues from a much different perspective than most...a true breath of fresh air fro the talking heads on the news channels.
His spin on information,.bureacracy, stories, Nationalism and AI and how each affects and could possibly affect our future is at once an eye opener and, dare I say, a bit worrisome (to say the least)
This is an important book and a must read. I highly recomend.Nexus and all of his books.