Nexus: A Brief History of Information Networks from the Stone Age to AI
TLDR: “Interesting-ish”
This is about “information networks” – the sources of information that humans use to understand their world.
The author’s premise is that information is about connecting people, not necessarily communicating truth. We come together as communities through shared information. this information might be true or it might not, but the larger value is that it’s shared between us, and that brings us together.
The author talks about “intersubjective realities,” by which he means things that we believe together, whether or not they’re true. The author is an atheist, so he hammers on religion quite a bit. He claims that none if it is true, but that doesn’t matter, because a bunch of people believe it, so it serves as a way to bind us together into a community of believers.
Another example: fiat money. The only reason is works because because we shared the “intersubjective reality” that it’s worth something. In purely factual terms, it’s not – it’s just paper and metal and numbers. But because we all agree to live with the fiction that it has value, then it does have value.
The author also posits “the naive view of information,” which says that the more information someone has, the more get get to the truth. However, misinformation and disinformation, along with a lack of ability to understand the information someone is exposed to, combined with a lack of desire or determination to consume and analyze it all (we love confirmation bias), mean that we’re swimming in the information but lacking truth.
Good information networks are self-correcting. If there’s an error, some other source in the network corrects it. However, this has become a problem with the Internet, social media, and AI. Despite being exposed to more information than we could ever dream of, these systems tend to be self-perpetuating, not self-correcting. Both LLMs and our friends on Facebook tend to double-down on mistakes, rather than fix them.
The second half of the book is kind of a big polemic against AI and how we need to regulate it. I get the concern, but I don’t know that I was looking for it in a book like this.
It was interesting. He writes well. However, I do think I should read his prior books – Sapiens and Homo Deux before embarking on this one.
Book Info
- Author
- Yuval Noah Harari
- Year
- Pages
- 528
- Acquired
- I have not read this book yet.
- A hardcover copy of this book is currently in my home library.