Review: How to Create a Mind

How would you go about creating a human brain? Do we know enough about the brain that we could – at least theoretically – create something comparable with technology?
These are the kinds of questions explored in this interesting book by Ray Kurzweil, and my first read of 2025.
Kurzweil’s book narrows in on the possibilities of technology to create an equivalence of the human brain.
Written just over 10 years ago, it was fascinating to consider how many of Kurzweils predictions for the past decade (and the next) have already progressed. Kurzweil has been writing on the subject of technological advancement since the 1980’s, and has made many predictions – impressively accurately – about what he thought might be achieved in the subsequent decades. He has been at the cutting edge of AI research and development, well before most people had any idea what it was, or what it could do (by today’s standards).
The book spends a fair amount of time outlining how brains store, sort and process information through hierarchical patterns. It’s quite interesting, and important for laying the groundwork to how technology can mimic how the brain functions. There’s a few technical concepts, but I found it relatively easy to follow. (It’s possible that may not be the case for you if all of this is very unfamiliar to you.)
Along the way, Kurzweil shares various stories of his own career, and important developments he has been involved with (or that have been significant for the AI industry as a whole). They are interesting, and help keep the book from becoming too abstract.
Who wrote it
Ray Kurzweil has been working in the tech industry for decades, chiefly in roles around front-line AI research and development. He was hired by Google in 2012 to bring “natural language understanding” to Googles products.
Why I read it
After reading (and thoroughly enjoying) Lisa Genova’s book on memory, I was keen to explore some of the ideas further.
What I liked
I enjoyed the explanation of how the brain categorises and stores memories as series of patterns. It was a very accessible presentation of a complex concepts. I appreciated the personal stories and the historical stories that helped to paint a picture of how we have arrived at the present day.
I appreciated that the book had a narrow focus and didn’t stray from it. It made it easy to follow and stay interested.
I was intrigued by the final chapter which presented some of his predictions for the next twenty years (which we are halfway through already) and beyond. Hearing him describe AI chatbots in 2014, and seeing where we are now, it is quite remarkable the progress that continues to be made in this space.
What I didn’t
There wasn’t much not to like. There were a few moments it was hard to stay focussed, but overwhelmingly an enjoyable read. Some of his future predictions seemed a little out there, but then I’m sure his 1980s predictions would have equally sounded ridiculous, and much of that proved correct.
Major Takeaway
There was a few interesting takeaways.
- You don’t actually store images in your mind as memories. You store many pieces of small data, which your mind puts together when you need to recall something. Effectively, it is recreating the scene for you. This is why (as Lisa Genova’s book also showed) every time you remember something, you are actually creating a new memory, not merely retrieving an archived replica of an old one.
- AI as we know it today (ChatGPT and the like) has been in slow development for may decades. Though it feels like it sprung up overnight, it’s taken a very long time to get here.
- Our minds have much redundancy in our memories. This is extremely helpful, as it means that there are multiple ways you can retrieve the information you search for in your mind. It also means that, when one aspect gets overwritten or discarded, there are still other parts of your brain storing the same information.
- While different parts of your brain are dedicated to different tasks (speech, etc), those who are born with significant parts of their brain missing can use other parts as an effective replacement. This is much easier if you are born with such an abnormality. While it is possible if you lose part of your brain, at say 40 years old, for other parts of your brain to step up, in reality, you would be learning these things like a newborn. It might take you another 40 years to get back to where you were. And that’s best case scenario.
- There was a fascinating experiment described where different words and images were projected to either the left or right eye/ear of an individual. And then they had to make a response to what they saw/heard. The trick of the experiment was that the choice people made was effectively unconscious, as they had separated the hemispheres of the brain. The result was that people made a choice without their brain fully recognising that choice, and then they shifted to make up an explanation for their choice, rather than being able to identity that they hadn’t made the choice. (This is a terrible summary of a very interesting story. Apologies)
Who should read it
Worth a read if you have any interest in the mind, history and development of AI, memory, or learning.
How to Create a Mind
How would you go about creating a human brain? Do we know enough about the brain that we could - at least theoretically - create something comparable with technology? These are the kinds of ...