What do you think?
Rate this book
How AI Thinks: How we built it, how it can help us, and how we can control it
THE #2 TIMES BESTSELLER
'Artificial intelligence is going to have a massive impact on everyone’s lives... an accessible and sensible read that helps demystify AI' Deborah Meaden, entrepreneur and star of Dragon's Den
'Nigel Toon is a visionary leader in the field of artificial intelligence... a must-read' Marc Tremblay, Distinguished Engineer, Microsoft
Those who understand how AI thinks are about to win big.
We are used to thinking of computers as being a step up from calculators - very good at storing information, and maybe even at playing a logical game like chess. But up to now they haven't been able to think in ways that are intuitive, or respond to questions as a human might. All that has changed, dramatically, in the past few years.
Our search engines are becoming answer engines. Artificial intelligence is already revolutionising sectors from education to healthcare to the creative arts. But how does an AI understand sentiment or context? How does it play and win games that have an almost infinite number of moves? And how can we work with AI to produce insights and innovations that are beyond human capacity, from writing code in an instant to unfolding the elaborate 3D puzzles of proteins?
We stand at the brink of a historic change that will disrupt society and at the same time create enormous opportunities for those who understand how AI thinks. Nigel Toon shows how we train AI to train itself, so that it can paint images that have never existed before or converse in any language. In doing so he reveals the strange and fascinating ways that humans think, too, as we learn how to live in a world shared by machine intelligences of our own creation.
'Artificial intelligence is going to have a massive impact on everyone’s lives... an accessible and sensible read that helps demystify AI' Deborah Meaden, entrepreneur and star of Dragon's Den
'Nigel Toon is a visionary leader in the field of artificial intelligence... a must-read' Marc Tremblay, Distinguished Engineer, Microsoft
Those who understand how AI thinks are about to win big.
We are used to thinking of computers as being a step up from calculators - very good at storing information, and maybe even at playing a logical game like chess. But up to now they haven't been able to think in ways that are intuitive, or respond to questions as a human might. All that has changed, dramatically, in the past few years.
Our search engines are becoming answer engines. Artificial intelligence is already revolutionising sectors from education to healthcare to the creative arts. But how does an AI understand sentiment or context? How does it play and win games that have an almost infinite number of moves? And how can we work with AI to produce insights and innovations that are beyond human capacity, from writing code in an instant to unfolding the elaborate 3D puzzles of proteins?
We stand at the brink of a historic change that will disrupt society and at the same time create enormous opportunities for those who understand how AI thinks. Nigel Toon shows how we train AI to train itself, so that it can paint images that have never existed before or converse in any language. In doing so he reveals the strange and fascinating ways that humans think, too, as we learn how to live in a world shared by machine intelligences of our own creation.
305 pages, Kindle Edition
Published February 8, 2024
103 people are currently reading
438 people want to read
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 19 of 19 reviews
April 19, 2025
“Can you adjust the tone of the following book review to be more considerate of the author’s feelings were they to read it, while retaining the critique and criticism of their work?”
“Absolutely. Here’s a version of your review with a more considerate tone, aimed at preserving the core critiques while softening the language and avoiding any ad hominem implications. It offers the same thoughtful analysis, just in a way that the author might find more constructive:”
This book is an accessible and at times thought-provoking introduction to an important and fast-evolving field. It raises big, timely questions and is clearly written with a broad audience in mind. For readers entirely new to the terrain, it may serve as a useful jumping-off point. However, its ambitions outpace its depth, and for anyone with even moderate familiarity with the topics discussed, the book is likely to feel frustratingly superficial.
The treatment of core concepts—like intelligence and information—is where this becomes most evident. The attempt to define these terms slides into circular reasoning, seemingly without the author’s awareness, and without engaging with the rich body of existing literature that has wrestled with these definitions for decades. The definitions offered come across more like quick personal takes than the result of sustained engagement or original insight.
One moment that particularly jarred was the invocation of Einstein’s famous aphorism about problem definition, followed by an example using Moore’s Law. The author first presents Moore’s Law as illustrative of a defined problem, then immediately calls it a prediction and an industry challenge. The internal contradiction here is never resolved, and it typifies a kind of glibness that runs through much of the book’s reasoning.
Similar issues appear elsewhere. The chapter on consciousness glosses over the ‘hard problem’ with a take that is, frankly, just wrong. The discussion of Bayes’ Theorem misrepresents both its logic and its significance. In areas where I have more domain knowledge, the author’s approach often seemed to cherry-pick or lean heavily on sources that aligned with their thesis, without seriously engaging counterarguments or complexity. That in turn made me cautious about accepting the book’s claims in areas where I was less well-read.
Most worryingly, there is a pervasive and unexamined human exceptionalism when it comes to volition, motive, ‘free will’, and our supposed capacity for ethical self-regulation—all of which lends a troubling naivety to the book’s treatment of AI risk. Its tone is breezily reassuring, but for readers who are more skeptical of humanity’s track record of acting in our own best interests—or who are attuned to the increasingly tenuous distinction between biological and silicon-based information processing—that reassurance is unlikely to land. In fact, given the author’s influential role within the AI industry, it may have quite the opposite effect.
Analogies to the domestication of wolves or early industrial-era anxieties feel similarly misplaced. They seem designed to cast concern about AI as little more than modern-day Luddism. But it’s entirely possible—and necessary—to recognise the potential benefits of AI while remaining clear-eyed about its risks. History has shown how transformative technologies can bring harm alongside progress, especially when their unintended consequences are dismissed in the name of optimism.
Perhaps most disappointing is the mismatch between the book’s confident tone and the author’s background. I expected a deeper foundation in philosophy, cognitive science, or technical AI. Instead, the author appears to come from a primarily entrepreneurial and policy-facing background. That’s not inherently disqualifying, but it does help explain the book’s tendency to favour broad statements over close reasoning. In a moment when public understanding and policy are being shaped by books like this, it’s crucial to distinguish between influence and expertise.
That said, books that provoke strong disagreement can still be valuable. This one certainly made me reflect more deeply, prompted some fact-checking, and encouraged me to revisit and expand my understanding of certain concepts. For all its flaws, I’m glad I read it—just not for the reasons the author might have hoped.
“Thanks!”
“Absolutely. Here’s a version of your review with a more considerate tone, aimed at preserving the core critiques while softening the language and avoiding any ad hominem implications. It offers the same thoughtful analysis, just in a way that the author might find more constructive:”
This book is an accessible and at times thought-provoking introduction to an important and fast-evolving field. It raises big, timely questions and is clearly written with a broad audience in mind. For readers entirely new to the terrain, it may serve as a useful jumping-off point. However, its ambitions outpace its depth, and for anyone with even moderate familiarity with the topics discussed, the book is likely to feel frustratingly superficial.
The treatment of core concepts—like intelligence and information—is where this becomes most evident. The attempt to define these terms slides into circular reasoning, seemingly without the author’s awareness, and without engaging with the rich body of existing literature that has wrestled with these definitions for decades. The definitions offered come across more like quick personal takes than the result of sustained engagement or original insight.
One moment that particularly jarred was the invocation of Einstein’s famous aphorism about problem definition, followed by an example using Moore’s Law. The author first presents Moore’s Law as illustrative of a defined problem, then immediately calls it a prediction and an industry challenge. The internal contradiction here is never resolved, and it typifies a kind of glibness that runs through much of the book’s reasoning.
Similar issues appear elsewhere. The chapter on consciousness glosses over the ‘hard problem’ with a take that is, frankly, just wrong. The discussion of Bayes’ Theorem misrepresents both its logic and its significance. In areas where I have more domain knowledge, the author’s approach often seemed to cherry-pick or lean heavily on sources that aligned with their thesis, without seriously engaging counterarguments or complexity. That in turn made me cautious about accepting the book’s claims in areas where I was less well-read.
Most worryingly, there is a pervasive and unexamined human exceptionalism when it comes to volition, motive, ‘free will’, and our supposed capacity for ethical self-regulation—all of which lends a troubling naivety to the book’s treatment of AI risk. Its tone is breezily reassuring, but for readers who are more skeptical of humanity’s track record of acting in our own best interests—or who are attuned to the increasingly tenuous distinction between biological and silicon-based information processing—that reassurance is unlikely to land. In fact, given the author’s influential role within the AI industry, it may have quite the opposite effect.
Analogies to the domestication of wolves or early industrial-era anxieties feel similarly misplaced. They seem designed to cast concern about AI as little more than modern-day Luddism. But it’s entirely possible—and necessary—to recognise the potential benefits of AI while remaining clear-eyed about its risks. History has shown how transformative technologies can bring harm alongside progress, especially when their unintended consequences are dismissed in the name of optimism.
Perhaps most disappointing is the mismatch between the book’s confident tone and the author’s background. I expected a deeper foundation in philosophy, cognitive science, or technical AI. Instead, the author appears to come from a primarily entrepreneurial and policy-facing background. That’s not inherently disqualifying, but it does help explain the book’s tendency to favour broad statements over close reasoning. In a moment when public understanding and policy are being shaped by books like this, it’s crucial to distinguish between influence and expertise.
That said, books that provoke strong disagreement can still be valuable. This one certainly made me reflect more deeply, prompted some fact-checking, and encouraged me to revisit and expand my understanding of certain concepts. For all its flaws, I’m glad I read it—just not for the reasons the author might have hoped.
“Thanks!”
August 15, 2025
3.5 stars.
The main attraction that made me pick up this book was the author’s background as the founder and CEO of Graphcore, the company that makes AI semiconductor. While everyone is thinking and talking about what AI can do and the impact it has on humanity, I find it interesting to deep dive into the hardware that makes AI possible to learn the way it does. AI semiconductor is basically a learning brain, which is designed specifically to process micro tasks involved in learning, just like the neural systems in our brains.
Unfortunately this book spent a lot of time relaying the history of computers and internet - while relevant and might provide great introduction into the concept of computing operations and powers, it took a bit of time before we finally reached the AI point. At the end of the day, I didn’t learn too much new information beyond what I had known about AI.
Having said that, this would be a great introduction of AI to non technical audience.
The main attraction that made me pick up this book was the author’s background as the founder and CEO of Graphcore, the company that makes AI semiconductor. While everyone is thinking and talking about what AI can do and the impact it has on humanity, I find it interesting to deep dive into the hardware that makes AI possible to learn the way it does. AI semiconductor is basically a learning brain, which is designed specifically to process micro tasks involved in learning, just like the neural systems in our brains.
Unfortunately this book spent a lot of time relaying the history of computers and internet - while relevant and might provide great introduction into the concept of computing operations and powers, it took a bit of time before we finally reached the AI point. At the end of the day, I didn’t learn too much new information beyond what I had known about AI.
Having said that, this would be a great introduction of AI to non technical audience.
July 30, 2024
Inspiring summary of the AI modus operandi and its contemplated impact upon humanity. Interesting comparative examples of other revolutions which impacted humanity with concrete references on several domains of human life. If you are into AI and its impact or you want to start to know more, this book is for you as it simply written but lack some level of audacity.
August 1, 2025
This book offers a rather unserious engagement with how AI thinks, it lacks insight for someone already familiar with the topic, but I think also as an introductory text, it jumps between concepts and analogies in a way that would just be confusing and not elucidating. And the explanations to me signal lack of deep engagement with the topics.
It rather just offers some light narrative stories about AI which isn't worthless, gives some space for ones own thoughts, but the information contained is not elucidatory. As one positive though I will say I liked the paragraph in the introduction when he rather accurately described the experience of reading the book which made me excited to look for serious engagement with the topics of consciousness and its intersection with AI
Rather than explaining my issues in the abstract I will simply paste a couple critiques that jumped at me while reading the book till page 38. (I wrote these notes in obsidian for myself that's why the formatting is what it is)
p.20
- "GPT-3 has around 175 billion paramaters — which are equivalent to the synaptic connections that structure information in your brain"
- equivalent, not analogous? Very very sloppy sentence signals unseriousness
p. 34
- "The hidden layers of neurons learn to recognize a hierarchy of generalized features (called 'parameters'). These parameters are filtered from the input information and are then mixed and filtered again to build up even higher-level parameters."
- Parameters are the weights and biases, not the feature representations, non-standard and misleading
p. 35
- Talks about the human tendency to ad hoc reason an untrue justification for an action, and then continues that this is exemplified by a human training tennis and improving their technique by iteration
- The tennis analogy is fine introductory analogy for backprop ML techniques, but it doesn't elucidate the tendency of humans for ad-hoc reasoning.
p. 36
- Introduces the concept of inductive reasoning by an example of deciding on what apartment to rent with imperfect information, but this seems much more like probabilistic reasoning (could be better used to introduce bayesian reasoning) than induction, doesn't really give a good intuition of what is induction, but I'd let this slide
p. 38
- "A modern AI system that is able to beat a world champion at chess or Go can't deduce what moves are possible: instead it learns how to develop a winning strategy.
This mathematical complexity problem was explained by Kurt Gödel in his 'incompleteness theorems'"
- This is what prompted me to write this review, the engagement here with what is incompleteness theorem is borderline misinformation, just smashing together exponentially increasing search spaces and incompleteness theorem.
- The author then continues that "Turing showed that a machine applying a sequence of instructions to a set of information can perform any operation, but that some operations might take an infeasibly long amount of time. Gödel called this problem the 'Entschedungsproblem', or the 'decision problem'"
- again, undeniably misleading. It isn't about them taking infeasibly long, but that they cannot be decided at all
It rather just offers some light narrative stories about AI which isn't worthless, gives some space for ones own thoughts, but the information contained is not elucidatory. As one positive though I will say I liked the paragraph in the introduction when he rather accurately described the experience of reading the book which made me excited to look for serious engagement with the topics of consciousness and its intersection with AI
Rather than explaining my issues in the abstract I will simply paste a couple critiques that jumped at me while reading the book till page 38. (I wrote these notes in obsidian for myself that's why the formatting is what it is)
p.20
- "GPT-3 has around 175 billion paramaters — which are equivalent to the synaptic connections that structure information in your brain"
- equivalent, not analogous? Very very sloppy sentence signals unseriousness
p. 34
- "The hidden layers of neurons learn to recognize a hierarchy of generalized features (called 'parameters'). These parameters are filtered from the input information and are then mixed and filtered again to build up even higher-level parameters."
- Parameters are the weights and biases, not the feature representations, non-standard and misleading
p. 35
- Talks about the human tendency to ad hoc reason an untrue justification for an action, and then continues that this is exemplified by a human training tennis and improving their technique by iteration
- The tennis analogy is fine introductory analogy for backprop ML techniques, but it doesn't elucidate the tendency of humans for ad-hoc reasoning.
p. 36
- Introduces the concept of inductive reasoning by an example of deciding on what apartment to rent with imperfect information, but this seems much more like probabilistic reasoning (could be better used to introduce bayesian reasoning) than induction, doesn't really give a good intuition of what is induction, but I'd let this slide
p. 38
- "A modern AI system that is able to beat a world champion at chess or Go can't deduce what moves are possible: instead it learns how to develop a winning strategy.
This mathematical complexity problem was explained by Kurt Gödel in his 'incompleteness theorems'"
- This is what prompted me to write this review, the engagement here with what is incompleteness theorem is borderline misinformation, just smashing together exponentially increasing search spaces and incompleteness theorem.
- The author then continues that "Turing showed that a machine applying a sequence of instructions to a set of information can perform any operation, but that some operations might take an infeasibly long amount of time. Gödel called this problem the 'Entschedungsproblem', or the 'decision problem'"
- again, undeniably misleading. It isn't about them taking infeasibly long, but that they cannot be decided at all
July 1, 2025
***We should be terrified of what the future holds***
The striking thing about this book was understanding that AI has been developed with the human brain as the base model. This seems counter productive because humans follow biological intelligence, meaning we are intelligent in a way that helps us to adapt and survive. If AI has also been built in this way, who is to say that it won't adapt by re-writing its own code in order to survive if a human decides to shut it down or not use it anymore. This a very real potential threat of machine learning and how AI adapts and learns. There's many parts to understanding and using AI safely and this book was a very comprehensive overview of all the different sides of the discussion. One thing I strongly agree with is that people who work in technical roles to develop AI should be mandated to pass educational courses in ethics or history/economics so that AI can be built on safe, sustainable principles. I believe developers have a certain line of thinking that is linked to (and limited to) innovation, whereas the business people that they work for are motivated by one thing only - profit. As someone who advocates for responsible consumption, AI poses a huge risk against the idea of humans being able to stay curious, creative and possess critical thinking skills because nowadays AI is seen as a one stop shop for all the answers. This is wildly problematic because AI is trained on human data, some of which is sourced from the internet and it is programmed with probabilistic models to spot trends in the data. Do you see the problem? It means that AI is just parroting back to us, that which we already know. And humans get thing wrong all the time! If you need to make a resignation letter or a breakup text more polite, then AI can assist because humans have already done this correctly many times before. But, if you are lonely and are getting emotionally comforted by ChatGPT then this should alarm you because humans have not biologically evolved to connect to mere simulated human responses from a machine. We cannot expect vulnerable people with fragile mental states or those who are experiencing difficult periods of life to not want that simulated experience of someone caring about them. In this example, a licensed, experienced, trained and most importantly, human, therapist would not get it wrong, but AI definitely could. AI is like that really smart but overconfident friend who can never accept they are wrong. There are unprecedented risks that governments need to get ahead of, otherwise AI will be death of the fundamental human experience. AI was created by humans and therefore humans hold the responsibility to safe, regulated AI that should always benefit humankind. This book was so informative and clearly demonstrated the different risks and arguments for and against AI and there is no group of people that I wouldn't recommend this book to. Happy reading!
The striking thing about this book was understanding that AI has been developed with the human brain as the base model. This seems counter productive because humans follow biological intelligence, meaning we are intelligent in a way that helps us to adapt and survive. If AI has also been built in this way, who is to say that it won't adapt by re-writing its own code in order to survive if a human decides to shut it down or not use it anymore. This a very real potential threat of machine learning and how AI adapts and learns. There's many parts to understanding and using AI safely and this book was a very comprehensive overview of all the different sides of the discussion. One thing I strongly agree with is that people who work in technical roles to develop AI should be mandated to pass educational courses in ethics or history/economics so that AI can be built on safe, sustainable principles. I believe developers have a certain line of thinking that is linked to (and limited to) innovation, whereas the business people that they work for are motivated by one thing only - profit. As someone who advocates for responsible consumption, AI poses a huge risk against the idea of humans being able to stay curious, creative and possess critical thinking skills because nowadays AI is seen as a one stop shop for all the answers. This is wildly problematic because AI is trained on human data, some of which is sourced from the internet and it is programmed with probabilistic models to spot trends in the data. Do you see the problem? It means that AI is just parroting back to us, that which we already know. And humans get thing wrong all the time! If you need to make a resignation letter or a breakup text more polite, then AI can assist because humans have already done this correctly many times before. But, if you are lonely and are getting emotionally comforted by ChatGPT then this should alarm you because humans have not biologically evolved to connect to mere simulated human responses from a machine. We cannot expect vulnerable people with fragile mental states or those who are experiencing difficult periods of life to not want that simulated experience of someone caring about them. In this example, a licensed, experienced, trained and most importantly, human, therapist would not get it wrong, but AI definitely could. AI is like that really smart but overconfident friend who can never accept they are wrong. There are unprecedented risks that governments need to get ahead of, otherwise AI will be death of the fundamental human experience. AI was created by humans and therefore humans hold the responsibility to safe, regulated AI that should always benefit humankind. This book was so informative and clearly demonstrated the different risks and arguments for and against AI and there is no group of people that I wouldn't recommend this book to. Happy reading!
June 25, 2025
This book is a great introduction to AI and all the possibilities that it holds, I’ve used AI tools in the past but been able to learn about the intricacies of how they were developed and the history behind them was very interesting. I also liked when the book delved into the more emerging technologies like Quantum and Molecular Computing.
It was great to learn more about the human body and see how much AI engineers have tried to replicate the human body in the creation of Neural Artificial Networks and the comparisons between what the current best AI software can do and how it still is so much worse than the human body
I liked the books distinction between intelligence and consciousness and it is very reassuring to be told (and have it broken down) exactly why we are right to be scared of some of the things AI could do but we don’t need to worry about the way the films portray it because it will always be based on human purpose
It was also great that the book was willing to touch on the genuine challenges that AI will bring rather than pretending it’s a perfect solution. There are genuine concerns about malicious actors using it as well as how it can entrench our own human biases (both conscious and unconscious)
Rating - 9/10
It was great to learn more about the human body and see how much AI engineers have tried to replicate the human body in the creation of Neural Artificial Networks and the comparisons between what the current best AI software can do and how it still is so much worse than the human body
I liked the books distinction between intelligence and consciousness and it is very reassuring to be told (and have it broken down) exactly why we are right to be scared of some of the things AI could do but we don’t need to worry about the way the films portray it because it will always be based on human purpose
It was also great that the book was willing to touch on the genuine challenges that AI will bring rather than pretending it’s a perfect solution. There are genuine concerns about malicious actors using it as well as how it can entrench our own human biases (both conscious and unconscious)
Rating - 9/10
August 21, 2024
Paints a less apocalyptic/dramatic picture of the future of AI. He offers a more positive view which covers the various areas in which AI can hugely improve our quality of life.
Toon is of the opinion that a Super-intelligence explosion is unlikely for various reasons; the risks of AI are not the AIs themselves but human misuse of them. He discusses the risks and the ways in which they could be regulated.
I found the historical context into which he puts AI to be particularly interesting. In his telling, the earlier impact of the Industrial Revolution upon Europe and the US is a key reason they overtook China to become the World's wealthiest and most powerful nations into the 20th century (apparently known as "the century of humiliation" in China).
China has embraced AI and been very proactive in innovation (and surprisingly to me, AI regulation), seeing the opportunity to embed a similar historic advantage.
Toon sees the biggest risk to the West as not so much an AI apocalypse, but rather a loss of economic competitiveness and the attached social and security advantages that many in the West take for granted.
Toon is of the opinion that a Super-intelligence explosion is unlikely for various reasons; the risks of AI are not the AIs themselves but human misuse of them. He discusses the risks and the ways in which they could be regulated.
I found the historical context into which he puts AI to be particularly interesting. In his telling, the earlier impact of the Industrial Revolution upon Europe and the US is a key reason they overtook China to become the World's wealthiest and most powerful nations into the 20th century (apparently known as "the century of humiliation" in China).
China has embraced AI and been very proactive in innovation (and surprisingly to me, AI regulation), seeing the opportunity to embed a similar historic advantage.
Toon sees the biggest risk to the West as not so much an AI apocalypse, but rather a loss of economic competitiveness and the attached social and security advantages that many in the West take for granted.
May 4, 2025
After reading this book I still don’t know how AI thinks.
Couldn’t help but feel the book glosses over many concepts of AI, which made it read like a lit review of the How AI came about and the current use cases for AI. I could see how this book would be mind blowing for a someone who has never heard of AI at all and is only introduced to it through the book; but for anyone else that has a basic understanding and is looking for something more, this book will put you to sleep.
Parallels and analogies felt weirdly out of place. And not to mention that the constant reference to “intelligence” lowkey started to get on my nerves. Certain concepts are repeated ad nausem, whereas other points of contention such as ethics, risks and governance of AI seem raw and not fully explored.
Overall, the book was ok as a primer into AI, but otherwise I’d rather put it into ChatGPT for a summary instead.
Couldn’t help but feel the book glosses over many concepts of AI, which made it read like a lit review of the How AI came about and the current use cases for AI. I could see how this book would be mind blowing for a someone who has never heard of AI at all and is only introduced to it through the book; but for anyone else that has a basic understanding and is looking for something more, this book will put you to sleep.
Parallels and analogies felt weirdly out of place. And not to mention that the constant reference to “intelligence” lowkey started to get on my nerves. Certain concepts are repeated ad nausem, whereas other points of contention such as ethics, risks and governance of AI seem raw and not fully explored.
Overall, the book was ok as a primer into AI, but otherwise I’d rather put it into ChatGPT for a summary instead.
May 27, 2025
لقد كان هذا الكتاب رحلة فكرية ثرية تأخذك من البدايات المتواضعة للذكاء الاصطناعي إلى آفاقه المستقبلية اللامحدودة. تخاطبك أفكار نايجل تون بأسلوب علمي عميق دون أن تفقد بساطتها، فتُدرك من خلالها أن الذكاء الاصطناعي ليس مجرد أداة تقنية، بل تحوّل جذري في علاقتنا بالمعرفة والقرارات.
ستري كيف قارن الكاتب بين التفكير البشري والإحصائي، وكيف وضع النقاط على الحروف في مسائل أخلاقية مثل التحيز والخصوصية. تشعر وأنت تقرأ أن الذكاء الاصطناعي هو مرآة لنا، لا خصمًا لنا.
إن كنت ممن يسعون لفهم هذا العصر الرقمي بروح ناقدة ونظرة مستقبلية، فإن هذا الكتاب يستحق أن يكون ضمن مكتبتك. فهو لا يجيب فقط عن سؤال: كيف يفكر الذكاء الاصطناعي؟ بل يجعلك تتأمل: كيف يجب علينا نحن أن نفكر فيه؟
تابعني على شبكات التواصل الاجتماعي
سليمان العوشن
http://about.me/aloshan
ستري كيف قارن الكاتب بين التفكير البشري والإحصائي، وكيف وضع النقاط على الحروف في مسائل أخلاقية مثل التحيز والخصوصية. تشعر وأنت تقرأ أن الذكاء الاصطناعي هو مرآة لنا، لا خصمًا لنا.
إن كنت ممن يسعون لفهم هذا العصر الرقمي بروح ناقدة ونظرة مستقبلية، فإن هذا الكتاب يستحق أن يكون ضمن مكتبتك. فهو لا يجيب فقط عن سؤال: كيف يفكر الذكاء الاصطناعي؟ بل يجعلك تتأمل: كيف يجب علينا نحن أن نفكر فيه؟
تابعني على شبكات التواصل الاجتماعي
سليمان العوشن
http://about.me/aloshan
September 25, 2024
This post on AI is incredibly insightful! Understanding how AI thinks and operates is crucial for harnessing its potential. With the rapid advancements in technology, it’s essential to have tools that help us navigate this evolving landscape. For anyone looking to improve their understanding of AI or even assist with their academic writing, I highly recommend checking out https://academized.com/coursework-wri... writing service. They provide excellent resources and support for those wanting to enhance their knowledge and writing skills. Overall, balancing AI's benefits with effective control strategies will ensure we maximize its positive impact on society.
August 13, 2025
I felt it’s a bit too broad and explorative instead of explaining the true way how LLMs and other models “think” — for that found the videos by Andrej Karpathy on YouTube better suited.
Nevertheless the references to Turing machines were a good fresh up, with some questionable references.
Nevertheless the references to Turing machines were a good fresh up, with some questionable references.
August 28, 2024
Good ideas on the positives of AI and how it can be a useful tool.
October 14, 2024
not reference to graphcore, if you are expert on AI and hardware, this book don't add much.
May 30, 2025
Gave up halfway through. Very basic.
June 27, 2025
Excellent book
Ttis is a very good book for someone who wants to understand AI
I recommend it for any one interested to learn about IA
Ttis is a very good book for someone who wants to understand AI
I recommend it for any one interested to learn about IA
July 1, 2025
It’s a nice book to read with generic content from all over the computer science world. I like the proposals on how we should use AI, still was expecting something more deep.
July 7, 2025
Just ok
July 20, 2025
I was probably expecting too much from what is a paperback.
July 20, 2025
3.5 stars
Displaying 1 - 19 of 19 reviews