What do you think?
Rate this book
The Creativity Code: How AI is learning to write, paint and think
Will a computer ever compose a symphony, write a prize-winning novel, or paint a masterpiece? And if so, would we be able to tell the difference?
As humans, we have an extraordinary ability to create works of art that elevate, expand and transform what it means to be alive.
Yet in many other areas, new developments in AI are shaking up the status quo, as we find out how many of the tasks humans engage in can be done equally well, if not better, by machines. But can machines be creative? Will they soon be able to learn from the art that moves us, and understand what distinguishes it from the mundane?
In The Creativity Code, Marcus du Sautoy examines the nature of creativity, as well as providing an essential guide into how algorithms work, and the mathematical rules underpinning them. He asks how much of our emotional response to art is a product of our brains reacting to pattern and structure, and exactly what it is to be creative in mathematics, art, language and music.
Marcus finds out how long it might be before machines come up with something creative, and whether they might jolt us into being more imaginative in turn. The result is a fascinating and very different exploration into both AI and the essence of what it means to be human.
333 pages, Kindle Edition
First published February 18, 2019
313 people are currently reading
4268 people want to read
About the author
Marcus du Sautoy
33 books492 followersMarcus Peter Francis du Sautoy, OBE is the Simonyi Professor for the Public Understanding of Science and a Professor of Mathematics at the University of Oxford.
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 30 of 209 reviews
May 18, 2019
Before I start in on the criticism (there will be criticism), let me say at once that if you're a geek like me you'll find this damn book unputdownable. It is stuffed full of excellent anecdotes about AI, computer science, mathematics and allied subjects, and if you're the sort of person who wants to know how software can fake a Rembrandt, compose a Bach-style chorale or prove the correctness of the Feit-Thompson theorem then you're going to enjoy it as much as I did. Last week, we were visiting Melbourne, catching up with friends we hadn't seen for two years. We had a wall-to-wall social calendar which usually started with an early breakfast and concluded late at night, and all the reading time I had available was odd moments on public transport when we were en route from A to B: nonetheless, I finished The Creativity Code in two days flat. It's that addictive.
And yet... I feel dissatisfied. The author, a mathematician with a wide range of interdisciplinary interests, says the event that pushed him towards writing the book was AlphaGo's already historic match against Lee Sedol. Like me, he'd swallowed the extremely plausible story according to which Go was the ultimate test of AI. Machines can play chess, we were told, because chess is mainly about accurate calculation. But Go is different. To become a world-class Go player, calculation isn't enough. You have to understand. You have to intuit. You have to create. The top Go players, according to the intense mystique that surrounds the game, are a kind of cross between idiot savants, sports stars, philosophers, and Zen monks. They spend their whole lives training, starting from when they are small children, to become one with the Tao of Go. A machine can't do this. We don't even begin to grasp the principles that would make it possible. Except, of course, that now they have. Not only that, but the improved version, AlphaZero, used the exact same methods to learn not just Go, but also chess and Shogi, starting with the bare rules of the three games and training itself up to superhuman level in a few dozen hours. De Sautoy was shaken. Go players like Lee Sedol are out of a job today; tomorrow, perhaps mathematicians like him. He decided to find out more. The book chronicles many examples of machines doing things that you certainly might think would require creativity, and explains roughly how they do it. As noted, it's very interesting stuff.
But here's the odd thing. Careful examination shows that most, perhaps all of the "creative" bots which appear here are really just tricks. They use clever versions of mixing and matching to produce an illusion of creativity, when in fact they're no more than next-generation clockwork automata. AlphaZero, however, is not a trick. It's the real deal; if you want the details, check out Sadler and Regan's Game Changer, which spends four hundred pages analysing AZ's extraordinary playing style, a style it's developed entirely on its own with no human input whatsoever. But de Sautoy says very little about either the chess or the Go versions of AZ, ostensibly the thing which caused him to write the book in the first place. I suspect the reason is that it's difficult to understand what AZ is doing, even in very vague terms, if you aren't yourself a strong games player. Similarly, Game Changer's authors don't say much about the AI aspects. I asked the series editor why they hadn't; he said it was beyond the scope of their publishing house, which specialises in chess, and that what I was asking for was a different book.
Well, I accept that, but I still want it. I feel like the guys in Hitchhiker's Guide who are told the Ultimate Answer is 42, but don't know what the Ultimate Question is. As far as I can see, we have managed to construct a machine with superhuman understanding and creativity, but we can't quite figure out what we did; we have many pieces of the puzzle, and now we need to put them together. Can't someone assemble a suitably qualified team to write the real story of AlphaZero? A million geeks are waiting to read it.
And yet... I feel dissatisfied. The author, a mathematician with a wide range of interdisciplinary interests, says the event that pushed him towards writing the book was AlphaGo's already historic match against Lee Sedol. Like me, he'd swallowed the extremely plausible story according to which Go was the ultimate test of AI. Machines can play chess, we were told, because chess is mainly about accurate calculation. But Go is different. To become a world-class Go player, calculation isn't enough. You have to understand. You have to intuit. You have to create. The top Go players, according to the intense mystique that surrounds the game, are a kind of cross between idiot savants, sports stars, philosophers, and Zen monks. They spend their whole lives training, starting from when they are small children, to become one with the Tao of Go. A machine can't do this. We don't even begin to grasp the principles that would make it possible. Except, of course, that now they have. Not only that, but the improved version, AlphaZero, used the exact same methods to learn not just Go, but also chess and Shogi, starting with the bare rules of the three games and training itself up to superhuman level in a few dozen hours. De Sautoy was shaken. Go players like Lee Sedol are out of a job today; tomorrow, perhaps mathematicians like him. He decided to find out more. The book chronicles many examples of machines doing things that you certainly might think would require creativity, and explains roughly how they do it. As noted, it's very interesting stuff.
But here's the odd thing. Careful examination shows that most, perhaps all of the "creative" bots which appear here are really just tricks. They use clever versions of mixing and matching to produce an illusion of creativity, when in fact they're no more than next-generation clockwork automata. AlphaZero, however, is not a trick. It's the real deal; if you want the details, check out Sadler and Regan's Game Changer, which spends four hundred pages analysing AZ's extraordinary playing style, a style it's developed entirely on its own with no human input whatsoever. But de Sautoy says very little about either the chess or the Go versions of AZ, ostensibly the thing which caused him to write the book in the first place. I suspect the reason is that it's difficult to understand what AZ is doing, even in very vague terms, if you aren't yourself a strong games player. Similarly, Game Changer's authors don't say much about the AI aspects. I asked the series editor why they hadn't; he said it was beyond the scope of their publishing house, which specialises in chess, and that what I was asking for was a different book.
Well, I accept that, but I still want it. I feel like the guys in Hitchhiker's Guide who are told the Ultimate Answer is 42, but don't know what the Ultimate Question is. As far as I can see, we have managed to construct a machine with superhuman understanding and creativity, but we can't quite figure out what we did; we have many pieces of the puzzle, and now we need to put them together. Can't someone assemble a suitably qualified team to write the real story of AlphaZero? A million geeks are waiting to read it.
April 1, 2019
You might expect a book about the potential of AI to be a little dry, maybe a little too niche for a general reader... instead, it's a fascinating multidisciplinary journey into the notion and expression of creativity across time and through varied forms, from music and art, to mathematical theory and games. The principal question: could AI ever truly create something on its own? Could a machine ever step into the realm that some consider an essentially human characteristic: creativity? It's hardly a new question, but with the advances in tech, it seems ever more relevant, maybe even essential. Anyone who has even dipped a toe into sci fi genre might have a few concerns about truly autonomous machines... So, could it happen??
Well, maybe not. Or, at least, not yet. For now, no need to worry too much. Probably.
What machines can do, whether directed via top down human created algorithms or bottom up systems that allow for machine learning, is amazing. If the advances in tech sometimes feel like just another part of modern living, then this book works to illuminate the incredible drive and imagination and hard work that goes into each step forward. The author's ability to convey the excitement of creative energy, whether human, machine, or a combination of both, makes the book a genuine page turner. There's a real sense of exploration, of what has been achieved by humans and/or machines in the past, what's going on right now, and what the pathways might be in the future. It feels like possibilities, like change, like a whole new way of thinking and being. This is a book about people challenging themselves and others, a whole history of innovation that keeps pushing to see what else is achievable, a world of creators that includes coders alongside the greatest artists and musicians. Lifting it all is the author's obvious enthusiasm and his clear desire to make the subject approachable without sacrificing its complexity. It's full of surprises and hints at even more to come.
Overall, the author's reasonable caution about AI's capacity for genuine creativity is somewhat determined by a definition and concept of creativity that seems entirely human. Until AI can feel emotion, can it truly create in the same way humans can? Perhaps these types of boundaries limit too far what we understand to be creative. Would it be enough for the human to bring their own emotion to the piece, as we do in our own individual reactions to art, music, even tech? There's a creative place in the human code that seems almost magical, but if nothing else this books ensures you never forget the sometimes years of work which underlies every invention. Until recently, those hours were determined by human effort, but increasingly the hard slog can be handed over to machines. How different will the creative process become as more and more of what we used to have to do ourselves becomes the realm of machines/AI? The evaluation of machine creativity is often framed by whether it's good enough to fool us, but the more machines become involved in the process, the more they become us, and we them. The question of whether that collaboration could become anything more is tantalising, but not entirely answered here. Creativity is a path, not a moment, so who can really know what is to come? The future is wide open.
ARC via Netgalley
Well, maybe not. Or, at least, not yet. For now, no need to worry too much. Probably.
What machines can do, whether directed via top down human created algorithms or bottom up systems that allow for machine learning, is amazing. If the advances in tech sometimes feel like just another part of modern living, then this book works to illuminate the incredible drive and imagination and hard work that goes into each step forward. The author's ability to convey the excitement of creative energy, whether human, machine, or a combination of both, makes the book a genuine page turner. There's a real sense of exploration, of what has been achieved by humans and/or machines in the past, what's going on right now, and what the pathways might be in the future. It feels like possibilities, like change, like a whole new way of thinking and being. This is a book about people challenging themselves and others, a whole history of innovation that keeps pushing to see what else is achievable, a world of creators that includes coders alongside the greatest artists and musicians. Lifting it all is the author's obvious enthusiasm and his clear desire to make the subject approachable without sacrificing its complexity. It's full of surprises and hints at even more to come.
Overall, the author's reasonable caution about AI's capacity for genuine creativity is somewhat determined by a definition and concept of creativity that seems entirely human. Until AI can feel emotion, can it truly create in the same way humans can? Perhaps these types of boundaries limit too far what we understand to be creative. Would it be enough for the human to bring their own emotion to the piece, as we do in our own individual reactions to art, music, even tech? There's a creative place in the human code that seems almost magical, but if nothing else this books ensures you never forget the sometimes years of work which underlies every invention. Until recently, those hours were determined by human effort, but increasingly the hard slog can be handed over to machines. How different will the creative process become as more and more of what we used to have to do ourselves becomes the realm of machines/AI? The evaluation of machine creativity is often framed by whether it's good enough to fool us, but the more machines become involved in the process, the more they become us, and we them. The question of whether that collaboration could become anything more is tantalising, but not entirely answered here. Creativity is a path, not a moment, so who can really know what is to come? The future is wide open.
ARC via Netgalley
March 11, 2019
At first glance this might just be another 'What AI is good at and not so good at' title. And in a way, it is. But, wow, what a brilliant book! Marcus du Sautoy takes us on a tour of what artificial intelligence has achieved (and possibly can in the future achieve) in a range of fields from his own of mathematics, through game playing, music, art and more.
After a little discussion of what creativity is, we start off with the now very familiar story of DeepMind's AlphaGo and its astonishing ability to take on the hugely challenging game of Go. Even though I've read about this many times before, du Sautoy, as a Go player and mathematician, gives a real feel for why this was such a triumph - and so shocking. Not surprisingly he is also wonderful on what mathematicians actually do, how computers have helped them to date and how they have the potential to do far more in the future. After all, mathematics is by far the closest science to game playing, as it has strict rules established beforehand, where with most other sciences we really don't know what the rules are and have to try to work them out as we go along.
My only slight moan about the mathematical aspect of the content is that there are so many references to the mathematical discipline of symmetry and group theory that it would have been nice to have had an a little backgrounder on these to put what is being talked about into context. Having said that, though, given the inevitably slightly scary aspect of AI potentially taking over some of the roles of mathematicians, du Sautoy gives us an excellently balanced and fair assessment of how some mathematical AI projects really do very little to advance the field, while others really do eat into the mathematician's work.
Perhaps my favourite chapters of all were those on music. There is, of course, a distinctly algorithmic feel to composition theory, and du Sautoy really makes his exploration of this delightful. I'm familiar with a lot of Bach's work, but not The Musical Offering - which du Sautoy argues well was a particularly algorithmic feat. Apparently, Bach was asked if he could improvise a fugue from a very tricky 'tune' dreamed up as a challenge by Frederick the Great - not only did Bach manage a three part version on the spot, he went on to compose 11 variants using all sorts of remarkable modifications. A particular bonus was being able to summon up The Musical Offering on Spotify (itself an algorithmic marvel on a good day) and listen to it as I read the book.
I wasn't quite as taken with the sections on art. With modern art it's arguable that it's pretty much impossible to define what good art is (du Sautoy points out how difficult it is to fake a Jackson Pollack if you know technically what to look for... but I still find his paintings a visual mess that I wouldn't hang in the toilet, so I don't really care that it's hard to reproduce). With such an arbitrary borderline between creativity and randomness, it's hard to worry too much about what an AI can do - it all seems to be a matter of fashion anyway. These chapters were still interesting, just less significant for me. My interest revived, though, when he got onto approaches to writing, even if there was nothing there that seemed likely to displace human work any time soon.
All in all, a great crossover title between computing, mathematics and creativity, presented with du Sautoy's usual charm and clarity. Excellent.
After a little discussion of what creativity is, we start off with the now very familiar story of DeepMind's AlphaGo and its astonishing ability to take on the hugely challenging game of Go. Even though I've read about this many times before, du Sautoy, as a Go player and mathematician, gives a real feel for why this was such a triumph - and so shocking. Not surprisingly he is also wonderful on what mathematicians actually do, how computers have helped them to date and how they have the potential to do far more in the future. After all, mathematics is by far the closest science to game playing, as it has strict rules established beforehand, where with most other sciences we really don't know what the rules are and have to try to work them out as we go along.
My only slight moan about the mathematical aspect of the content is that there are so many references to the mathematical discipline of symmetry and group theory that it would have been nice to have had an a little backgrounder on these to put what is being talked about into context. Having said that, though, given the inevitably slightly scary aspect of AI potentially taking over some of the roles of mathematicians, du Sautoy gives us an excellently balanced and fair assessment of how some mathematical AI projects really do very little to advance the field, while others really do eat into the mathematician's work.
Perhaps my favourite chapters of all were those on music. There is, of course, a distinctly algorithmic feel to composition theory, and du Sautoy really makes his exploration of this delightful. I'm familiar with a lot of Bach's work, but not The Musical Offering - which du Sautoy argues well was a particularly algorithmic feat. Apparently, Bach was asked if he could improvise a fugue from a very tricky 'tune' dreamed up as a challenge by Frederick the Great - not only did Bach manage a three part version on the spot, he went on to compose 11 variants using all sorts of remarkable modifications. A particular bonus was being able to summon up The Musical Offering on Spotify (itself an algorithmic marvel on a good day) and listen to it as I read the book.
I wasn't quite as taken with the sections on art. With modern art it's arguable that it's pretty much impossible to define what good art is (du Sautoy points out how difficult it is to fake a Jackson Pollack if you know technically what to look for... but I still find his paintings a visual mess that I wouldn't hang in the toilet, so I don't really care that it's hard to reproduce). With such an arbitrary borderline between creativity and randomness, it's hard to worry too much about what an AI can do - it all seems to be a matter of fashion anyway. These chapters were still interesting, just less significant for me. My interest revived, though, when he got onto approaches to writing, even if there was nothing there that seemed likely to displace human work any time soon.
All in all, a great crossover title between computing, mathematics and creativity, presented with du Sautoy's usual charm and clarity. Excellent.
December 10, 2024
A well-written and fascinating dive into the world of math, algorithms, Ai, and—of course—art. du Sautoy takes complicated technical ideas and articulates them through understandable analogies and descriptions without completely cutting out the math (like many other authors resort to). He does an excellent job combing his understanding of the technical side with his interdisciplinary knowledge of art and literature. A truly engaging read that I highly recommend to those interested in AI (both the code and the philosophical).
September 8, 2019
The book claims to explore how Artificial Intelligence (AI) will or will not replace human creativity, in poetry, music, and mathematics. It begins with AlphaGo, an AI program that taught itself to play Go and become the strongest player in the world in very short order.
I was disappointed. The book equates AI with deep learning and never seems to recognize that there are decades more work in the field that deep learning ignores. For those who don't know, "deep learning" uses very big, complicated programs based on simulations of neural networks, which were inspired by (loosely) brains. Deep learning has made a big splash, because it can recognize images of cats and drive cars, and because it has the hype and money of Google behind it. But much of the research in DL is presented in this book, and in academia, as if it were the first and only attempt to use computers to solve complicated problems. The more we learn about DL, the more we learn its limitations: it basically re-codes huge datasets in such as way as to make it easier to classify new instances. This is not creativity, as du Sautoy eventually concludes.
But there are areas in AI which do generate creativity. Evolutionary computation, specifically the evolution of computer programs known as "genetic programming" (GP), for example, has produced visual art and music, and has helped discover mathematical laws underlying physical phenomena. GP has been around at least since the late 70s, and arguably since the mid 50s. Undergraduate students and research professors alike are routinely surprised by the creative and unexpected artifacts created by GP. However, du Sautoy never mentions GP, and rarely mentions other AI techniques except to say they are old and limited and can't do what deep learning can do.
It's a shame du Sautoy didn't explore AI a little more, so that he could appreciate what techniques other than deep learning have to say about creativity.
I was disappointed. The book equates AI with deep learning and never seems to recognize that there are decades more work in the field that deep learning ignores. For those who don't know, "deep learning" uses very big, complicated programs based on simulations of neural networks, which were inspired by (loosely) brains. Deep learning has made a big splash, because it can recognize images of cats and drive cars, and because it has the hype and money of Google behind it. But much of the research in DL is presented in this book, and in academia, as if it were the first and only attempt to use computers to solve complicated problems. The more we learn about DL, the more we learn its limitations: it basically re-codes huge datasets in such as way as to make it easier to classify new instances. This is not creativity, as du Sautoy eventually concludes.
But there are areas in AI which do generate creativity. Evolutionary computation, specifically the evolution of computer programs known as "genetic programming" (GP), for example, has produced visual art and music, and has helped discover mathematical laws underlying physical phenomena. GP has been around at least since the late 70s, and arguably since the mid 50s. Undergraduate students and research professors alike are routinely surprised by the creative and unexpected artifacts created by GP. However, du Sautoy never mentions GP, and rarely mentions other AI techniques except to say they are old and limited and can't do what deep learning can do.
It's a shame du Sautoy didn't explore AI a little more, so that he could appreciate what techniques other than deep learning have to say about creativity.
March 27, 2021
For someone who loves science but whose math skills = suspected dyscalculia, this book is a joy. Du Sautoy explains the math in such a way that I was able to try out the problems before engaging in the concepts, and believe me, that’s a feat right there. I could have done with fewer words on Google’s fancy office space in an age of rampant homelessness and glassholery, but overall, a very enjoyable read, packed with ideas and generating quite a few along the way.
May 31, 2024
Li dois livros seguidos sobre a febre da IA Generativa — o primeiro “Creativity Code” (2019), escrito por Marcus du Sautoy como resposta ao impacto do GPT-1, e “Brave New Words” (2024), escrito por Sal Khan, o guru que criou a Khan Academy, como resposta ao impacto do GPT-4 — e confesso ter ficado boquiaberto, não com os feitos do GPT, mas com o deslumbramento dos autores, principalmente no segundo caso, com o livro a funcionar como uma declaração de fé na IA Generativa, suportada diretamente pela Open AI e Bill Gates. Se acalentava dúvidas quanto ao efetivo impacto da IA Generativa fora do domínio criativo, o fervor apresentado por estes livros tornou bastante claro para mim que o rei vai nu.
EN: https://medium.com/@nzagalo/ia-genera...
PT: https://narrativax.blogspot.com/2024/...
EN: https://medium.com/@nzagalo/ia-genera...
PT: https://narrativax.blogspot.com/2024/...
May 18, 2021
Bastante irregular. Empieza muy bien con la prueba de Lovelace y los tipos de creatividad y termina de manera regular, con algunas reflexiones sobre los límites y sesgos de la inteligencia artificial. Entre medio hay 300 páginas que podrían haberse resumido a la mitad. En ellas el autor escribe en modo piloto automático, como si fuera un robot que tuviera que llegar a las 360 páginas para recibir un anticipo mayor; o como vaca sin cencerro, sin dirección alguna y repitiendo la misma idea. Sí, el aprendizaje automático es muy prometedor, pero, de momento, solo hay tentativas en el campo de la pintura, la música y la narración. Por si esto no fuera poco, el autor está obsesionado con las matemáticas y nos da la chapa intercalando capítulos sobre la creatividad de los números y equiparando las ecuaciones y teoremas con las artes. Una lástima.
March 20, 2021
Maybe because I am familiar with deep learning, i found this book the worst kind of insufferable dry musing with no apparent destination. You get a retelling of the AlphaGo game like it was a political drama, pages and pages about how the author counted tiles at the Serpentine, yet no actual depth to any of the algorithms beyond 'it adjusts based on my new input' or 'has a random seed'. No rigourous philosophical comparison of underlying approaches beyond the 'Lovelace test' - something that is more for publicity than actual relevance imho. Neither is there any actual social commentary beyond 'Cambridge Analytica bad! Spotify mediocre!'
It is just such a bizarre collage of shallow info on disjointed topics that never makes a bigger picture that I wonder who the audience is.
It is just such a bizarre collage of shallow info on disjointed topics that never makes a bigger picture that I wonder who the audience is.
Read
September 30, 2019Within creativity code algorithms shape modern life and also Math, music and algorithms are closely connected.
The advent of bottom-up machine learning has revolutionized the field of AI.
کد خلاقیت : هوش مصنوعی در نوشتن،هنر و تفکر چگونه یاد می گیرد
Thanks to machine learning, the talents of modern AI exceed many of our previous expectations. Although they’re still struggling to recognize images and understand language, computers already create fascinating pieces of art, music and literature. However, until they learn to do so consciously and with purpose, they remain creative tools rather than being creative agents themselves.
In these book you’ll learn about:
the three different types of human creativity;
the common language that unites math, music and computing; and
a computer program able to compose symphonies like Bach.
Vision and language still provide big challenges for AI.
Consider this sentence: “The children won’t eat the grapes because they are old.” Who or what is old? Unlike you – a human with some experience with children and grapes – a computer couldn’t easily answer this question. It has no definitive way of knowing who is meant by “they” in this sentence.
Actionable advice:
Learn about the algorithms that control your life.
Big companies like Google, Netflix and Amazon use ever-evolving algorithms to direct your consumer choices. They even track your browsing habits outside of their websites to calculate what to sell you. By learning about which data these companies collect about you, and the principles by which their algorithms operate, you’ll be able to gauge the influence they have on your life – and make conscious decisions to circumvent it.
(Blinkist.com Summary)
The advent of bottom-up machine learning has revolutionized the field of AI.
کد خلاقیت : هوش مصنوعی در نوشتن،هنر و تفکر چگونه یاد می گیرد
Thanks to machine learning, the talents of modern AI exceed many of our previous expectations. Although they’re still struggling to recognize images and understand language, computers already create fascinating pieces of art, music and literature. However, until they learn to do so consciously and with purpose, they remain creative tools rather than being creative agents themselves.
In these book you’ll learn about:
the three different types of human creativity;
the common language that unites math, music and computing; and
a computer program able to compose symphonies like Bach.
Vision and language still provide big challenges for AI.
Consider this sentence: “The children won’t eat the grapes because they are old.” Who or what is old? Unlike you – a human with some experience with children and grapes – a computer couldn’t easily answer this question. It has no definitive way of knowing who is meant by “they” in this sentence.
Actionable advice:
Learn about the algorithms that control your life.
Big companies like Google, Netflix and Amazon use ever-evolving algorithms to direct your consumer choices. They even track your browsing habits outside of their websites to calculate what to sell you. By learning about which data these companies collect about you, and the principles by which their algorithms operate, you’ll be able to gauge the influence they have on your life – and make conscious decisions to circumvent it.
(Blinkist.com Summary)
December 27, 2020
I couldn't put this down. Fascinating, even if (at least so far) most of these machines are only imitating creativity. Loved chapter 11 on music, in particular, but every chapter was wonderful.
January 26, 2022
I find the book a fun read as the topics are related to math, to music, visual arts, even novel writing! AI checking proofs is my favorite:
But of course:
The overall assessment of du Sautoy is that at the current state certainly AI can help assist humans in the creative process, but it in itself doesn't yet have consciousness and thus perhaps no desire from its own to create and to express. Artists, and mathematicians, are not yet at a lot of risk
Just by its sheer computation ability, codes can explore and generate a behemoth amount of possibilities, what notes to arrange in music, what steps to take in math proofs or in a go play, or what nouns what verbs what sentences to continue in writings. Yet it is still humans who make choices, who decide what are of values. The comparison with Jorge Luis Borges's Library of Babel is apt:
du Sautoy also makes a sharp remarks about the marketing/selling of AI:
And lastly I find it particularly beautiful as du Sautoy suggests if we humans wish to understand computers
Stories like this have fuelled mathematicians' fears that computers might lead us to build elaborate edifices on top of programs that are structurally sound. But frankly a human has more chance of making a mistake than a computer. It may be a heresy to admit it, but there are probably thousands of proofs with gaps or mistakes that have been missed.
But of course:
How do we know that the computer program doing the checking doesn't have a bug? You could get another computer to check that program for bugs, but where would this end? Scence and mathematics have always been dogged by this dilemma. How can you be certain that your methods are leading you to true knowledge? Any attempt to prove that it is invariably depends on the methodology you are trying to show produces truth.
The overall assessment of du Sautoy is that at the current state certainly AI can help assist humans in the creative process, but it in itself doesn't yet have consciousness and thus perhaps no desire from its own to create and to express. Artists, and mathematicians, are not yet at a lot of risk
At the moment all the creativity in machines is being initiated and driven by the human code. We are not seeing machiens compelled to express themselves. They don't really seem to have anything to say beyond what we are getting them to do. They are the ventriloquist's dummy providing the mouthpiece for our urge still to express ourselves.
...
And that creative urge is an expression of our belief in free will. We can live our lives like automata or we can suddenly make the choice to stop and break out of the routine and create something that is new. Our creativity is intimately ound up with our free will, something that it seems impossible to automate. To program free will would be to contradict what free will means.
...
Although, then again, we might end up asking whether our free will is an illusion which just masks the complexity of our underlying algorithmic process.
Just by its sheer computation ability, codes can explore and generate a behemoth amount of possibilities, what notes to arrange in music, what steps to take in math proofs or in a go play, or what nouns what verbs what sentences to continue in writings. Yet it is still humans who make choices, who decide what are of values. The comparison with Jorge Luis Borges's Library of Babel is apt:
[] the Library of Babel, where because it contains everything it ends up containing nothing
du Sautoy also makes a sharp remarks about the marketing/selling of AI:
There is a huge amount of hype about AI. There are too many initiatives that are branded as AI but which are little more than statistics or data science. Just as any company wishing to make it at the turn of the millennium would put .com on the end of its name, today it is the addition of the tag AI or Deep which is what companies are using to jump on the bandwagon.
And lastly I find it particularly beautiful as du Sautoy suggests if we humans wish to understand computers
If they [computers] become conscious, it's unlikely to be something that humans will initially understand. Ultimately it will be their paintings, their music, their novels, their creative output, eventheir mathematics that will give us any chance to crack the machine's code and feel what it's like to be a machine.
September 16, 2021
Back in the 1980s I did my PhD in a sub domain of Artificial Intelligence called Machine Learning. Little did I know then that some day it will be much more than an academic curiosity, and there will be billion dollar industries based on it, and people will make lucratively profitable careers practicing it. Now it is almost impossible to find a company that does not claim that they use Machine Learning in their products, even though those claims are too often little more than marketing hypes.
As a result of its spectacular rise, Machine Learning is now a buzzword that everyone likes to use, and most people believe that they know what it is and what it can do. But these popular notions are shrouded in myths and misconceptions. There are now dozens of books that try to explain this technology to ordinary people, but I am yet to see another book that explains the technology better than this book.
However, this is much more than a what-is-machine-learning book. It goes deep into trying to define what we mean by creativity, and then try to show examples where this technology has produced artifacts that we would have called creative if they were created by humans. It is a profoundly thought provoking book, that should be appealing to anyone who is curious about creativity, art, beauty in mathematics, limits of computing, consciousness, and about the near future.
As a result of its spectacular rise, Machine Learning is now a buzzword that everyone likes to use, and most people believe that they know what it is and what it can do. But these popular notions are shrouded in myths and misconceptions. There are now dozens of books that try to explain this technology to ordinary people, but I am yet to see another book that explains the technology better than this book.
However, this is much more than a what-is-machine-learning book. It goes deep into trying to define what we mean by creativity, and then try to show examples where this technology has produced artifacts that we would have called creative if they were created by humans. It is a profoundly thought provoking book, that should be appealing to anyone who is curious about creativity, art, beauty in mathematics, limits of computing, consciousness, and about the near future.
April 9, 2019
This is a book by a mathematician about maths. Much more so than about AI creativity. I was expecting the book to focus on artistic, writing and musical developments so I persevered up until chapter 13. But by then I just couldn't stand listening to any more about prime numbers, equations or algebra. There's a few interesting chapters in there but overall just felt too much off topic and not at all what I wanted to listen to.
September 22, 2020
While I did appreciate feeling like I was back at my liberal arts college (math, computer science, music, art, and literature--all in one place!), the book was rather dry and didn't reveal much I hadn't read elsewhere.
January 30, 2021
What an amazing way to start my reading journey in 2021. Sautoy, an outstanding matemathician in areas concerning objects' symmetry, guides us through the impact of AI in areas related to poetry, music, art, among others. It allows you to understand the impact of the new emerging technologies, always supported by real examples starting with the one of the IBM computer (Deepblue) and its historical win against the world chess champion (Kasparov) to the one responsible for copying and improving Rembrant's art, or Bach compositions, or many others.
A clarified discussion related to consciousness of machines in the future, the conceptual meaning of art, or the unavoidable adoption of the new technological paradigms by the most powerful enterprises such as Google, Netflix, Spotify, etc. are some key points of this book.
You are worth as much as your data is.
A clarified discussion related to consciousness of machines in the future, the conceptual meaning of art, or the unavoidable adoption of the new technological paradigms by the most powerful enterprises such as Google, Netflix, Spotify, etc. are some key points of this book.
You are worth as much as your data is.
April 18, 2024
Завдання цієї книги — перевірити можливості нових моделей штучного інтелекту (ШІ), аби дізнатися, чи зможуть вони коли-небудь створити щось таке, що можна було б порівняти б з дивами нашого людського коду або що навіть перевершувало їх.
Аналітична машина не претендує на створення чогось нового. Вона може робити тільки те, що ми їй накажемо. І все ж існує царина людської діяльності, яка ніколи не стане доступною машинам, — це творчість.
Те, що відбувається в наших головах, залишається таємницею, але останнім роками зʼявилося нове бачення компʼютерного коду: відбувся перехід від «низхідної» концепції програмування до «висхідної», яка, передбачає, що компʼютер має сам знаходити шлях. Тобто створити метапрограму, яка писатиме програму. Метапрограма має бути створена так, аби вона могла вчитися на своїх власних помилках у міру того, як програма, виконує свої функції.
Що було іскрою, із якої розгорілася нова революція штучного інтелекту? Проста відповідь — дані. 90% даних у світі було створено за останні пʼять років. Дані такі важливі для цієї нової революції, що багато хто називає їх новою нафтою.
Недоліком творчості ШІ є те, що він не може рефлексувати над власним результатом і виносити судження про те, їх він хорошим чи поганим, варто цим ділитися — чи видалити. Наразі вся творчість у машинах ініціюється та керується людським кодом. Ми не бачимо, щоб самі машини прагнули самовираження. Насправді їм, мабуть, нема чого сказати за межами того, що доручаємо їм ми. Вони — лялька черевомовця, що слугує рупором для нашого бажання до самовираження.
Поки машина не набуде свідомості, вона не стане чимось більшим, ніж інструмент, що розширює можливості творчості людини. Можливо, нам слід вважати твори наших нових складних алгоритмів якоюсь мірою емерджентними явищами. Якщо ви не готові до поразки — ви не наважитися на ризик, який би дозволив вам відкрити й створити щось радикальне нове. Висадка на Місяць не дала нам якоїсь приголомшливої нової інформації про Всесвіт, але її дали технології, які ми розробили для отримання цього результату.
5/5 Супер🤩
«Код творчості. Як штучний інтелект учиться писати, малювати, думати»
Маркус де Сейтуа
ArtHuss, 2023
#що_я_зараз_читаю
https://t.me/tetiana_garanenko
Аналітична машина не претендує на створення чогось нового. Вона може робити тільки те, що ми їй накажемо. І все ж існує царина людської діяльності, яка ніколи не стане доступною машинам, — це творчість.
Те, що відбувається в наших головах, залишається таємницею, але останнім роками зʼявилося нове бачення компʼютерного коду: відбувся перехід від «низхідної» концепції програмування до «висхідної», яка, передбачає, що компʼютер має сам знаходити шлях. Тобто створити метапрограму, яка писатиме програму. Метапрограма має бути створена так, аби вона могла вчитися на своїх власних помилках у міру того, як програма, виконує свої функції.
Що було іскрою, із якої розгорілася нова революція штучного інтелекту? Проста відповідь — дані. 90% даних у світі було створено за останні пʼять років. Дані такі важливі для цієї нової революції, що багато хто називає їх новою нафтою.
Недоліком творчості ШІ є те, що він не може рефлексувати над власним результатом і виносити судження про те, їх він хорошим чи поганим, варто цим ділитися — чи видалити. Наразі вся творчість у машинах ініціюється та керується людським кодом. Ми не бачимо, щоб самі машини прагнули самовираження. Насправді їм, мабуть, нема чого сказати за межами того, що доручаємо їм ми. Вони — лялька черевомовця, що слугує рупором для нашого бажання до самовираження.
Поки машина не набуде свідомості, вона не стане чимось більшим, ніж інструмент, що розширює можливості творчості людини. Можливо, нам слід вважати твори наших нових складних алгоритмів якоюсь мірою емерджентними явищами. Якщо ви не готові до поразки — ви не наважитися на ризик, який би дозволив вам відкрити й створити щось радикальне нове. Висадка на Місяць не дала нам якоїсь приголомшливої нової інформації про Всесвіт, але її дали технології, які ми розробили для отримання цього результату.
5/5 Супер🤩
«Код творчості. Як штучний інтелект учиться писати, малювати, думати»
Маркус де Сейтуа
ArtHuss, 2023
#що_я_зараз_читаю
https://t.me/tetiana_garanenko
November 13, 2023
Wonderful insight into what AI is and is not capable of in the domains of art, music, writing and (most importantly;) maths.
April 26, 2024
I enjoyed and understood it a lot more than I thought I would hehe
September 10, 2020
A must read by anyone who is interested in the progress being made by intelligently programmed algorithms that have the slightest chance of competing with good-old fashioned human intelligence.
I was really worried when I picked up this book that it will delve into the philosophical aspects of creativity and that it might have a self-help tone to it. It did not. This is a well researched, analytical and most importantly, comprehensive narrative about what is currently being done in the machine learning field to produce projects that could be perceived as creative.
Du Sautoy begins by explaining how for centuries we’ve romanticised the concept of creativity as a godly or external pulse of inspiration and goes on to explore the creative process by simply categorising creative outputs into three basic modes of creativity: exploratory, collaborative and transformative and than splendidly moves on to provide examples from the fields of art, music, mathematics and literature. He cleverly arrives at the conclusion that the process behind each product of human intelligence has an explicit, but mostly hidden algorithm running and controlling it. I will admit that he sent me to shivers when he spoke about Bob in the Serpentine Gallery deciding to project light beyond its operational hour constrains, the group of robots in the Sony laboratory in Paris that created their own language and communicated with one another and the story an algorithm titled “The Seeker” which wrote about the algorithm’s own struggles to communicate with humans, but he goes on to explain why and how exactly these moments that appear as algorithmic ingenuities occurred.
One of things I most enjoyed about this book and that confirmed how well written it is, was that he did not leave any loose ends to any of his statements. With the large breadth he explores integrating concepts from how we evolved as creative beings, why Plato is so dismissive of poetry, how Ada Lovelace’s commentary on the analytical engine has become a phenomenon, why the Turing test is so critical to any algorithm, the duality of the Library of Babel and how Cambridge Analytica came to its own demise; to name a few, he manages to fuses all these critical incidents in such a concise and comprehensive manner in his analysis. One of the other things that stood out to me, is that instead of focusing on Deep Blue’s epic chess game or Watson’s Jeopardy, he chose AlphaGo as his AI heroine, with good mathematical reasoning behind his choice.
The only chapter that I had a hard time following even though, I classify myself as a classic music and Jazz lover was ‘Chapter 11: Music: The Process of Sounding Mathematics’ and that is only because I did not find myself as interested in uncovering the delicate process of composing that he brings to the surface. I found it a bit more dry than the others but still remained charmed by his personal passion towards the art. He revealed in the end of the book that he used the assistance of the Great Automatic Grammatizator to help him write 350 words which we readers are invited to guess belonged to which section. My bet would be on that one.
To Be a Machine and In Our Own Image are two of my absolute favourite books that take on the deeply beloved topic of human vs the machine, and it’s been a long time since I’ve read something on their level. Now, I can add The Creativity Code to that list. For Marcus Du Sautoy, creativity is not this abstract romanticised concept but instead “creativity is about humans asserting they are not machines”. This alone should give you a snippet to the overarching tone of this book.
I was really worried when I picked up this book that it will delve into the philosophical aspects of creativity and that it might have a self-help tone to it. It did not. This is a well researched, analytical and most importantly, comprehensive narrative about what is currently being done in the machine learning field to produce projects that could be perceived as creative.
Du Sautoy begins by explaining how for centuries we’ve romanticised the concept of creativity as a godly or external pulse of inspiration and goes on to explore the creative process by simply categorising creative outputs into three basic modes of creativity: exploratory, collaborative and transformative and than splendidly moves on to provide examples from the fields of art, music, mathematics and literature. He cleverly arrives at the conclusion that the process behind each product of human intelligence has an explicit, but mostly hidden algorithm running and controlling it. I will admit that he sent me to shivers when he spoke about Bob in the Serpentine Gallery deciding to project light beyond its operational hour constrains, the group of robots in the Sony laboratory in Paris that created their own language and communicated with one another and the story an algorithm titled “The Seeker” which wrote about the algorithm’s own struggles to communicate with humans, but he goes on to explain why and how exactly these moments that appear as algorithmic ingenuities occurred.
One of things I most enjoyed about this book and that confirmed how well written it is, was that he did not leave any loose ends to any of his statements. With the large breadth he explores integrating concepts from how we evolved as creative beings, why Plato is so dismissive of poetry, how Ada Lovelace’s commentary on the analytical engine has become a phenomenon, why the Turing test is so critical to any algorithm, the duality of the Library of Babel and how Cambridge Analytica came to its own demise; to name a few, he manages to fuses all these critical incidents in such a concise and comprehensive manner in his analysis. One of the other things that stood out to me, is that instead of focusing on Deep Blue’s epic chess game or Watson’s Jeopardy, he chose AlphaGo as his AI heroine, with good mathematical reasoning behind his choice.
The only chapter that I had a hard time following even though, I classify myself as a classic music and Jazz lover was ‘Chapter 11: Music: The Process of Sounding Mathematics’ and that is only because I did not find myself as interested in uncovering the delicate process of composing that he brings to the surface. I found it a bit more dry than the others but still remained charmed by his personal passion towards the art. He revealed in the end of the book that he used the assistance of the Great Automatic Grammatizator to help him write 350 words which we readers are invited to guess belonged to which section. My bet would be on that one.
To Be a Machine and In Our Own Image are two of my absolute favourite books that take on the deeply beloved topic of human vs the machine, and it’s been a long time since I’ve read something on their level. Now, I can add The Creativity Code to that list. For Marcus Du Sautoy, creativity is not this abstract romanticised concept but instead “creativity is about humans asserting they are not machines”. This alone should give you a snippet to the overarching tone of this book.
October 5, 2022
It's pretty hard to read about topics like AI that are at the vanguard of technology. I think I finished this book abt 2 months ago and in that time there's already been an explosion in commercial AI art usage (Dall-E, Midjourney, NovelAI, Stable Diffusion etc..) that has wiped the floor with many ideas explored here. AI creation is developing at such a breakneck pace that new programs are kicking out any novelty by the time a book has been written. I feel like twitter is a better place to look at new stuff
Regardless, the book such a slog to read. The author gives anecdotes such as the AlphaGo game with this sterile ass prose presenting everyday people as a monolith fighting against progress. Interpretation is rendered binary; so much about composition is seen through a procedural lens of algorithms + huge sample sizes = out comes product.
+ A lot of surface level technical stuff about AI deep learning that taps into a v large database with a lot of computation..
I'm disappointed because I was clickbaited by the title. I feel like the advent of AI is this insanely gripping topic in an apocalyptic existential crisis about becoming obsolete sort of way (like what is the role of the artist, what use do they serve?). This is more of a dry mathematical catalogue of anecdotes.
Regardless, the book such a slog to read. The author gives anecdotes such as the AlphaGo game with this sterile ass prose presenting everyday people as a monolith fighting against progress. Interpretation is rendered binary; so much about composition is seen through a procedural lens of algorithms + huge sample sizes = out comes product.
+ A lot of surface level technical stuff about AI deep learning that taps into a v large database with a lot of computation..
I'm disappointed because I was clickbaited by the title. I feel like the advent of AI is this insanely gripping topic in an apocalyptic existential crisis about becoming obsolete sort of way (like what is the role of the artist, what use do they serve?). This is more of a dry mathematical catalogue of anecdotes.
December 13, 2020
Read this in preparation for an interview with Marcus (side note: the BBC series he hosted - ‘Story of Maths’ - was partly the inspiration for me studying maths at uni, so I was pretty excited to get to chat with him on stage..!)
I enjoyed this book. It’s about AI and creativity, and yes it does the kind of standard AI art stuff - ‘will AI replace artists etc etc’ - but once you get past that (fair play if you like that stuff; imo it’s extremely overdone), there’s some wonderful ideas in here about mathematical creativity and ‘why we even do pure maths’, which is 100% worth the read.
I feel like I spend a lot of time convincing people that doing pure maths is like doing art - it’s about the thought process and the pursuit of proof and the indescribable ‘a-ha’ moments when something just clicks. And this book, paired with Hardy’s ‘A Mathematician’s Apology’, does a superb job of not only bringing this to life, but questioning what it’s really about and what the future of it might look like in a world seemingly obsessed with automating even the most pure forms of human delight.
Read Hardy, then read this.
(Oh and you can watch back my interview with Marcus on some of these big topics - just Google CogX 2020 and our names and it’ll come up 😇)
I enjoyed this book. It’s about AI and creativity, and yes it does the kind of standard AI art stuff - ‘will AI replace artists etc etc’ - but once you get past that (fair play if you like that stuff; imo it’s extremely overdone), there’s some wonderful ideas in here about mathematical creativity and ‘why we even do pure maths’, which is 100% worth the read.
I feel like I spend a lot of time convincing people that doing pure maths is like doing art - it’s about the thought process and the pursuit of proof and the indescribable ‘a-ha’ moments when something just clicks. And this book, paired with Hardy’s ‘A Mathematician’s Apology’, does a superb job of not only bringing this to life, but questioning what it’s really about and what the future of it might look like in a world seemingly obsessed with automating even the most pure forms of human delight.
Read Hardy, then read this.
(Oh and you can watch back my interview with Marcus on some of these big topics - just Google CogX 2020 and our names and it’ll come up 😇)
July 31, 2019
This one is definitely going to become one of my top 5 books read in the past 10 years.
November 11, 2020
Poderão as máquinas inteligentes ser criativas? Mas o que é isso de Inteligência e criatividade? Poderemos realmente definir esses conceitos? E não seremos nós, como entidades biológicas, uma espécie de máquinas naturais, com a inteligência a surgir como propriedade emergente de elementos biológicos que se auto-organizam? O matemático Marcus du Sautoy coleciona neste livro exemplos de algoritmos e robôs que permitem explorar as fronteiras da criatividade artificial. Dos algoritmos aleatórios concebidos por artistas às inteligências Artificiais aparentemente capazes de gerar obras de arte, Sautoy analisa a criatividade com um olho no espírito da máquina, e outro no impulso humano.
Discutir se alguma vez teremos máquinas verdadeiramente inteligentes e criativas é futurismo longínquo. O que dispomos agora é de interfaces, que variam entre simples programação ao processamento por redes neurais, que permitem aos humanos explorar novas formas de criar, utilizando ou em diálogo com mecanismos e software. Gostamos de acreditar na ilusão que o output destes mecanismos é uma expressão de criatividade artificial, mas na verdade são o resultado de algoritmos aleatórios, ou regras iterativas, ou processamento de conjuntos de treino que apesar dos resultados surpreendentes, dependem sempre das escolhas das bases de dado de treinamento de Inteligência Artificial. Gostamos de as criações produzidas ver como produtos de uma mente artificial, mas na verdade dizem mais sobre as sensibilidades, ideias e sensações daqueles que estão envolvidos no seu desenvolvimento, do que realmente espíritos da máquina.
Discutir se alguma vez teremos máquinas verdadeiramente inteligentes e criativas é futurismo longínquo. O que dispomos agora é de interfaces, que variam entre simples programação ao processamento por redes neurais, que permitem aos humanos explorar novas formas de criar, utilizando ou em diálogo com mecanismos e software. Gostamos de acreditar na ilusão que o output destes mecanismos é uma expressão de criatividade artificial, mas na verdade são o resultado de algoritmos aleatórios, ou regras iterativas, ou processamento de conjuntos de treino que apesar dos resultados surpreendentes, dependem sempre das escolhas das bases de dado de treinamento de Inteligência Artificial. Gostamos de as criações produzidas ver como produtos de uma mente artificial, mas na verdade dizem mais sobre as sensibilidades, ideias e sensações daqueles que estão envolvidos no seu desenvolvimento, do que realmente espíritos da máquina.
April 16, 2023
Marcus Du Sautoy bespreekt in zijn kenmerkende vlotte schrijfstijl (prima vertaling) de vraag of AI aanleg heeft voor creativiteit. Hij onderscheidt daarbij drie soorten creativiteit: verkennende , getransplanteerde uit andere vakgebieden, en disruptieve (eigen omschrijving).
Als voorbeelden geeft hij de AI toepassingen bij schaak, go, schilderkunst, literatuur en zijn eigen vakgebied de wiskunde. Elk voorbeeld is op zichzelf al fascinerend, maar Du Sautoy maakt er een samenhangend verhaal van (en dit sluit aan bij een vorig boek dat ik las van Harari die stelt dat Homo sapiens zijn sprong naar bewustzijn maakte als verhalenverteller).
Marcus Du Sautoy werpt de mogelijkheid op dat AI code onze “biologische code” zal overtreffen. Het struikelblok voor AI code is een“bewustzijn”. In mijn eigen woorden vertaald: Heeft een AI entiteit een innerlijke stem die reflecteert op de door de AI entiteit geproduceerde output? Mijn persoonlijke interpretatie van het boek is dat Marcus Du Sautoy dit inderdaad voor mogelijk houdt maar dat het AI bewustzijn anders zal zijn dan ons biologisch bewustzijn. Voor wat het waard is, ik ben dezelfde mening toegedaan.
Het boek is een aanrader. Mijn betere helft zei dat ze mij zelden een boek zo snel zag uitlezen. Voor wie graag experimenteert met GPT 3 (en ondertussen 4) een aanrader (het boek dateert van 2019).
Als voorbeelden geeft hij de AI toepassingen bij schaak, go, schilderkunst, literatuur en zijn eigen vakgebied de wiskunde. Elk voorbeeld is op zichzelf al fascinerend, maar Du Sautoy maakt er een samenhangend verhaal van (en dit sluit aan bij een vorig boek dat ik las van Harari die stelt dat Homo sapiens zijn sprong naar bewustzijn maakte als verhalenverteller).
Marcus Du Sautoy werpt de mogelijkheid op dat AI code onze “biologische code” zal overtreffen. Het struikelblok voor AI code is een“bewustzijn”. In mijn eigen woorden vertaald: Heeft een AI entiteit een innerlijke stem die reflecteert op de door de AI entiteit geproduceerde output? Mijn persoonlijke interpretatie van het boek is dat Marcus Du Sautoy dit inderdaad voor mogelijk houdt maar dat het AI bewustzijn anders zal zijn dan ons biologisch bewustzijn. Voor wat het waard is, ik ben dezelfde mening toegedaan.
Het boek is een aanrader. Mijn betere helft zei dat ze mij zelden een boek zo snel zag uitlezen. Voor wie graag experimenteert met GPT 3 (en ondertussen 4) een aanrader (het boek dateert van 2019).
July 6, 2020
Very interesting read! The author picks up many intriguing examples of algorithms that can beat human Go masters, create new Rembrandt paintings, compose music, write poetry and much more, all the while explaining how it works, without going too far into the technical details and explaining in a fashion easy to follow for someone who is not acquainted with the subject. But is this art? The book doesn't give a "yes or no" answer, but rather many thought-provoking ideas and thoughts about this matter, spanning across various fields of science from mathematics and computer science to neuroscience, linguistics and more.
April 18, 2021
Who's this for:
In the end this book, or better its main question, is more philosophical than technical or creative. It's more about what we deem to be creative. So I guess the book is best suited for people that have some background on the technical part and want to broaden their horizons, adding some meat to the philosophical bone.
If you are a creative (and assume that machines can't be creative) and want to understand more about the technical part to open your thinking, the book probably doesn't deliver convincing enough its points. Superintelligence: Paths, Dangers, Strategies delivers a more robust fundament for that and should be read first.
I liked:
- The combination of mathematics and creativity. Gives a widening of horizon for both angles.
- That the author gave a good overview and enough, but not too much or detailed, history about AI, it's earliest developments, important happenings and more or less a current state.
- That I got some additional insights and appreciation of art, music, and in particular classical music, which has surprisingly much to do with mathematics. On the other hand it's not that surprising either.
- I really liked that the author actually did some experiments and tests himself and had a good hands-on approach. It shows that he actually got deep into the topic.
- That you can learn besides how computer work and think also a bit how humans work and think. That we developed the system of Generative Adversarial Network to judge the vast possible options of possible generations (of an image, idea or else) sounds remarkably similar to how we come up with ideas, find them ingenious and after a night of sleep we judge it ourselves as maybe not such a good idea.
I didn't like:
- It's a bit basic on the technical side if you familiar with certain terms in AI, neuronal networks, machine learning, etc. Some examples are even more confusing to get. A layman might get a too different understanding because it's a tad too dumbed down. Not sure how to do that better though without repel the broader audience.
- While I liked the basic introduction of the topics, it missed a bit to develop really the core insight he wanted to deliver. Or at least I expected a bit more a "scientific approach" to answering the question and not just explanatory history.
- Only on the very last pages he takes a stance that he believes that computers can get conscious and with that become "creative" (whatever that means in the end). I wished a bit more of a combination of logical reasoning and philosophical discussion why and how that could be (but maybe Superintelligence: Paths, Dangers, Strategies is a better read for this). The book is mostly of the form: "Computers cannot do X because it can't think creative. Followed by an example and a group that successfully developed an algorithm or system that did X." This will go on and on until we get to a strong AI and then all of those questions get irrelevant. While this seems obvious for someone that has spent a bit time thinking about this topic (or has read Superintelligence) for a reader that hasn't arrived there already, it's not that obvious to grasp that point (from the last couple pages).
- I loved the comparison for the proof about the infinity of prime numbers from the Mizar system with the "prose proof" he gave. But this is an unfair comparison as the prose is not rigorous enough to be a proof. It certainly helps to understand how it works and is enough to follow a proof but it's not sufficient. Again as a layman you easily could get fooled by believing this was enough.
Things I learned:
- About classical music and how it's composed and that it had more to do with understanding the mathematical and structural side of music than "knowing what good music is and what not". Humans love patterns and structures so we prefer it also in music. Bach definitely understood that not only intuitively. The author also hints at current pop music and it's based on a proven model music-wise (and that's why it's boring from that perspective). At this point I'd like to link to The Song Machine: Inside the Hit Factory which goes into the details how that happened and how modern, at least pop-, music works.
- Babylonians started to calculate with symbols to abstract away some tediousness. They observed more and came up with certain rules and structures how numbers behaved. They invented the basics of Algebra. They didn't proof it though.
- The Greeks later stared to proof those observed behaviours. This based in their general development of rhetorics and logic. They wanted to to have a logical model to proof to other about a certain fact based on arguments and not only convincing.
- The book also shows again our collective chauvinism how we think we are, as a species, so special because we are "creative" and a machine can't be. As the author delivers towards the end, it's just a question of time that we can develop a machine that can also be better at "creative" on any scale we want to measure it. It's probably also better at defining what that is at all.
In the end this book, or better its main question, is more philosophical than technical or creative. It's more about what we deem to be creative. So I guess the book is best suited for people that have some background on the technical part and want to broaden their horizons, adding some meat to the philosophical bone.
If you are a creative (and assume that machines can't be creative) and want to understand more about the technical part to open your thinking, the book probably doesn't deliver convincing enough its points. Superintelligence: Paths, Dangers, Strategies delivers a more robust fundament for that and should be read first.
I liked:
- The combination of mathematics and creativity. Gives a widening of horizon for both angles.
- That the author gave a good overview and enough, but not too much or detailed, history about AI, it's earliest developments, important happenings and more or less a current state.
- That I got some additional insights and appreciation of art, music, and in particular classical music, which has surprisingly much to do with mathematics. On the other hand it's not that surprising either.
- I really liked that the author actually did some experiments and tests himself and had a good hands-on approach. It shows that he actually got deep into the topic.
- That you can learn besides how computer work and think also a bit how humans work and think. That we developed the system of Generative Adversarial Network to judge the vast possible options of possible generations (of an image, idea or else) sounds remarkably similar to how we come up with ideas, find them ingenious and after a night of sleep we judge it ourselves as maybe not such a good idea.
I didn't like:
- It's a bit basic on the technical side if you familiar with certain terms in AI, neuronal networks, machine learning, etc. Some examples are even more confusing to get. A layman might get a too different understanding because it's a tad too dumbed down. Not sure how to do that better though without repel the broader audience.
- While I liked the basic introduction of the topics, it missed a bit to develop really the core insight he wanted to deliver. Or at least I expected a bit more a "scientific approach" to answering the question and not just explanatory history.
- Only on the very last pages he takes a stance that he believes that computers can get conscious and with that become "creative" (whatever that means in the end). I wished a bit more of a combination of logical reasoning and philosophical discussion why and how that could be (but maybe Superintelligence: Paths, Dangers, Strategies is a better read for this). The book is mostly of the form: "Computers cannot do X because it can't think creative. Followed by an example and a group that successfully developed an algorithm or system that did X." This will go on and on until we get to a strong AI and then all of those questions get irrelevant. While this seems obvious for someone that has spent a bit time thinking about this topic (or has read Superintelligence) for a reader that hasn't arrived there already, it's not that obvious to grasp that point (from the last couple pages).
- I loved the comparison for the proof about the infinity of prime numbers from the Mizar system with the "prose proof" he gave. But this is an unfair comparison as the prose is not rigorous enough to be a proof. It certainly helps to understand how it works and is enough to follow a proof but it's not sufficient. Again as a layman you easily could get fooled by believing this was enough.
Things I learned:
- About classical music and how it's composed and that it had more to do with understanding the mathematical and structural side of music than "knowing what good music is and what not". Humans love patterns and structures so we prefer it also in music. Bach definitely understood that not only intuitively. The author also hints at current pop music and it's based on a proven model music-wise (and that's why it's boring from that perspective). At this point I'd like to link to The Song Machine: Inside the Hit Factory which goes into the details how that happened and how modern, at least pop-, music works.
- Babylonians started to calculate with symbols to abstract away some tediousness. They observed more and came up with certain rules and structures how numbers behaved. They invented the basics of Algebra. They didn't proof it though.
- The Greeks later stared to proof those observed behaviours. This based in their general development of rhetorics and logic. They wanted to to have a logical model to proof to other about a certain fact based on arguments and not only convincing.
- The book also shows again our collective chauvinism how we think we are, as a species, so special because we are "creative" and a machine can't be. As the author delivers towards the end, it's just a question of time that we can develop a machine that can also be better at "creative" on any scale we want to measure it. It's probably also better at defining what that is at all.
January 28, 2022
This is a brilliantly written, interesting book about AI and its implications on the realm of creativity. This is the first book I've read by Marcus du Sautoy and after reading this book, I'm going to read his other books. His ability to explain complex mathematical and AI concepts in a way I could understand is phenomenal. The concepts that most resonated with me are his explanations and stories surrounding the three types of creativity (Exploratory, Combinational, and Transformative) and how AI and humans use them.
April 8, 2025
(3.5) An interesting perspective on the futures for AI from a modern-day mathematician, however the theoretical side of the AI concepts were still a little more basic than I was expecting/hoping for (this is an unfair bias with my own CompSci background though). The philosophical points raised throughout the book were interesting, as it dived into our pillars and basic understanding of what we call things like Art and Music, but there were times in the musical section that the book became slightly harder to get through (the reason for 3.5 rather than 4). Overall a cool book to open your minds into the basics of AI and how it’s moulding our future
September 20, 2019
A great book when it comes to outlining the status quo and advances of AI in creative areas such as playing Go, composing music, writing novels or proving mathematical theorems. The author is leading us on a journey through the extensive AI projects such as AlphaGo or DeepBach whose results leave people speechless and also lead to the question if algorithms will develop a form of consciousness one day.
It is definitely a very refreshing book that will keep you engaged but it fails to capture the essence of AI/neural networks; how they work and evolve. The author merely scratches the surface and stops elaborating further when it starts to become really interesting. But anything else would probably go beyond the scope of this book, as its intention is to provide an overview of the world of AI rather than giving a detailed and complex explanation of how it reaches its "creative" output.
It is definitely a very refreshing book that will keep you engaged but it fails to capture the essence of AI/neural networks; how they work and evolve. The author merely scratches the surface and stops elaborating further when it starts to become really interesting. But anything else would probably go beyond the scope of this book, as its intention is to provide an overview of the world of AI rather than giving a detailed and complex explanation of how it reaches its "creative" output.
Displaying 1 - 30 of 209 reviews