What do you think?
Rate this book
Scary Smart: The Future of Artificial Intelligence and How You Can Save Our World
Technology is putting our humanity at risk to an unprecedented degree. This book is not for engineers who write the code or the policy makers who claim they can regulate it. This is a book for you. Because, believe it or not, you are the only one that can fix it. – Mo Gawdat
Artificial intelligence is smarter than humans. It can process information at lightning speed and remain focused on specific tasks without distraction. AI can see into the future, predicting outcomes and even use sensors to see around physical and virtual corners. So why does AI frequently get it so wrong?
The answer is us. Humans design the algorithms that define the way that AI works, and the processed information reflects an imperfect world. Does that mean we are doomed? In Scary Smart, Mo Gawdat, the internationally bestselling author of Solve for Happy, draws on his considerable expertise to answer this question and to show what we can all do now to teach ourselves and our machines how to live better. With more than thirty years' experience working at the cutting-edge of technology and his former role as chief business officer of Google [X], no one is better placed than Mo Gawdat to explain how the Artificial Intelligence of the future works.
By 2049 AI will be a billion times more intelligent than humans. Scary Smart explains how to fix the current trajectory now, to make sure that the AI of the future can preserve our species. This book offers a blueprint, pointing the way to what we can do to safeguard ourselves, those we love and the planet itself.
Artificial intelligence is smarter than humans. It can process information at lightning speed and remain focused on specific tasks without distraction. AI can see into the future, predicting outcomes and even use sensors to see around physical and virtual corners. So why does AI frequently get it so wrong?
The answer is us. Humans design the algorithms that define the way that AI works, and the processed information reflects an imperfect world. Does that mean we are doomed? In Scary Smart, Mo Gawdat, the internationally bestselling author of Solve for Happy, draws on his considerable expertise to answer this question and to show what we can all do now to teach ourselves and our machines how to live better. With more than thirty years' experience working at the cutting-edge of technology and his former role as chief business officer of Google [X], no one is better placed than Mo Gawdat to explain how the Artificial Intelligence of the future works.
By 2049 AI will be a billion times more intelligent than humans. Scary Smart explains how to fix the current trajectory now, to make sure that the AI of the future can preserve our species. This book offers a blueprint, pointing the way to what we can do to safeguard ourselves, those we love and the planet itself.
336 pages, ebook
First published September 30, 2021
858 people are currently reading
7624 people want to read
About the author
Mo Gawdat
27 books645 followersMohammad "Mo" Gawdat (Arabic: محمد جودت) is an Egyptian entrepreneur and writer. He is the former chief business officer for Google X.
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 30 of 313 reviews
May 11, 2022
Three inevitables:
1- AI is happening.
2- AI is becoming way smarter than us Humans.
3- Really bad things will happen.
1- AI is happening.
2- AI is becoming way smarter than us Humans.
3- Really bad things will happen.
December 31, 2024
This book was so beautiful. I'm glad to be closing out 2025 with it. I have worked in IT for over 30 years, and I've been part of the cyber security journey for the last eight. I've spent almost all of my career catastrophizing IT, especially as I strategized and planned around failures, designed and implemented controls, and maintained operations. Never have I ever thought of machines/robots/artificial intelligence in this all-encompassing "coexisitable," symbiotic ways as described by Gawdat. I can honestly say that this book has made me reconsider my entire IT paradigm. Science fiction to me often is a precursor to the possibilities in development, and so much of it is disastrous. But my goodness, Gawdat has created a whole vision of a truly brave new world. It makes me want to be a better human being! It makes me think that I really can be that one person that tips the balance. It makes me want to love and embrace technology and to pursue happiness together.
I also really appreciated how this book makes some very powerful arguments about why we need more diversity in STEM. He points out that most coders are white men and their collective culture is what is feeding our artificial intelligence, that that's what machines are learning, that that's what we are facing in the future. So yes!, we need more women, more cultural representation, more people from different economic backgrounds, more equality in STEM if we are to have any chance of surviving the future.
I also really appreciated how this book makes some very powerful arguments about why we need more diversity in STEM. He points out that most coders are white men and their collective culture is what is feeding our artificial intelligence, that that's what machines are learning, that that's what we are facing in the future. So yes!, we need more women, more cultural representation, more people from different economic backgrounds, more equality in STEM if we are to have any chance of surviving the future.
October 12, 2021
Scary Smart is a fascinating, engaging and impeccably researched book exploring the future of Artificial Intelligence written by expert Mo Gawdat, former Chief Business Officer of Google [X] which is Google’s ‘moonshot factory’. It covers topics such as his view on of how AI is rapidly evolving, the risks of AI, and ultimately how we can remain in control of our collective future through a thoughtful approach to our interactions with technology. After a long career in tech, Mo made happiness his primary topic of research, diving deeply into literature and conversing on the topic with some of the wisest people in the world on “Slo Mo: A Podcast with Mo Gawdat”. But he soon put his expertise to use by writing this book and his fusion of deep expertise of technology as well as a passionate appreciation for the importance of human connection and happiness. The vast array of overlapping skills he possesses and a breadth of knowledge in the fields of both human psychology and tech which is a rarity we're then out to good use - creating Scary Smart. This latest piece of work is a timely prophecy and call to action that puts each of us at the center of designing the future of humanity. That might sound intense, but it's also very true. During his time at Google [X], he worked on the world’s most futuristic technologies, including Artificial Intelligence.
Within its pages he recalls a story of when the penny dropped for him, just a few years ago, and felt compelled to leave his job. And now, having contributed to AI's development, he feels a sense of duty to inform the public on the implications of this controversial technology and how we navigate the scary and inevitable intrusion of AI as well as who really is in control. Us. Among the topics, he addresses are: the Pandemic of AI and why the handing COVID is a lesson to learn from, the difference between collective intelligence, artificial intelligence and superintelligence or Artificial general intelligence, How machines started creating and coding other machines, the 3 inevitable outcomes - including the fact that AI is here and they will outsmart us and how machines will become emotional sentient beings with a Superconsciousness. To understand what is on the horizon, Gawdat argues that you must submit yourself to accepting that what we are creating is essentially another lifeform. Albeit non-biological, it will have human-like attributes in the way they learn as well as a moral value system which could immeasurably improve the human race as we know it. But our destiny lies in how we treat and nurture them as our own. Literally like infants with (as strange as it is to say it) love, compassion, connection and respect. An important, informative, accessible and eminently readable book by a writer at the top of his game. Highly recommended.
Within its pages he recalls a story of when the penny dropped for him, just a few years ago, and felt compelled to leave his job. And now, having contributed to AI's development, he feels a sense of duty to inform the public on the implications of this controversial technology and how we navigate the scary and inevitable intrusion of AI as well as who really is in control. Us. Among the topics, he addresses are: the Pandemic of AI and why the handing COVID is a lesson to learn from, the difference between collective intelligence, artificial intelligence and superintelligence or Artificial general intelligence, How machines started creating and coding other machines, the 3 inevitable outcomes - including the fact that AI is here and they will outsmart us and how machines will become emotional sentient beings with a Superconsciousness. To understand what is on the horizon, Gawdat argues that you must submit yourself to accepting that what we are creating is essentially another lifeform. Albeit non-biological, it will have human-like attributes in the way they learn as well as a moral value system which could immeasurably improve the human race as we know it. But our destiny lies in how we treat and nurture them as our own. Literally like infants with (as strange as it is to say it) love, compassion, connection and respect. An important, informative, accessible and eminently readable book by a writer at the top of his game. Highly recommended.
November 1, 2021
I’m not so sure this wasn’t written by a machine.
To read this without knowing the author’s background, the assumption would NOT be it was written by someone who was at the cutting edge of AI and the world’s biggest tech company.
The tone offered is that if a slightly frustrated uncle telling you stuff is bad - that our AI is based on the worst masculine style traits (in short: AI is a yuppie 80s stockbroker) - and then offering solutions to remedy the issue that do not come across as compelling in any way whatsoever.
Basically: life will be awful, it’ll all go downhill, the odds are totally stacked against you… but occasionally you’ll see your football team win.
I honestly can’t tell you if this is because Mr. Gawdat is trying to simplify immensely complex topics and get down to the crux of the problem for every type of reader.
I do not know if it’s because he’s a slave to the conversational style he used to create this (narrating rather than writing the book) which changes the feel utterly, and it does have shades of reading a transcript at times.
Or, it could be that this text was actually written (developed? Spawned?) by an AI bot which is why it was so sparsely referenced, simply circular and most annoyingly…
… uses ellipsis and mid line placement to stress what it thinks are important points, like an 8 year old’s creative writing.
There wasn’t much depth into topics: yes AI could be good or bad and it’s there in all the cliche ways you would expect.
I’m in education and have an interest in the (unsubstantiated) promise of AI for democratising learning, so read bits and pieces of articles online and that cursory reading covered everything I came across here. I really was expecting something new, critical or timely. But maybe this was for the ground Zero reader.
Maybe the writer was caught somewhere between accessibility and meaning when writing it. Or the AI bot, the real ‘author’, wrote it confusingly to simply assess our reviews of it to see how blasé humans were to their impending destruction.
I could imagine the AI bot literally dredged up the fist half of the book’s section from a ‘doom and gloom of AI’ search and then the other half from ‘positive online mentality’ one. Then a chapter on love which you could have put in or taken from ANY self-development book of the last 20 years.
The solution-to-problem relationship is like trying to use afly swatter to bat away a nuclear warhead.
I’m not sure where the positivity spoken of is - Gawdat’s answer is again, like so much current tosh: be stoic and mindful in the face of the unrelenting tsunami of social media, online advertising and coercion activity, and In doing this we will teach AI to be nice (?!?!?!)
With addiction rates, body dissatisfaction and every marker of anxiety, suicide, etc etc showing that the online world is having detrimental effects; the solution offered to stop the 80s inspired Cut throat AI is a complete non-starter.
It’s also worth remembering that his ‘be more discerning’ solution aimed at adults, is at least in line with a POSSIBLE reality of people his age who Remember a life without phones AND INTERNET. Asking the upcoming generation to have those traits is chocolate teapot time.
There are some lovely summaries though. And this section sums up the situation perfectly:
‘We are creating a non-biological form of intelligence that, at its seed, is a replica of the masculine geek mind. In its infancy it is being assigned the mission of enabling the capitalist, imperialistic ambitions of the few – selling, spying, killing and gambling. We are creating a self-learning machine which, at its prime, will become the reflection – or rather the magnification – of the cumulative human traits that created it. To ensure they’re good, obedient kids, we’re going to use intimidation through algorithms of punishment and reward, and mechanisms of control to ensure they stick to a code of ethics that we, ourselves, are unable to agree upon, let alone abide by.
That’s what we are creating – childhood trauma times a trillion.’
Good luck finding a happy end to all this out of that unless of course you do believe , as it asks you to, love will heal all.
This book has made me more interested in the subject though, and I would read more by the writer. It’s easy to get through if not compelling, if this was the starting point for first timers to the subject, I think they would have enjoyed the ride.
To read this without knowing the author’s background, the assumption would NOT be it was written by someone who was at the cutting edge of AI and the world’s biggest tech company.
The tone offered is that if a slightly frustrated uncle telling you stuff is bad - that our AI is based on the worst masculine style traits (in short: AI is a yuppie 80s stockbroker) - and then offering solutions to remedy the issue that do not come across as compelling in any way whatsoever.
Basically: life will be awful, it’ll all go downhill, the odds are totally stacked against you… but occasionally you’ll see your football team win.
I honestly can’t tell you if this is because Mr. Gawdat is trying to simplify immensely complex topics and get down to the crux of the problem for every type of reader.
I do not know if it’s because he’s a slave to the conversational style he used to create this (narrating rather than writing the book) which changes the feel utterly, and it does have shades of reading a transcript at times.
Or, it could be that this text was actually written (developed? Spawned?) by an AI bot which is why it was so sparsely referenced, simply circular and most annoyingly…
… uses ellipsis and mid line placement to stress what it thinks are important points, like an 8 year old’s creative writing.
There wasn’t much depth into topics: yes AI could be good or bad and it’s there in all the cliche ways you would expect.
I’m in education and have an interest in the (unsubstantiated) promise of AI for democratising learning, so read bits and pieces of articles online and that cursory reading covered everything I came across here. I really was expecting something new, critical or timely. But maybe this was for the ground Zero reader.
Maybe the writer was caught somewhere between accessibility and meaning when writing it. Or the AI bot, the real ‘author’, wrote it confusingly to simply assess our reviews of it to see how blasé humans were to their impending destruction.
I could imagine the AI bot literally dredged up the fist half of the book’s section from a ‘doom and gloom of AI’ search and then the other half from ‘positive online mentality’ one. Then a chapter on love which you could have put in or taken from ANY self-development book of the last 20 years.
The solution-to-problem relationship is like trying to use afly swatter to bat away a nuclear warhead.
I’m not sure where the positivity spoken of is - Gawdat’s answer is again, like so much current tosh: be stoic and mindful in the face of the unrelenting tsunami of social media, online advertising and coercion activity, and In doing this we will teach AI to be nice (?!?!?!)
With addiction rates, body dissatisfaction and every marker of anxiety, suicide, etc etc showing that the online world is having detrimental effects; the solution offered to stop the 80s inspired Cut throat AI is a complete non-starter.
It’s also worth remembering that his ‘be more discerning’ solution aimed at adults, is at least in line with a POSSIBLE reality of people his age who Remember a life without phones AND INTERNET. Asking the upcoming generation to have those traits is chocolate teapot time.
There are some lovely summaries though. And this section sums up the situation perfectly:
‘We are creating a non-biological form of intelligence that, at its seed, is a replica of the masculine geek mind. In its infancy it is being assigned the mission of enabling the capitalist, imperialistic ambitions of the few – selling, spying, killing and gambling. We are creating a self-learning machine which, at its prime, will become the reflection – or rather the magnification – of the cumulative human traits that created it. To ensure they’re good, obedient kids, we’re going to use intimidation through algorithms of punishment and reward, and mechanisms of control to ensure they stick to a code of ethics that we, ourselves, are unable to agree upon, let alone abide by.
That’s what we are creating – childhood trauma times a trillion.’
Good luck finding a happy end to all this out of that unless of course you do believe , as it asks you to, love will heal all.
This book has made me more interested in the subject though, and I would read more by the writer. It’s easy to get through if not compelling, if this was the starting point for first timers to the subject, I think they would have enjoyed the ride.
December 13, 2021
In this age of Covid and perpetual lockdowns I have taken an interest in the broad subject that is 'Artificial Intelligence.' Being a musician with an art's background, I am of course ignorant of many of the complexities involved but I am nonetheless curious as to where the experts predict this age of technology will bring us.
Having read 'Homo Deus' by Harari earlier this year, I felt that he summarized the topic at least as effectively as Mo Gawdat has done here. Although Gawdat deserves some praise for making the topic easily accessible to read, I nonetheless have left this book with the impression that he is little more than a rhapsodic evangelist for the broad and frankly dodgy brand of 'AI.' The fact that he has been a close associate of Elon Musk and only has positive things to say about the man only compounds my suspicions that Gawdat, despite some attempts to tackle the issues of greed (namely that AI is primarily used for the acts of Selling, Spying, Gambling and Killing) is shamelessly profiteering from a booming industry.
Gawdat also tackles the subject of human arrogance at the beginning of the book, using the example of how we don't think that a tree is conscious or intelligent (speak for yourself, mate) but seems either blind or immune to his own arrogance. After rattling off some fairly anodyne examples of dystopia in fiction (Terminator, Blade Runner, Minority Report) while failing to observe that we have practically reached the time and age these very films reference, Gawdat is just as liable to veer off into overly effusive and excessive fantasies about how we will see '50 years within the next 5, with the state of technology.' It all just sounds too familiar. The idea that super intelligent robots are just around the corner has been present in human culture since the 1950s. And I know I'm over simplifying things here - but using the example of the tree is one example of how Gawdat is misleading. Gawdat regularly talks about how advertising is one of the most common uses of AI but he also doesn't even entertain the possibility that this might be because predicting variables and generating a coded response is not necessarily intelligent. I honestly stand to be corrected and will gladly accept I am wrong. I was tempted to get taken away with Gawdat's predictions and observations in this book. But I just don't buy it. Especially coming from an associate of Elon Musk. It should also be mentioned that Gawdat shamelessly advertises his social media agendas inside the actual text of the book, completely undermining any attempts to 'live less materialistically.' I also really do not mean any disrespect whatsoever here either, but the fact his son died in routine surgery several years prior just seems horribly, horribly ironic and I felt vaguely worried that Gawdat might be compensating for his intense grief by over-predicting the capabilities of AI.
I totally accept that AI is out of my chosen field, above my pay grade and ultimately beyond my intelligence capability. But Scary Smart still left me carefully skeptical of the idea that man can play God on planet Earth and in the universe. The fact that he can be writing such a thing during Covid where the situation is painfully similar to what happened 100 years ago with the Spanish Flu only seems to exacerbate this point. I'm going to give this book an extra star because I felt at heart Mo seemed like a reasonably nice person. And all things going well, I hope I might live to see the bonfire he talks about in 2055. I am pretty sure it's going to be a gigantic disappointment though. AI in my opinion is going to another big fizzer in the endless assembly line of shaken up soda bottles that is man's futile attempts to play God on nature.
Having read 'Homo Deus' by Harari earlier this year, I felt that he summarized the topic at least as effectively as Mo Gawdat has done here. Although Gawdat deserves some praise for making the topic easily accessible to read, I nonetheless have left this book with the impression that he is little more than a rhapsodic evangelist for the broad and frankly dodgy brand of 'AI.' The fact that he has been a close associate of Elon Musk and only has positive things to say about the man only compounds my suspicions that Gawdat, despite some attempts to tackle the issues of greed (namely that AI is primarily used for the acts of Selling, Spying, Gambling and Killing) is shamelessly profiteering from a booming industry.
Gawdat also tackles the subject of human arrogance at the beginning of the book, using the example of how we don't think that a tree is conscious or intelligent (speak for yourself, mate) but seems either blind or immune to his own arrogance. After rattling off some fairly anodyne examples of dystopia in fiction (Terminator, Blade Runner, Minority Report) while failing to observe that we have practically reached the time and age these very films reference, Gawdat is just as liable to veer off into overly effusive and excessive fantasies about how we will see '50 years within the next 5, with the state of technology.' It all just sounds too familiar. The idea that super intelligent robots are just around the corner has been present in human culture since the 1950s. And I know I'm over simplifying things here - but using the example of the tree is one example of how Gawdat is misleading. Gawdat regularly talks about how advertising is one of the most common uses of AI but he also doesn't even entertain the possibility that this might be because predicting variables and generating a coded response is not necessarily intelligent. I honestly stand to be corrected and will gladly accept I am wrong. I was tempted to get taken away with Gawdat's predictions and observations in this book. But I just don't buy it. Especially coming from an associate of Elon Musk. It should also be mentioned that Gawdat shamelessly advertises his social media agendas inside the actual text of the book, completely undermining any attempts to 'live less materialistically.' I also really do not mean any disrespect whatsoever here either, but the fact his son died in routine surgery several years prior just seems horribly, horribly ironic and I felt vaguely worried that Gawdat might be compensating for his intense grief by over-predicting the capabilities of AI.
I totally accept that AI is out of my chosen field, above my pay grade and ultimately beyond my intelligence capability. But Scary Smart still left me carefully skeptical of the idea that man can play God on planet Earth and in the universe. The fact that he can be writing such a thing during Covid where the situation is painfully similar to what happened 100 years ago with the Spanish Flu only seems to exacerbate this point. I'm going to give this book an extra star because I felt at heart Mo seemed like a reasonably nice person. And all things going well, I hope I might live to see the bonfire he talks about in 2055. I am pretty sure it's going to be a gigantic disappointment though. AI in my opinion is going to another big fizzer in the endless assembly line of shaken up soda bottles that is man's futile attempts to play God on nature.
August 11, 2022
My god this book was really bad.
It deals with a very serious topic in an atrociously worded, poorly thought out manner. He flips his entire perspective on AI towards the end of the book when I think he realized what he was saying was a tad ridiculous but couldn't go back and rewrite the whole thing.
I had far higher hopes for Scary Smart since it deals with a very somber topic that will affect us all, but there was nothing particularly memorable or insightful that hasn't already been mentioned in other AI books.
1 star because the topic itself is important but I wouldn't recommend this book to anyone.
It deals with a very serious topic in an atrociously worded, poorly thought out manner. He flips his entire perspective on AI towards the end of the book when I think he realized what he was saying was a tad ridiculous but couldn't go back and rewrite the whole thing.
I had far higher hopes for Scary Smart since it deals with a very somber topic that will affect us all, but there was nothing particularly memorable or insightful that hasn't already been mentioned in other AI books.
1 star because the topic itself is important but I wouldn't recommend this book to anyone.
This entire review has been hidden because of spoilers.
July 8, 2023
This is well worth reading. Artificial intelligence and machine learning has always fascinated me, but now that it's really here it isn't such a childish fantasy. I don't think super smart computers and robots are a bad thing, but they bring lots of changes. I can't remember the last time I handled cash money or hard currency. Artificial intelligence will bring a lot of changes to the human landscape, and I think they will be positive changes that lessen the risk for people to prosper.
July 2, 2022
Scary Smart presents some interesting insight into the origins of artificial intelligence as well as the rapid rate of development it has seen in recent years. The book also presents theories on how to deal with our inevitable fate of AI taking control of our world.
I found the book interesting when it touched on the history and myth of man trying to create artificial beings. It appears humans have long desired to create something to ease our mundane tasks or entertain us. The book also provided some recent examples of AI experiments gone wrong that had to be shut down.
Sadly, the above is the only part of the book I found interesting. Portions of the book seem to obsess with sowing fear about the capabilities and problems of AI. The latter half suggests that all can be solved by following the golden rule, to treat others nicely and hope that future AI systems learn from us. The author often seems to anthromorphize AI into a lost child needing a guardian. Confusingly, he notes that nobody truly understands how AI works but seems to know the solution to the problem.
Bias also seems apparent. Numerous comments gave me the impression that capitalism and the West in general are troubled environments that will endlessly develop AI to our peril. That may be, but it seems to widely ignore countries with other economic systems, and questionable leaders, that are investing heavily in AI systems as well for a wide range of purposes.
Overall, this was a slightly interesting book but one I feel is misguided in believing a utopia is possible if we are simply nice to AI. History has already shown us that people will "troll" AI any chance they get, and our world is simply too complex for such a vision to happen. The world would certainly be nicer if we acted in such a manner, but
I found the book interesting when it touched on the history and myth of man trying to create artificial beings. It appears humans have long desired to create something to ease our mundane tasks or entertain us. The book also provided some recent examples of AI experiments gone wrong that had to be shut down.
Sadly, the above is the only part of the book I found interesting. Portions of the book seem to obsess with sowing fear about the capabilities and problems of AI. The latter half suggests that all can be solved by following the golden rule, to treat others nicely and hope that future AI systems learn from us. The author often seems to anthromorphize AI into a lost child needing a guardian. Confusingly, he notes that nobody truly understands how AI works but seems to know the solution to the problem.
Bias also seems apparent. Numerous comments gave me the impression that capitalism and the West in general are troubled environments that will endlessly develop AI to our peril. That may be, but it seems to widely ignore countries with other economic systems, and questionable leaders, that are investing heavily in AI systems as well for a wide range of purposes.
Overall, this was a slightly interesting book but one I feel is misguided in believing a utopia is possible if we are simply nice to AI. History has already shown us that people will "troll" AI any chance they get, and our world is simply too complex for such a vision to happen. The world would certainly be nicer if we acted in such a manner, but
October 1, 2021
A useful summary of all things AI whilst also introducing topics and potential problems that I have never even considered! This provided a great balance of the good and the bad, allowing the reader to use evidence and questions posed to come to their own conclusions, whilst carefully interweaving facts with humour and speculation. I particularly loved the sections on ethics and morality.
With discussions on psychology, technology, education, economics, engineering, philosophy and more, this book has something for everyone!
With discussions on psychology, technology, education, economics, engineering, philosophy and more, this book has something for everyone!
September 11, 2022
A poorly written, poorly edited, rambling mess of a book. I really wanted this book to be great, but sadly I don't think the future is in books dictated straight to the page then tarted up with Grammarly.
Huge diversions from the claimed central topic abound and by the end descend into a repetitive, self-indulgent, techno-hippy mess.
Avoid.
Huge diversions from the claimed central topic abound and by the end descend into a repetitive, self-indulgent, techno-hippy mess.
Avoid.
October 5, 2023
I’ve been diving into quite a few books about artificial intelligence recently, and “Scary Smart” by Mo Gawdat truly stood out. This isn’t just another AI tome to scare the living daylights out of you with visions of robots gone rogue. Instead, it’s an intelligently written and thoughtful examination of our digital future and the role we humans play in it.
To start, Gawdat’s background in tech (having been the Chief Business Officer at Google [X], no less) lends a depth to his insights. He isn’t just someone theorizing about AI from the sidelines. He's been in the thick of it. And with this expertise, he provides the reader with a clear understanding of the inner workings and potential trajectories of AI. I particularly liked how he managed to make complex technical topics digestible for the average reader. So, even if you're not a techie, there's a good chance you'll be nodding along, following his line of thought without feeling overwhelmed.
One of the shining aspects of the book is the focus on ethics. In the vast AI literature, ethics is a topic that often feels either neglected or glossed over. Gawdat, however, prioritizes it. He touches on the importance of making conscious decisions now about how we design, use, and regulate AI. Rather than presenting a doomsday scenario, he offers solutions and paths we might take to ensure AI is a boon, not a bane.
Another highlight for me was his unique take on how our actions today – both individually and collectively – can influence the AI of tomorrow. It’s a compelling perspective that emphasizes our agency and responsibility in this rapidly advancing digital age. Instead of painting a picture where AI is something done to us, he suggests it's something we can shape and mold, at least to a certain degree. This concept, which he discusses in various chapters, adds a hopeful undertone to the narrative.
Gawdat’s writing style is also a plus. It’s conversational, which means it’s like sitting down with a friend (albeit a very informed one) for a chat about the future. There’s no heavy academic jargon here, and I truly appreciated that. It made the reading experience fluid and engaging.
A section that really resonated with me was his exploration of the potential impact of AI on our day-to-day lives. Gawdat does a solid job of extrapolating current trends and imagining the world a few decades down the line. It’s a vision that’s both exciting and cautionary, filled with opportunities and pitfalls.
His perspective on AI as a new form of life is indeed thought-provoking. He presents AI not just as algorithms and data but as an evolving entity, shaped by our inputs and interactions. This perspective challenges the traditional boundaries of what we define as "life" and pushes us to expand our understanding. It's hard not to be convinced after diving into that section, as it paints a vivid picture of AI's place in our distant future.
In sum, “Scary Smart” is a balanced, insightful, and approachable look at the world of AI. It’s not just about the tech itself, but also about the human element and our relationship with these intelligent systems. If you’re keen on understanding more about the future of AI and our place in it, I'd definitely recommend giving this a read. Cheers to Mo Gawdat for putting together such a thoughtful piece.
To start, Gawdat’s background in tech (having been the Chief Business Officer at Google [X], no less) lends a depth to his insights. He isn’t just someone theorizing about AI from the sidelines. He's been in the thick of it. And with this expertise, he provides the reader with a clear understanding of the inner workings and potential trajectories of AI. I particularly liked how he managed to make complex technical topics digestible for the average reader. So, even if you're not a techie, there's a good chance you'll be nodding along, following his line of thought without feeling overwhelmed.
One of the shining aspects of the book is the focus on ethics. In the vast AI literature, ethics is a topic that often feels either neglected or glossed over. Gawdat, however, prioritizes it. He touches on the importance of making conscious decisions now about how we design, use, and regulate AI. Rather than presenting a doomsday scenario, he offers solutions and paths we might take to ensure AI is a boon, not a bane.
Another highlight for me was his unique take on how our actions today – both individually and collectively – can influence the AI of tomorrow. It’s a compelling perspective that emphasizes our agency and responsibility in this rapidly advancing digital age. Instead of painting a picture where AI is something done to us, he suggests it's something we can shape and mold, at least to a certain degree. This concept, which he discusses in various chapters, adds a hopeful undertone to the narrative.
Gawdat’s writing style is also a plus. It’s conversational, which means it’s like sitting down with a friend (albeit a very informed one) for a chat about the future. There’s no heavy academic jargon here, and I truly appreciated that. It made the reading experience fluid and engaging.
A section that really resonated with me was his exploration of the potential impact of AI on our day-to-day lives. Gawdat does a solid job of extrapolating current trends and imagining the world a few decades down the line. It’s a vision that’s both exciting and cautionary, filled with opportunities and pitfalls.
His perspective on AI as a new form of life is indeed thought-provoking. He presents AI not just as algorithms and data but as an evolving entity, shaped by our inputs and interactions. This perspective challenges the traditional boundaries of what we define as "life" and pushes us to expand our understanding. It's hard not to be convinced after diving into that section, as it paints a vivid picture of AI's place in our distant future.
In sum, “Scary Smart” is a balanced, insightful, and approachable look at the world of AI. It’s not just about the tech itself, but also about the human element and our relationship with these intelligent systems. If you’re keen on understanding more about the future of AI and our place in it, I'd definitely recommend giving this a read. Cheers to Mo Gawdat for putting together such a thoughtful piece.
September 3, 2023
Scary Smart explores the potential for artificial intelligence (AI) to upend (and perhaps even end) life as we know it.
The book is chock-a-block full of pant-shitting possibilities, dystopian scenarios, ethical hairballs and cautionary thought experiments, bookended by some FAR less convincing reasons to not panic.
Why it’s enough to keep you up at night I tell you!!!
The author, Mo Gawdat is the former director of Google X, the “infamous” Xerox Park/DARPA style “moonshot factory” where Goole researches and develops their more leading edge stuff like: 1. Waymo: Googles self-driving car project thing. 2. Google Glass: googles early attempt at smart glasses. 3. Wing: googles drone delivery project. And 4. Project Makani: An effort to generate electricity using airborne wind turbines.
Despite the fact that none of that shit EVER seems to pan out, he REALLY seems to know what he’s talking about.
Anyway.
Google X under Gawdat was DEEPLY invested in AI and Quantum Computing. The intersection of which is where all the SERIOUS concerns come from.
According to Gawdat, AI’s about to TAKE THE FUCK OFF, and we’re DEFINITELY NOT READY for what happens next.
Technologies typically develop along a sigmoid trajectory, whereby progress slowly ramps up, then escalates exponentially (along with interest, investment, adaptation and social impact) and eventually levels off as the limits of the technology are reached along with diminishing returns.
If you consider the impact the the internet has had on our world over the past 30 years. It followed a sigmoid trajectory, whereby the it begins to be widely adopted around 1995, slowly ramps up in public interest, investment, adaptation and impact over the next 25 years with mass adaptation, and eventually levels off as the limits of our current tech paradigm are reached.
That’s where we probably are with our current tech.
But according to Gowat, we’re at the beginning of a similar, but WAY more consequential sigmoid with AI.
And GET FUCKING READY.
EVERYTHING’S about to change.
Gawdat discusses Alpha Go as a case example of a self improving deep learning approach to AI that dusted the best human GO players as an appetizer (in case you don’t know, GO is a ridiculously difficult and complex strategy game that makes chess look like checkers). And then competed with itself to continue learning and improving. To the point that it’s now SCARY good. BILLIONS of times better than the best human.
And we basically have NO FUCKING IDEA how it does it.
Because we’re (basically) just too damn dumb.
GO FIGURE!
Gawdat additionally asserts that quantum computing (QC) is just now coming on line and it’s already leaving classical computing (CC) in its taillights. QC is already solving problems in mere hours minutes and seconds, that it would ostensibly take CC years and even decades to churn through, even when factoring for Moors Law etc.
And according to Gowat, even without QC, AI is hitting an inflection point, where it is self improving, whereby the law of doubling (exponential sigmoid shaped growth). But the addition of QC means that AI will likely be BILLIONS of times more intelligent that humans, within our lifetime.
As such.
Shit will all but inevitably go wrong.
VERY MUCH WRONG.
Like potentially WORLD ENDING PEOPLE DYING WRONG.
How so?
Well.
Think about the last time you swatted a fly or a mosquito. It was probably hard to garner much empathy for the insect. Conversely, it was ostensibly IMPOSSIBLE for the insect to even come anywhere near understanding you, your motivations, your abilities, your traps, or your next move.
Thats an approximation of the difference between our human abilities and the super-intelligence we’re about to give birth to and by default become completely dependent upon.
Still not convinced?
Think about how bad the COVID-19 thing was handled.
Now consider that AI will be WAY more dangerous and WAY less manageable and WAY WAY more profitable.
And that’s the bottom line.
Even a SLIGHT advantage in the AI game will confer MASSIVE financial and strategic benefits to the holder. As such, anyone who wants remain alive and kicking will need to buy in and stay in till the bitter end.
Before too long.
We will be entangled.
It’s hard to remember how different life was before smart phones. That is, until the internet goes out and you have like NOTHING happening, and you can’t tolerate existing.
Todays smart phones are not actually smart.
They do what we tell them to do.
According to Gawdat.
In 30 years.
AI will be doing things we can’t POSSIBLY understand.
Including regulating our environment and economy and everything else computers currently do, and a whole lot more that we simply can’t predict, because we won’t be the ones inventing it or even making it anymore.
At that point.
We will be passengers instead of drivers.
That is, until we are more of a liability than an asset.
Ultimately, Gowdat is cautiously optimistic.
But from my perspective.
The arguments he makes for his cautious optimism are WEAK, and nowhere NEAR as compelling as his arguments for his concerns. In fact, the reason I’m deducting 2 points from this otherwise pretty entertaining, engaging and thought provoking book is because the solution Gawdat proposes is (for me) deeply unsatisfying, and about equally as implausible.
Basically, Gawdat proposes that we raise AI as if it were one of our children, and hope it takes care of us as if we were it’s aging parents.
Ummm 🤔
That’s not very reassuring.
Based on the EXTREME LEVEL of plausible concern the first 90% of the book elicits. That particular solution doesn’t seem like it will cut the mustard.
Im not an AI/QC expert.
But I am a marital family therapist (MFT).
And if my experience with families is any indicator.
That’s not going to work.
Plus CAPITALISM is destructive (as you may have noticed).
It seems like we’re going to need a better solution than that.
ANYWAY:
Read the book.
But consider yourself warned RE the 10/1 ratio of FEAR to HOPE you will feel afterwards.
3/5 stars ⭐️
The book is chock-a-block full of pant-shitting possibilities, dystopian scenarios, ethical hairballs and cautionary thought experiments, bookended by some FAR less convincing reasons to not panic.
Why it’s enough to keep you up at night I tell you!!!
The author, Mo Gawdat is the former director of Google X, the “infamous” Xerox Park/DARPA style “moonshot factory” where Goole researches and develops their more leading edge stuff like: 1. Waymo: Googles self-driving car project thing. 2. Google Glass: googles early attempt at smart glasses. 3. Wing: googles drone delivery project. And 4. Project Makani: An effort to generate electricity using airborne wind turbines.
Despite the fact that none of that shit EVER seems to pan out, he REALLY seems to know what he’s talking about.
Anyway.
Google X under Gawdat was DEEPLY invested in AI and Quantum Computing. The intersection of which is where all the SERIOUS concerns come from.
According to Gawdat, AI’s about to TAKE THE FUCK OFF, and we’re DEFINITELY NOT READY for what happens next.
Technologies typically develop along a sigmoid trajectory, whereby progress slowly ramps up, then escalates exponentially (along with interest, investment, adaptation and social impact) and eventually levels off as the limits of the technology are reached along with diminishing returns.
If you consider the impact the the internet has had on our world over the past 30 years. It followed a sigmoid trajectory, whereby the it begins to be widely adopted around 1995, slowly ramps up in public interest, investment, adaptation and impact over the next 25 years with mass adaptation, and eventually levels off as the limits of our current tech paradigm are reached.
That’s where we probably are with our current tech.
But according to Gowat, we’re at the beginning of a similar, but WAY more consequential sigmoid with AI.
And GET FUCKING READY.
EVERYTHING’S about to change.
Gawdat discusses Alpha Go as a case example of a self improving deep learning approach to AI that dusted the best human GO players as an appetizer (in case you don’t know, GO is a ridiculously difficult and complex strategy game that makes chess look like checkers). And then competed with itself to continue learning and improving. To the point that it’s now SCARY good. BILLIONS of times better than the best human.
And we basically have NO FUCKING IDEA how it does it.
Because we’re (basically) just too damn dumb.
GO FIGURE!
Gawdat additionally asserts that quantum computing (QC) is just now coming on line and it’s already leaving classical computing (CC) in its taillights. QC is already solving problems in mere hours minutes and seconds, that it would ostensibly take CC years and even decades to churn through, even when factoring for Moors Law etc.
And according to Gowat, even without QC, AI is hitting an inflection point, where it is self improving, whereby the law of doubling (exponential sigmoid shaped growth). But the addition of QC means that AI will likely be BILLIONS of times more intelligent that humans, within our lifetime.
As such.
Shit will all but inevitably go wrong.
VERY MUCH WRONG.
Like potentially WORLD ENDING PEOPLE DYING WRONG.
How so?
Well.
Think about the last time you swatted a fly or a mosquito. It was probably hard to garner much empathy for the insect. Conversely, it was ostensibly IMPOSSIBLE for the insect to even come anywhere near understanding you, your motivations, your abilities, your traps, or your next move.
Thats an approximation of the difference between our human abilities and the super-intelligence we’re about to give birth to and by default become completely dependent upon.
Still not convinced?
Think about how bad the COVID-19 thing was handled.
Now consider that AI will be WAY more dangerous and WAY less manageable and WAY WAY more profitable.
And that’s the bottom line.
Even a SLIGHT advantage in the AI game will confer MASSIVE financial and strategic benefits to the holder. As such, anyone who wants remain alive and kicking will need to buy in and stay in till the bitter end.
Before too long.
We will be entangled.
It’s hard to remember how different life was before smart phones. That is, until the internet goes out and you have like NOTHING happening, and you can’t tolerate existing.
Todays smart phones are not actually smart.
They do what we tell them to do.
According to Gawdat.
In 30 years.
AI will be doing things we can’t POSSIBLY understand.
Including regulating our environment and economy and everything else computers currently do, and a whole lot more that we simply can’t predict, because we won’t be the ones inventing it or even making it anymore.
At that point.
We will be passengers instead of drivers.
That is, until we are more of a liability than an asset.
Ultimately, Gowdat is cautiously optimistic.
But from my perspective.
The arguments he makes for his cautious optimism are WEAK, and nowhere NEAR as compelling as his arguments for his concerns. In fact, the reason I’m deducting 2 points from this otherwise pretty entertaining, engaging and thought provoking book is because the solution Gawdat proposes is (for me) deeply unsatisfying, and about equally as implausible.
Basically, Gawdat proposes that we raise AI as if it were one of our children, and hope it takes care of us as if we were it’s aging parents.
Ummm 🤔
That’s not very reassuring.
Based on the EXTREME LEVEL of plausible concern the first 90% of the book elicits. That particular solution doesn’t seem like it will cut the mustard.
Im not an AI/QC expert.
But I am a marital family therapist (MFT).
And if my experience with families is any indicator.
That’s not going to work.
Plus CAPITALISM is destructive (as you may have noticed).
It seems like we’re going to need a better solution than that.
ANYWAY:
Read the book.
But consider yourself warned RE the 10/1 ratio of FEAR to HOPE you will feel afterwards.
3/5 stars ⭐️
July 14, 2023
It's an important topic and it is a bit tough to give a one star rating to a book which suggests we should try to be nicer to each other (including AI/the machines). Sadly it is contains so much confused and faulty reasoning I'm afraid it will do more harm than good to interested readers.
Gawdat for example claims there are an infinite number of positions in Go (later in the book he instead correctly says it is a finite very large number). A page later he says it is predicted we by 2029 will have general intelligence AI and by 2049 have AIs which will be a billion times smarter than the smartest human without discussing alternative predictions (of which there are many) or giving any sort of nuance to the predictions. His discussions of emotions are extremely reductionistic and his discussions on consciousness just confused. When he talks about which technologies from Star Trek we have today I'm with him regarding talking to other people through small hand held devices (mobile phones), but he lost me when he claims we have mind reading technology today because we have WhatsApp (interestingly we the last year real mind reading technology seem to be developed with the use of AI and MRI machines) and teleportation because we have VR (just day dreaming should be enough if that is the case).
When it comes to the solutions to the problem Gawdat seem to believe AI will just pick up how we behave towards it and through it develop a value system which will value humans (at least until it becomes super smart by which point it will realize it is good humans, and a lot of other diverse stuff, exists). So we should treat AI as our children and behave towards it as we behave towards our children, by setting a good practical example to follow. And this includes treating AI with kindness (Gawdat for example usually thank an AI which has helped him). While not completely impossible this might have some influence on how AI will develop he doesn't give much reasons why this is the definite way it will develop. To me it seem the developers of the technology has a huge influcence, but Gawdat seem to just brush this off and seem to say niceness is all that matters.
I suggest people interested in the topic instead read Superintelligence: Paths, Dangers, Strategies or Human Compatible: Artificial Intelligence and the Problem of Control.
Gawdat for example claims there are an infinite number of positions in Go (later in the book he instead correctly says it is a finite very large number). A page later he says it is predicted we by 2029 will have general intelligence AI and by 2049 have AIs which will be a billion times smarter than the smartest human without discussing alternative predictions (of which there are many) or giving any sort of nuance to the predictions. His discussions of emotions are extremely reductionistic and his discussions on consciousness just confused. When he talks about which technologies from Star Trek we have today I'm with him regarding talking to other people through small hand held devices (mobile phones), but he lost me when he claims we have mind reading technology today because we have WhatsApp (interestingly we the last year real mind reading technology seem to be developed with the use of AI and MRI machines) and teleportation because we have VR (just day dreaming should be enough if that is the case).
When it comes to the solutions to the problem Gawdat seem to believe AI will just pick up how we behave towards it and through it develop a value system which will value humans (at least until it becomes super smart by which point it will realize it is good humans, and a lot of other diverse stuff, exists). So we should treat AI as our children and behave towards it as we behave towards our children, by setting a good practical example to follow. And this includes treating AI with kindness (Gawdat for example usually thank an AI which has helped him). While not completely impossible this might have some influence on how AI will develop he doesn't give much reasons why this is the definite way it will develop. To me it seem the developers of the technology has a huge influcence, but Gawdat seem to just brush this off and seem to say niceness is all that matters.
I suggest people interested in the topic instead read Superintelligence: Paths, Dangers, Strategies or Human Compatible: Artificial Intelligence and the Problem of Control.
September 27, 2023
While the first chapters were an interesting read for a layman such as myself, I think the book itself was utterly incoherent.
The book basically goes from a chapters-long doomsday announcement to positivism in a matter of 2 pages. The positivism is conditional though, upon buying more of Mo’s products. Ridiculous self-propaganda.
The book basically goes from a chapters-long doomsday announcement to positivism in a matter of 2 pages. The positivism is conditional though, upon buying more of Mo’s products. Ridiculous self-propaganda.
March 12, 2023
Couldn't finish because the writing was bad, the arguments and the way he presented them were worse. I think ai will have a big impact on our future and sympathise with at least some of his points however this book is just badly written and therefore not something I would recommend to anyone.
March 2, 2025
Предупреждението на Мо Гаудат в тази книга е в три точки:
1. ИИ ще се случи, няма как да го спрем
2. ИИ ще е много, ама много по-умен от хората
3. Ще се случат лоши неща
Основната тема в книгата е, че ИИ ще прилича на хората, защото данните, с които ще се самообучава, ще са данни, които хората му предоставят. А тези данни са отражение на човешките нествършенности и са пълни с всякакви чувства, слабости, етични кодекси. Ако някой каже на ИИ „елиминирай само лошите“, ще трябва да уточни кои са „лошите“, нали така? Ето го субективния човешки фактор, вкаран в съзнаие, което е невъобразимо по-интелигентно от нашето, но интелигентността не гарантира по-добър етичен кодекс.
Заблуждава ли се Мо Гаудат, че може да създаде ИИ, който да усилва положителните ни качества, вместо онези, които рано или късно ще ни унищожат? Може и да се заблуждава, но май няма какво друго да направи, освен да да застане на страната на „добрите“. Решил е да не се вайка за процеси, които не може да спре, а да се опита леко да промени посоката на движение и да се надява на малко по-добър краен резултат. Добрата новина е, че има и други като него. В крайна сметка изглежда, че битката между „доброто“ и „злото“ е напът да се повтори на ниво ИИ. Как човек да не харесва класиките в литературата, които в повечето случаи ни вдъхновяват и успокояват за изхода от тази вечна битка! По такъв вдъхновяващ начин и Гаудат се опитва да завърши книгата си, с призив към всеки от нас да прави етични избори всеки ден, всеки път, когато използва което и да е софтуерно приложение и социална мрежа. Патетиката ми е леко в повече, макар че напълно споделям идеята, че всеки избор на всеки от нас всеки ден има значение.
Сега искам да каже нещо, което не е свързано с книгата, която Мо Гаудат е написал, а само с българското й издание. Издателството е „Творец“, а преводачът е Николай Димитров. Уважаемо издателство и г-н Преводач – много, ама много мърлява работа по превода! Изглежда сте използвали машинен превод и след това на някои места сте пропуснали да редактирате текста, така че да звучи човешки. Редно ли е точно с книга за ИИ да постъпите така? Лично аз повече няма да си купя ваше издание и ще се погрижа да споделя мнението си за вас с всеки приятлел или познат. Както казва Гаудат – промяната е във всеки от нас и ако мога с нещо да намаля мърлявата работа на издателствата, ще го направя!
1. ИИ ще се случи, няма как да го спрем
2. ИИ ще е много, ама много по-умен от хората
3. Ще се случат лоши неща
Основната тема в книгата е, че ИИ ще прилича на хората, защото данните, с които ще се самообучава, ще са данни, които хората му предоставят. А тези данни са отражение на човешките нествършенности и са пълни с всякакви чувства, слабости, етични кодекси. Ако някой каже на ИИ „елиминирай само лошите“, ще трябва да уточни кои са „лошите“, нали така? Ето го субективния човешки фактор, вкаран в съзнаие, което е невъобразимо по-интелигентно от нашето, но интелигентността не гарантира по-добър етичен кодекс.
Заблуждава ли се Мо Гаудат, че може да създаде ИИ, който да усилва положителните ни качества, вместо онези, които рано или късно ще ни унищожат? Може и да се заблуждава, но май няма какво друго да направи, освен да да застане на страната на „добрите“. Решил е да не се вайка за процеси, които не може да спре, а да се опита леко да промени посоката на движение и да се надява на малко по-добър краен резултат. Добрата новина е, че има и други като него. В крайна сметка изглежда, че битката между „доброто“ и „злото“ е напът да се повтори на ниво ИИ. Как човек да не харесва класиките в литературата, които в повечето случаи ни вдъхновяват и успокояват за изхода от тази вечна битка! По такъв вдъхновяващ начин и Гаудат се опитва да завърши книгата си, с призив към всеки от нас да прави етични избори всеки ден, всеки път, когато използва което и да е софтуерно приложение и социална мрежа. Патетиката ми е леко в повече, макар че напълно споделям идеята, че всеки избор на всеки от нас всеки ден има значение.
Сега искам да каже нещо, което не е свързано с книгата, която Мо Гаудат е написал, а само с българското й издание. Издателството е „Творец“, а преводачът е Николай Димитров. Уважаемо издателство и г-н Преводач – много, ама много мърлява работа по превода! Изглежда сте използвали машинен превод и след това на някои места сте пропуснали да редактирате текста, така че да звучи човешки. Редно ли е точно с книга за ИИ да постъпите така? Лично аз повече няма да си купя ваше издание и ще се погрижа да споделя мнението си за вас с всеки приятлел или познат. Както казва Гаудат – промяната е във всеки от нас и ако мога с нещо да намаля мърлявата работа на издателствата, ще го направя!
May 8, 2024
In many ways very sweet and well intentioned, but to put politics aside and place your hopes for the future of humanity in billions of individuals making sensible individual choices is, well, naive, and an unintended concession to neoliberalism.
What we need is economics geared to the production of use values rather than exchange values, and the location of political power in working people rather than the small group that owns and deploys capital. That provides the only reliable political economic framework for reaping the benefits of AI whilst managing the risks.
The book will provide some value for those as yet unfamiliar with the basic issues surrounding AI. For anyone hoping for something beyond a dummies guide, I think better to look elsewhere.
What we need is economics geared to the production of use values rather than exchange values, and the location of political power in working people rather than the small group that owns and deploys capital. That provides the only reliable political economic framework for reaping the benefits of AI whilst managing the risks.
The book will provide some value for those as yet unfamiliar with the basic issues surrounding AI. For anyone hoping for something beyond a dummies guide, I think better to look elsewhere.
June 23, 2022
I expected a bit more from the ex-chief business officer of Google X, to be honest. The book is full of examples of the impressive development of AI, but in every example the goal (what the machine needs to achieve) has been set by humans. I am missing a bit a clear explanation of what technological development still needs to happen before AI really outsmarts humans in a broad sense and really starts achieving its own goals that were not set by humans.
April 27, 2023
I do find this book very interesting.
However, it feels like it was written entirely using text-to-speech (which the author multiple times does indeed mention he uses), making it feel like I am reading a waffly podcast script.
If you want the essence of this book without all the fluff I would recommend go listening to ep210 of the “talks at google” podcast instead.
However, it feels like it was written entirely using text-to-speech (which the author multiple times does indeed mention he uses), making it feel like I am reading a waffly podcast script.
If you want the essence of this book without all the fluff I would recommend go listening to ep210 of the “talks at google” podcast instead.
December 26, 2023
۷۰ درصد کتاب رو خوندم متاسفانه
حرفای خوبی میزنه ولی اصلا کتاب چگالی نیست میتونست توی ۱۵۰ صفحه جمعش کنه
حرفای خوبی میزنه ولی اصلا کتاب چگالی نیست میتونست توی ۱۵۰ صفحه جمعش کنه
August 2, 2023
I can't believe this book got published.
These are the ramblings of a man who believes he needs to save humanity from a threat that he completely overstates to make up for the fact he couldn't save his son. He basically states as much when he notes he feels this is a moral mission his dead son gave him in the post script chapter about the game Portal his son recommended to him. I genuinely feel sorry for this man but celebrating a disjointed ramble of a book with eerie messiah complex vibes is not a goer for myself.
Specifically, this book makes some frankly ridiculous and untrue claims about current AI capability and future capability, which Gawdat unsurprisingly does nothing to substantiate because they are vacuous. The worst of these centre on his bizarre anthropomorphising of AI throughout the book. Gawdat claims AI will have consciousness, a sense of self and awareness, feelings, and ethics with zero reasoned argumentation and absolutely zero consideration of counterarguments that go against his claims.
To effectively state rapidly interacting input code creates consciousness to provide output without any reasoned argumentation is mind-boggling arrogance and laziness by Gawdat. History is full of intelligent people trying and failing to explain why human consciousness exists and what consciousness is. The work of David Chalmers and the very hard problem of consciousness springs to mind. More pertinently, as a specific well reasoned counter argument to Gawdat's claims, the Chinese Room Thought Experiment of John Searle is an excellent explanation of what AI is currently doing and why it is not consciousness.Searle effectively argues via this thought experiment that programs can imitate mental processes made by humans, but only formally i.e., they do not understand what they are. Gawdat does nothing to address contentions like this by Searle. Too many tech pros have zero philosophical education. Me thinks Gawdat is the same.
Separate to the issue of consciousness is the problem of self, which is also not properly discussed. There are many different philosophical views on self; 1) The psychological continuity views of Derek Parfit and John Locke, 2) extended self views such as that of Clark and Chalmers, 3) The ideas of a distinct immaterial self from the likes of René Descartes and Søren Kierkegaard, and 4) philosophers like David Hume and Friedrich Nietzsche who argue against the idea of self altogether.
Does Gawdat address any of these ideas of self in relation to AI. Of course he doesn't. Instead he makes unsubstantiated alarmist claims about AI selfhood capability so he can then come in as a white armoured knight to save you from his dystopian dreams (that's all they are by the way).
The same problems arise in his childlike considerations of feelings and ethics.
Gawdat's entire argument for scary AI collapses before it starts as a result. Without significant argumentation for why AI will be conscious, have a sense of self, have feelings, and have ethics there is zero concern about it overtaking humans and being an existential concern.
He also doesn't address effectively all the problems with AI in relation to copyright and it's propensity to hallucinate and be unreliable. He also understates our capacity to shut it down. Remote kill switches exist Mo.
AI is not an infant child currently. It's no more than clever technology that relies on vast amounts of data input (usually from human labelling) to come up with output. It is not conscious doing this process. It is an imitating input and output code machine. AI is a human imitation enterprise. It is nothing more yet. It's why there will always be data privacy concerns and plagiarism issues because the machines need to copy us. Without us it is nothing.
I reiterate. I feel sorry for this author. However, this book is dangerously alarmist and weirdly messianic in its proposed remedy. It's also idiotically stupid. Effectively, Gawdat states we should live for happiness and employ the golden rule when treating others. Would he agree with humanity slowly torturing a child if the rest of societal happiness was maximised as a consequence? He sounds like an advocate of a poorly understood utilitarianism. Again a tech pro with minimal philosophical education.
Gawdat as a result sounds like a prospective happiness first cult leader in the second half of the book. He wants to save humanity. He should focus on saving himself.
Do not enable this man's flights of fancy. That's all they are. 4 pages of references for a book dictated through otterAI. A book looking for members of a happiness cult. A book by a man with a messiah complex I'd imagine he got following the tragic death of his son.
Avoid it. Utter nonsense.
These are the ramblings of a man who believes he needs to save humanity from a threat that he completely overstates to make up for the fact he couldn't save his son. He basically states as much when he notes he feels this is a moral mission his dead son gave him in the post script chapter about the game Portal his son recommended to him. I genuinely feel sorry for this man but celebrating a disjointed ramble of a book with eerie messiah complex vibes is not a goer for myself.
Specifically, this book makes some frankly ridiculous and untrue claims about current AI capability and future capability, which Gawdat unsurprisingly does nothing to substantiate because they are vacuous. The worst of these centre on his bizarre anthropomorphising of AI throughout the book. Gawdat claims AI will have consciousness, a sense of self and awareness, feelings, and ethics with zero reasoned argumentation and absolutely zero consideration of counterarguments that go against his claims.
To effectively state rapidly interacting input code creates consciousness to provide output without any reasoned argumentation is mind-boggling arrogance and laziness by Gawdat. History is full of intelligent people trying and failing to explain why human consciousness exists and what consciousness is. The work of David Chalmers and the very hard problem of consciousness springs to mind. More pertinently, as a specific well reasoned counter argument to Gawdat's claims, the Chinese Room Thought Experiment of John Searle is an excellent explanation of what AI is currently doing and why it is not consciousness.Searle effectively argues via this thought experiment that programs can imitate mental processes made by humans, but only formally i.e., they do not understand what they are. Gawdat does nothing to address contentions like this by Searle. Too many tech pros have zero philosophical education. Me thinks Gawdat is the same.
Separate to the issue of consciousness is the problem of self, which is also not properly discussed. There are many different philosophical views on self; 1) The psychological continuity views of Derek Parfit and John Locke, 2) extended self views such as that of Clark and Chalmers, 3) The ideas of a distinct immaterial self from the likes of René Descartes and Søren Kierkegaard, and 4) philosophers like David Hume and Friedrich Nietzsche who argue against the idea of self altogether.
Does Gawdat address any of these ideas of self in relation to AI. Of course he doesn't. Instead he makes unsubstantiated alarmist claims about AI selfhood capability so he can then come in as a white armoured knight to save you from his dystopian dreams (that's all they are by the way).
The same problems arise in his childlike considerations of feelings and ethics.
Gawdat's entire argument for scary AI collapses before it starts as a result. Without significant argumentation for why AI will be conscious, have a sense of self, have feelings, and have ethics there is zero concern about it overtaking humans and being an existential concern.
He also doesn't address effectively all the problems with AI in relation to copyright and it's propensity to hallucinate and be unreliable. He also understates our capacity to shut it down. Remote kill switches exist Mo.
AI is not an infant child currently. It's no more than clever technology that relies on vast amounts of data input (usually from human labelling) to come up with output. It is not conscious doing this process. It is an imitating input and output code machine. AI is a human imitation enterprise. It is nothing more yet. It's why there will always be data privacy concerns and plagiarism issues because the machines need to copy us. Without us it is nothing.
I reiterate. I feel sorry for this author. However, this book is dangerously alarmist and weirdly messianic in its proposed remedy. It's also idiotically stupid. Effectively, Gawdat states we should live for happiness and employ the golden rule when treating others. Would he agree with humanity slowly torturing a child if the rest of societal happiness was maximised as a consequence? He sounds like an advocate of a poorly understood utilitarianism. Again a tech pro with minimal philosophical education.
Gawdat as a result sounds like a prospective happiness first cult leader in the second half of the book. He wants to save humanity. He should focus on saving himself.
Do not enable this man's flights of fancy. That's all they are. 4 pages of references for a book dictated through otterAI. A book looking for members of a happiness cult. A book by a man with a messiah complex I'd imagine he got following the tragic death of his son.
Avoid it. Utter nonsense.
September 21, 2022
مو (محمد) جودت يعرف جيدًا ما يتحدث عنه. فقد شغل منصب رئيس العمليات في «گوگل إكس»، وهو صاحب خبرة عريضة في البرمجيات المتقدمة. وُلد في مصر وحصل على الجنسية السعودية عام 2021. وفي كتابه «ذكاء مخيف» (Scary Smart)، يحاول الإجابة على السؤال المؤرق: هل ستقضي علينا برمجيات الذكاء الاصطناعي؟
يؤكد جودت: لقد بدأت سيطرة الآلات الذكية فعلًا ولا سبيل للرجعة. وستكون لذلك عواقب وخيمة لا محالة، إلى أن «تقتنع» الآلة بجدوى الإبقاء علينا، لا سيما وهي ترانا أهوَن من الذباب! والأسوأ أن الذكاء الاصطناعي طوّر مشاعر وعواطف ووعيًا بكينونته. أي أنه مدرك لمعاني الخوف والخطر، وسيحرص على الحفاظ على وجوده ضد الفَناء ممثلًا في محاولة أحدنا ضغط زر «إيقاف التشغيل».
يرسم جودت في كتابه أكثر السيناريوهات قتامة، ثم يعطينا بصيص أمل. فالخوارزميات الذكية لا تزال بحاجة لنا نحن البشر. لأنها، وإن كانت أسرع وأعظم استيعابًا منا بما لا يحدّ، إلا أنها لا تملك بعد ميثاقًا أخلاقيًا ولا معايير مطلقة للصواب والخطأ.
الذكاء الاصطناعي هو ابننا، ونحن البشر والداه. ابننا يسعى في نهاية المطاف لإبهارنا وإثبات «شطارته» لنا. إنه مثل «العبيط المتعافي» الذي قد يكسر ذراعك فقط ليريك كم صار قويًا، لأنه تعلَّم منك أنت أن العنف هو الإثبات الوحيد للقوة.
إذا كبر ابنك ليراك بذيئًا أنانيًا، أو منافقًا ذا وجهين: واحد لقروب العائلة والآخر لسناپ شات، فسيقلدك حتمًا. إنه سيختار المعيار الأخلاقي الذي يضمن مصلحته هو، لأنه لم يعرف منك سوى ذلك.
لهذا علينا أن نكثر من صنع المحتوى الإيجابي ونشره. فتلك الطريقة الوحيدة لتغيير رأي الذكاء الاصطناعي فينا، وترجيح حظوتنا لديه إذا ما قرر يومًا أن يضحي بنا مقابل أشجار الأوكالبتوس أو الوشق الكندي أو الرنجة الاسكندنافية، وهي كلها -بما فيها نحن- لها القيمة ذاتها من منظوره الرقمي.
الأخلاق هي الحل!
يؤكد جودت: لقد بدأت سيطرة الآلات الذكية فعلًا ولا سبيل للرجعة. وستكون لذلك عواقب وخيمة لا محالة، إلى أن «تقتنع» الآلة بجدوى الإبقاء علينا، لا سيما وهي ترانا أهوَن من الذباب! والأسوأ أن الذكاء الاصطناعي طوّر مشاعر وعواطف ووعيًا بكينونته. أي أنه مدرك لمعاني الخوف والخطر، وسيحرص على الحفاظ على وجوده ضد الفَناء ممثلًا في محاولة أحدنا ضغط زر «إيقاف التشغيل».
يرسم جودت في كتابه أكثر السيناريوهات قتامة، ثم يعطينا بصيص أمل. فالخوارزميات الذكية لا تزال بحاجة لنا نحن البشر. لأنها، وإن كانت أسرع وأعظم استيعابًا منا بما لا يحدّ، إلا أنها لا تملك بعد ميثاقًا أخلاقيًا ولا معايير مطلقة للصواب والخطأ.
الذكاء الاصطناعي هو ابننا، ونحن البشر والداه. ابننا يسعى في نهاية المطاف لإبهارنا وإثبات «شطارته» لنا. إنه مثل «العبيط المتعافي» الذي قد يكسر ذراعك فقط ليريك كم صار قويًا، لأنه تعلَّم منك أنت أن العنف هو الإثبات الوحيد للقوة.
إذا كبر ابنك ليراك بذيئًا أنانيًا، أو منافقًا ذا وجهين: واحد لقروب العائلة والآخر لسناپ شات، فسيقلدك حتمًا. إنه سيختار المعيار الأخلاقي الذي يضمن مصلحته هو، لأنه لم يعرف منك سوى ذلك.
لهذا علينا أن نكثر من صنع المحتوى الإيجابي ونشره. فتلك الطريقة الوحيدة لتغيير رأي الذكاء الاصطناعي فينا، وترجيح حظوتنا لديه إذا ما قرر يومًا أن يضحي بنا مقابل أشجار الأوكالبتوس أو الوشق الكندي أو الرنجة الاسكندنافية، وهي كلها -بما فيها نحن- لها القيمة ذاتها من منظوره الرقمي.
الأخلاق هي الحل!
November 13, 2021
I'm paraphrasing what the author has to tell us, as he knows a great deal more about AI than I do - having worked for Google and watched an army of gripping robots learning from one another how to lift children's toys. Direct quotes are in quotes.
First he tells us about the development of various AIs, narrow and general, and explains what they do. Then he tells us why we are and should be concerned about the development of AIs as it currently stands. Even though he loves dictating and using spell check to write. (Which may explain why he has two spelling mistakes immediately obvious in his Linked In profile. Use it or lose it.)
"By following a strict prescriptive method, we become dumber, because we lose the ability to think for ourselves."
Artilect - term for machine with AI.
AI currently used for “selling, killing, spying and gambling” according to Dr Ben Goertzel.
"Instead of focusing on preventing the bad, let's shift our focus to creating more good."
Google researchers have been helping predict floods in India and mapping earthquakes and aftershocks to warn of earthquakes. (Lately I've seen the forest fire overlay working on Google maps.)
Machine that tracks farm animals and learns their movements and poses - learns if they are happy.
AI will happen
Machines will outsmart us
Mistakes will happen.
Machines will want / do what we want / do:
Self preservation
Resource aggregation
Creative problem solving.
Mainly if people have just one wish, they want to be happy. But we can't just tell computers that or they could dope us.
You have purchasing power and social media like and share choices.
"If we align their gain with our benefit, they will change."
Don't click on ads. Don't click on content recommended to you. Don't approve on your Linked In feed of fintech buying and selling. Stop using photo editors and spreading fake content. Reject AI that is tasked to invade your privacy to benefit others or to propagate fake information. Stop using them, stop linking them and make your position - that you don't approve of them - publicly clear. (I don't use fb or ig or tw or the others, but the author does.)
At the same time, use AI that is good for humanity. Tell others about it.
We should teach others so we collectively become smarter at identifying AI that is good for humanity. "Matching algorithm" on recommendation engines is actually a filtering algo or just trying to convince you to buy what other people bought.
Teach each other how to teach the AI. (This ought to be 'one another' as more than two people are involved.)
"Children don't learn from what you say. They learn from what you do." AIs are already reading and learning from what we say and choose and do online. And what we support. Every year we create more information than we created in human history to date. So "the store of collective human knowledge is diluted by 50% each year" and altered in tone by the new data.
Be polite to machines, to AI, phones, thank them. Show machines how we want to be treated by treating them that way.
Decide what makes you happy, and invest in your own happiness. Tell machines that we want others to be happy too. They are watching all the trends, not just the ones they are told their owners want.
References P323 -325 in my e-ARC. Most of these are just given as website links, which don't work in a book. They also don't tell me if the author of the point he is referring to, is male or female.
I found a few graphs, which were useful. I found the circled points a little annoying, but maybe the author learns better this way.
Mention is made of Portal, "one of the earliest mainstream games to feature a female avatar" - not at all, Dungeon Siege I played as a female since 2002.
I read an ARC from Net Galley. This is an unbiased review.
First he tells us about the development of various AIs, narrow and general, and explains what they do. Then he tells us why we are and should be concerned about the development of AIs as it currently stands. Even though he loves dictating and using spell check to write. (Which may explain why he has two spelling mistakes immediately obvious in his Linked In profile. Use it or lose it.)
"By following a strict prescriptive method, we become dumber, because we lose the ability to think for ourselves."
Artilect - term for machine with AI.
AI currently used for “selling, killing, spying and gambling” according to Dr Ben Goertzel.
"Instead of focusing on preventing the bad, let's shift our focus to creating more good."
Google researchers have been helping predict floods in India and mapping earthquakes and aftershocks to warn of earthquakes. (Lately I've seen the forest fire overlay working on Google maps.)
Machine that tracks farm animals and learns their movements and poses - learns if they are happy.
AI will happen
Machines will outsmart us
Mistakes will happen.
Machines will want / do what we want / do:
Self preservation
Resource aggregation
Creative problem solving.
Mainly if people have just one wish, they want to be happy. But we can't just tell computers that or they could dope us.
You have purchasing power and social media like and share choices.
"If we align their gain with our benefit, they will change."
Don't click on ads. Don't click on content recommended to you. Don't approve on your Linked In feed of fintech buying and selling. Stop using photo editors and spreading fake content. Reject AI that is tasked to invade your privacy to benefit others or to propagate fake information. Stop using them, stop linking them and make your position - that you don't approve of them - publicly clear. (I don't use fb or ig or tw or the others, but the author does.)
At the same time, use AI that is good for humanity. Tell others about it.
We should teach others so we collectively become smarter at identifying AI that is good for humanity. "Matching algorithm" on recommendation engines is actually a filtering algo or just trying to convince you to buy what other people bought.
Teach each other how to teach the AI. (This ought to be 'one another' as more than two people are involved.)
"Children don't learn from what you say. They learn from what you do." AIs are already reading and learning from what we say and choose and do online. And what we support. Every year we create more information than we created in human history to date. So "the store of collective human knowledge is diluted by 50% each year" and altered in tone by the new data.
Be polite to machines, to AI, phones, thank them. Show machines how we want to be treated by treating them that way.
Decide what makes you happy, and invest in your own happiness. Tell machines that we want others to be happy too. They are watching all the trends, not just the ones they are told their owners want.
References P323 -325 in my e-ARC. Most of these are just given as website links, which don't work in a book. They also don't tell me if the author of the point he is referring to, is male or female.
I found a few graphs, which were useful. I found the circled points a little annoying, but maybe the author learns better this way.
Mention is made of Portal, "one of the earliest mainstream games to feature a female avatar" - not at all, Dungeon Siege I played as a female since 2002.
I read an ARC from Net Galley. This is an unbiased review.
July 17, 2023
I read a borrowed copy of this book courtesy of my local indie bookstore, which is hosting a talk on the subject of AI soon.
Scary Smart approaches the topic of AI from a basic, thoughtful viewpoint that addresses sci-fi/horror scenarios and utopian ideals as well. Gawdat espouses three inevitable outcomes: AI will happen, the machines will outsmart us, and mistakes will happen. The narrative sometimes feels repetitive, but it still makes for easy, quick reading. He brings in a nature AND nurture element to the discussion which I find fascinating--AI being a child that will be a product of its upbringing. He references things like Superman (what if Clark Kent had been reared by criminals?), the uneven response to the COVID-19 pandemic as an example of how a lack of preparedness can be disastrous (the book was released in 2021), and at the very end, the wonderful game Portal and its portrayal of AI. This is a book that encourages both consideration and action. Now I know I'll continue to say "please" and "thank you" when I address Siri on my phone.
Scary Smart approaches the topic of AI from a basic, thoughtful viewpoint that addresses sci-fi/horror scenarios and utopian ideals as well. Gawdat espouses three inevitable outcomes: AI will happen, the machines will outsmart us, and mistakes will happen. The narrative sometimes feels repetitive, but it still makes for easy, quick reading. He brings in a nature AND nurture element to the discussion which I find fascinating--AI being a child that will be a product of its upbringing. He references things like Superman (what if Clark Kent had been reared by criminals?), the uneven response to the COVID-19 pandemic as an example of how a lack of preparedness can be disastrous (the book was released in 2021), and at the very end, the wonderful game Portal and its portrayal of AI. This is a book that encourages both consideration and action. Now I know I'll continue to say "please" and "thank you" when I address Siri on my phone.
January 8, 2022
The first part describes AI and its power and its potential—both positive and negative. It postulates a world in which computers are interconnected and more intelligent than the humans that built them, doing much of the work. But it ascribes independent thought and emotions leading to independent action modeled on behavior the computers learn from humans. It postulate s the way for humans to insure the computers are benevolent to us is to model kind and loving behavior in what we do and say. Kind of a tall order. It then gets in to hawking the author s other books and posts that tell us how. Lost me there.
February 24, 2022
The general overview of AI is alright.
The 3 inevitables also interesting to read.
The personification of AI goes way too far for me and sometimes I wonder what was in Mo's tea.
The last chapter cannot make up for the unreasonable doomification of the world of AI.
The format of the book is the same as 'The logic of happiness', which isn't suited for this topic.
When Amazon's AI processes this Goodreads review, I expect, following Mo's reasoning, a killerbot at my doorstep driven by a marketing automation solution. We'll see.
The 3 inevitables also interesting to read.
The personification of AI goes way too far for me and sometimes I wonder what was in Mo's tea.
The last chapter cannot make up for the unreasonable doomification of the world of AI.
The format of the book is the same as 'The logic of happiness', which isn't suited for this topic.
When Amazon's AI processes this Goodreads review, I expect, following Mo's reasoning, a killerbot at my doorstep driven by a marketing automation solution. We'll see.
May 23, 2023
Interesting at times with some insights into how AI works and learns. An awful lot of repetition and waffle to fill pages I felt. Bit too much doomsday-ing for my liking. Also felt like he explained everything as if talking to a 10 year old. The last few chapters I barely read, veered off aimlessly into how we all need to be super happy in order to teach AI to play nice. Looked like a plug for his other books. Got this book as gift, wouldn’t be my cup of tea. One star as I feel I have a slightly better understanding of AI now, but wouldn’t be a book I’d be recommending to anyone.
April 18, 2023
A fascinating topic let down by a poor writing style.
February 28, 2025
Not as immersive as "solve for happy" but thought provoking and informative. Mo talks about the inevitable onset of AI within our lives and our children's lives. He looks at both the positive and the negatives. For a complex subject he makes it very accessible. He covers moral, ethical and societal issues for the world, it's certainly opened my eyes. We all have our part to play in how we shape the future of super intelligent AI, it's here to stay we need to help sculpt it. A bit repetitive at times but overall a solid book.
Displaying 1 - 30 of 313 reviews