Zdrowy umysł w sieci algorytmów

by Gerd Gigerenzer, Tadeusz Chawziuk (Translator)

Wyobraźmy sobie, że pewna kawiarnia pozbyła się całej konkurencji przez to, że oferuje darmową kawę, więc nie mamy innego wyboru niż umawiać się właśnie tam. Podczas gdy rozkoszujemy się miłymi rozmowami z przyjaciółmi, zamontowane w stołach podsłuchy i kamery dokładnie monitorują nasze konwersacje i rejestrują, z kim siedzimy. W pomieszczeniu jest także pełno sprzedawców, którzy płacą za naszą kawę, ale nieustannie nam przerywają, proponując zakup swoich produktów i usług. W istocie klientami kawiarni są ci sprzedawcy, a nie Ty i Twoi przyjaciele. Na takiej zasadzie funkcjonują media społecznościowe.

Czy powinniśmy beztrosko oddać programom komputerowym prawo podejmowania za nas osobistych decyzji? Zdecydowanie nie. Inteligencja nie oznacza pokładania ślepej ufności w technologii, ale nie oznacza także przejawiania względem niej chorobliwej nieufności. Chodzi o to, by zdawać sobie sprawę z rzeczywistych możliwości AI i wiedzieć, co jest tylko chwytem marketingowym i artykułem techno-religijnej wiary. Chodzi także o zdolność zachowania kontroli nad urządzeniami.

A przede wszystkim o to, żeby nie przestać myśleć w świecie, który próbuje myśleć za nas.

Genres: Psychology, Nonfiction, Science, Business, Artificial Intelligence, Technology, Self Help

400 pages, Paperback

Published April 12, 2023

About the author

Gerd Gigerenzer is a German psychologist who has studied the use of bounded rationality and heuristics in decision making, especially in medicine. A critic of the work of Daniel Kahneman and Amos Tversky, he argues that heuristics should not lead us to conceive of human thinking as riddled with irrational cognitive biases, but rather to conceive rationality as an adaptive tool that is not identical to the rules of formal logic or the probability calculus.

Gerd Gigerenzer ist ein deutscher Psychologe und seit 1997 Direktor der Abteilung „Adaptives Verhalten und Kognition“ und seit 2009 Direktor des Harding-Zentrum für Risikokompetenz, beide am Max-Planck-Institut für Bildungsforschung in Berlin. Er ist mit Lorraine Daston verheiratet.

Gigerenzer arbeitet über begrenzte Rationalität, Heuristiken und einfache Entscheidungsbäume, das heißt über die Frage, wie man rationale Entscheidungen treffen kann, wenn Zeit und Information begrenzt und die Zukunft ungewiss ist (siehe auch Entscheidung unter Ungewissheit). Der breiten Öffentlichkeit ist er mit seinem Buch Bauchentscheidungen, bekannt geworden; dieses Buch wurde in 17 Sprachen übersetzt und veröffentlicht.

[English bio taken from English Wikipedia article]

[Deutsche Autorenbeschreibung aus dem deutschen Wikipedia-Artikel übernommen]

Reviews

[Cliff M · Rating 4 out of 5]

Helpful but sobering. A non-technical description and discussion of how algorithms are being used by governments and corporations to keep us stupid (sorry... 'shape our lives'). I guess it never hurts to be reminded of the misogyny of Zuckerburg or the mendacity of Boris Johnson (so you can defend against them and their like) but it also can be depressing. I haven't yet recovered my equilibrium...

A good friend who works in IT consultancy read the book and immediately started again from the beginning. I was surprised by this, but I get it now. I too am going to reread selected chapters. In fact, I would say that you can pretty much treat each chapter as an essay in its own right, and just focus on those topics that interest you. Eg for me the fact the book opens with internet dating was disappointing as of no interest. I guess some principles of algorithmic matching in social media were outlined therein, but I felt like I could have skipped all the tackiness and not suffered in my understanding as a result.

If you are going to read it, read it soon as it will date quickly. Eg it still talks about FBook being used by young people, years after the last young person to have an account abandoned the platform to their grandparents.

[Moh. Nasiri · Rating 5 out of 5]

این کتاب خوب از معرفی های دکتر آذرخش در لایو اینستا درخصوص یافته های روانشناسی و علمی است. چگونه در جهان هوشمند، هوشمند باشیم.
لینک صوتی معرفی مفصل کتاب:
https://www.instagram.com/tv/CfrkKs7o...

[Daniel Taylor · Rating 4 out of 5]

Before this, I'd have described myself as an AI optimist.

But Gigerenzer piled on argument after convincing argument that AI is at its best in a stable, predictable environment. Faced with uncertainty, AI sucks. The idea of Artificial General Intelligence (AGI), that can teach itself how to do anything a human better than a human can, seems a far off possibility, if it's even possible at all.

To gain insight into the underlying philosophical viewpoints that shape each of our views about AI and what's possible, I recommend The Fourth Age: Smart Robots, Conscious Computers, and the Future of Humanity by Byron Reese.

While I think we do need to consider our choices around AI now, as William MacAskill argues in What We Owe the Future, I don't think we need to worry about "intelligent" technology replacing us any time soon.

[Pedro Correia · Rating 3 out of 5]

The book approaches several angles on the AI usage in the world.
The author writes in a very simple and provocative way, not becoming too heavy for the reader.

Unfortunately, the title of the book is not explained in it, I was hoping to find some hint on where AI is more of a complement than a replacement and how to leverage that.

Yet the author spends more than 2/3 of the book stating what AI cannot do as if it were something static. And in some examples (self driving cars) making bold assumptions as it won't evolve much more than that.

I don't think the book so much touches on how to stay smart in a smart world but more on how the world is not that smart at all.

Last but not least, the ending part where the main topic of the book is brought up feels short, with more to develop.

It's a thoughtful book, filled with interesting scenarios but does not deliver on what it said it would do.

[Vincze Andrada · Rating 4 out of 5]

"Soluționismul tehnologic este credința că orice problemă socială este o "eroare" care trebuie "rezolvată" printr-un algoritm. Paternalismul tehnologic este consecința firească, guvernarea prin algoritmi."

[Kai-Christian · Rating 4 out of 5]

Also, als jemand der mit vielen der Themen bereits Berührungspunkte hatte, waren Passage des Buches detailliert repetitiv geschrieben. Der Autor stößt mit diesen Beschreibungen sehr viele Türen zu verschiedensten Themen auf und lässt den Leser, ohne das so präsent platzierte "Wie" zu bestimmen, vor dem Thema stehen.

Die Frage nach dem anleitenden "Wie" wurde mir bis zum Schluss nicht deutlich. Das Buch endet wie es anfängt - der Leser wird auf Studien noch und nöcher gestoßen, eine Anleitung erhält er nicht. Schade!

Die 4 Sterne rühren daher, dass es vielleicht die Antwort und das "Wie" nicht gibt - gerade im Umgang mit KI ein heiß diskutiertes Thema. Der Autor gibt, gerade für themenfremde Leser, einen Überblick und regt dazu an, sich mit dem Thema zu beschäftigen.

[Lamia · Rating 4 out of 5]

Zimny prysznic dla wszystkich niespecjalistów zafascynowanych uczeniem maszynowym. Ale tez uspokajacz nastrojów apokaliptycznych rodem z Terminatora.

[Andrea · Rating 3 out of 5]

Ich denke das Buch soll vor allem eine Diskussion anregen.
Über Künstliche Intelligenz, Datenschutz, Soziale Medien.
Mir war es an manchen Stellen etwas zu überspitzt, aber im Ganzen sehr interessant und hat mich auf jeden Fall zum Nachdenken gebracht.

[Yehyeon Cho · Rating 5 out of 5]

How to Stay Smart in a Smart World - Why Human Intelligence still Beats Algorithms

by

Gerd Gigerenzer


Artificial intelligence is an incredibly powerful tool, the scale of which can be sometimes quite frightening. When AI beats our best in chess and Go, previously and erroneously thought unattainable, some may imagine a terminator like AI to be just in yonder future, ready to either annihilate or subjugate us.
Others may paint a different picture, in which algorithms will eventually make better decisions on every front, leaving us with the conundrum of having to choose between freedom to make bad choices and giving up that freedom for far superior choices.

We may be standing at a juncture that could be decisive for our future and well being. Technology has transformed our lives for the better in a myriad of ways, but at what cost? We often, albeit inadvertently, pay with our personal data for the luxury that technology affords us. We are constantly subject to a barrage of psychological tricks and surveillance from groups that may have a hidden agenda that we know little to nothing about.

So what does that mean? Will technology be boon or bane? Will we be redundant in the face of superior AI?

Well, the author gives reassurances and warnings. In his opinion, AI will still need us because algorithms outperform us under stable conditions on any day of the week, but the world is riddled with too much uncertainty all the time. But he warns us to be wary of technology, as it is far more subversive to our democracy and freedoms than one might think.

All in all a great read, even for self proclaimed digital natives.

[Shanoe · Rating 4 out of 5]

Ein sehr spannendes Sachbuch zum Thema Digitalisierung und Künstlicher Intelligenz. Ob des Titels hätte ich mir etwas mehr Tipps oder Ratschläge erwartet, aber insgesamt gesehen wurde dieses umfangreiche und spannende Thema sehr strukturiert, unterhaltsam und anschaulich beleuchtetet. Wirklich sehr empfehlenswerd, wenn man sich mit Schlagwörtern wie selbstfahrende Autos, Big Data oder auch klassisch Datenschutz und Privatsphäre auseinander setzen möchte. Von mir gibts dafür vier Sterne und eine Leseempfehlung.

[Jan Slupski · Rating 4 out of 5]

Autor częściowo uspokaja nas przed technologią podkreślając różnice pomiędzy naszym myśleniem a 'myśleniem' maszyn i programów.
Książka bardzo popularnonaukowa, co dla mnie było wadą (wolałbym miejscami więcej konkretów) ale dla innych mogłoby to być zaletą. Miejscami monotonna, jednak podkreśla bardzo istotne i aktualne problemy społeczne z którymi mierzymy się jako społeczeństwo.

[Osama Qandeel · Rating 5 out of 5]

One of rare opinions about AI that makes sense and is joyful to mind. Many "Intellectuals" take AI as granted without feeling the need for further understanding or deep thinking about it.

The author in this book gives an interesting point of view that puts back some rationality into AI discussions. It opened my appetite for reading more works of him.

[Adrianaa · Rating 5 out of 5]

Bardzo mi się podobała ta książka. W ciekawy i przystępny sposób opisuje mechanizmy trenowania AI i zahacza o nasze sposoby podejmowania decyzji.

[Julia :) · Rating 4 out of 5]

It’s amazing how quickly things become out of date with the speed of technology. I’d love to get a follow-on to this book with the advent of ChatGPT. Great overview on the topic!

[Mangoo · Rating 4 out of 5]

A not-so-early and not-so-distant warning to remain alert and vigil in the forthcoming world dominated by so-called artificial intelligence. Gigerenzer builds profusely on his life-long studies on decision under uncertainty to reason that machine learning works best in stable conditions, which are not the unpredictable conditions we typically find in the real world. Therefore in his opinion AI should be complemented by psychological (human) intelligence to face daily risks and uncertainties, for which the past and correlations are by definition not predictive enough. Besides dismantling some extremes of the AI hype - most likely that part of the book will die by new AI achievements - Gigerenzer iterates the power of fast and frugal heuristics which add the advantage of transparency and need of little data. Gigerenzer voices openly worries about surveillance capitalism and its nefarious effects on our privacy, freedom, and mental health. He remains proud of the achievements of AI, but skeptical of exaggerated claims mixed with marketing and still argues that a better internet should be based on pay-per-service and not pay-by-your-data.
AI is going to be pervasive, and seemingly echoing McLuhan one like Gigerenzer could anticipate that this enormous prothesis will coincide with disinterest or dismissal of organic intelligence, whereas there are ever more reasons to become ever smarter and awake in view of risks and treats.

[Alex · Rating 2 out of 5]

This is a really poorly organised book with little discursive drive, problems which act against its goals of educating the reader to stay data-savvy (it usually forgets, anyway) and communicating the author’s ideas about the AI-human interface. With a shrewd editor there could have been a good short book or a few good magazine articles in place of this - these are interesting ideas - but expert readers will be frustrated and novices will probably come away with quite a few misconceptions. A real shame as I really love the perspectives and insights I was able to pull out. I will be looking up the author’s research with great interest.

[Emil O. W. Kirkegaard · Rating 4 out of 5]

Poorly organized, reads like a book length semi connected blog posts. Author is very old. But it has some interesting contrarian takes so I'm giving 4 instead of 3.

[Max Ungar · Rating 4 out of 5]

there is much useful knowledge and i would recommend the book, but it has many flaws; what i would call a "tome of semi-inverted wisdom". he does discuss principles of when tech, or complex decisionmaking, is useful or not, relative to human reasoning, and simplified reasoning and perception, respectively. there were a lot of things i was delightfully surprised or enlightened by, like degrees of automation in driving (and the discussion of marketing propaganda); healthcare technology propaganda; diminishing and regressing returns on computational usefulness in complex systems, and fast 'n' frugal trees; common conceptions across time on the premise of "creativity"; second order consequences of regulation, resulting in pseudo-malpractice; error culture sociology; and more. however, he never seems to explicitly clarify general principles on the time and place for human reasoning. he spends a lot of time discussing the value of heuristics for making all kinds of practical decisions, but heuristics are extremely bad at making many kinds of practical decisions. the human mind was crafted in an evolutionary context and is not calibrated for the modern world. the negative consequences of such thinking may be seen on different social tiers; individual level, small collective levels, larger collective levels. for example, the popularity heuristic of popular = good; there are many cases of popular people who are incompetent at their roles!, and this is a major principle at play in politics, a realm that most people recognize as being absurdly flawed. he doesn't discuss how industries and the media zeitgeist are largely constructed to game our heuristical decisionmaking. again, lots of useful knowledge, but if you don't spend a lot of time critically assessing the ideas in the book for yourself, it could do more harm than good. my life has been heavily influenced for the worse by people who only know how to think heuristically, so i feel strongly about this topic.

an interesting topic was intuitive thinking with lack of conscious justification, or gut feelings. he doesn't mention the idea that in collaborative scenarios, people make decisions premised on the ideas of other people. when push comes to shove, maybe you're forced to make gut decisions in crises, but it's not a long term sustainable way to run an organization. critically analyzing ideas is one of the best ways to generate even better ideas. it is a really terrible experience, with many terrible outcomes, to live or work with people who assert their opinions and choices solely based on their *feeling* that something is right.

another example is satisficing. he discusses its value, i think in the context of what show to watch, but doesn't discuss the meta-context; that is, only discusses the selection process for choosing a show, but doesn't discuss the selection process for choosing a mode of entertainment. you might say that the underlying principle of satisficing may hold statistically true, but the book is oriented toward applied contexts, which demands sourcing ideas from outside a pure, single-subject theoretical framework. i've spent a lot of time browsing netflix, and while there are some good shows, most of it is factory-churned bullshit. depending on the general viability, a person or couple might alternatively consider video, board, or other games, reading, walking, sports, or even do research on modes of entertainment, and consider how well they might be fulfilling along different parameters, like type 1 vs type 2 enjoyment, or instrumental functional utility relative to other goals (eg broader cultural knowledge, exercising strategic thinking ability, or social bonding), or exercising a lifestyle ethic of active curationism. it's important to recognize and point out that certain options seem like the obvious choice because of, e.g., the bias of normalcy, but that the are many statistically normal things that are counterproductive in many ways.
the discussion on satisficing is also interesting to compare with the first-page bias in search results, where he says you *shouldn't* be satisfied with initial results that seem acceptable. this reminds me of how sowell fails to point out how in the case of newspapers, market forces produce some bad results. (check out my Basic Economics: Fifth Edition review.)

[Susanna Bloderer · Rating 3 out of 5]

In 2025, society is faced with particular challenges such as AI and the shortening of attention spans, social media and smartphone addiction, as well as discerning facts from false information. Gerd Gigerenzer not only warns of the risks of modern technology, but also tries to arm the reader with useful advice on how to navigate these modern tools.

Full disclosure, I was reading this book from a lens of already having been deeply skeptical of AI and refusing to use it, so a lot of what Mr. Gigerenzer wrote on that topic was confirming my negative attitude towards it.

For me, it wasn't difficult to guess what the author's political opinions are - which is most certainly liberal or at the very least center-left, as there were strong criticisms of Qanon or opponents of the Covid lockdowns. There was, to Mr. Gigerenzer's credit, criticism of the Black Mirror-esque Social Credit system and, inevitably, China, but more from a liberal German well-to-do Baby Boomer rather than an alternative media journalist.

What I found impressive was that so many different topics and statistics were covered - I didn't actually think that it was too much - and that it's such a relevant book to modern times. With technology evolving so rapidly, many works that are even only five years old, are already outdated - but not this one. You might wonder what traffic accidents might have to do with the end of democracy as we know it, but Mr. Gigerenzer finds a way of connecting these seemingly unrelated issues.

My takeaway from this book is to stay away from AI and smart technology as much as I can, to be weary of sacrificing my privacy on the altar of convenience and to not use most social media websites. It was also revealing - although not entirely surprising - that even if you don't accept a website's cookies, that the provider might override your consent anyway ("if you're not paying for the product, you are the product").

In summary I would say that I've learned quite a few things that I didn't know about, my skeptical stance on AI obviously hasn't changed (I don't think I'd have chosen to read this book, had I not agreed with the premise) and I felt the book was well-researched on most topics. I recommend reading "How to Stay Smart in a Smart World", but may I add, with a lot of discernment and critical thinking.

[Doug Sleeter · Rating 5 out of 5]

I've been obsessing about the dangers of "big bad data" for over a decade. This book helps put into words many of those dangers, and explains how the bad small data becomes BIG problems when it's aggregated, reported, or otherwise (mis)interpreted by humans.

The author separates problems into two large groups. Finite, or "stable" problems such as playing chess where fixed rules are followed, vs. problems that involve "uncertainty" where the environment is ever changing. For example driving a car, human relationships, or perceiving nuances in spoken languages.

The book discusses how the tech industry has gone hog-wild towards AI systems that promise to provide better solutions to healthcare (IBM's Deep Blue that was fired by many early adopters), or to provide surveillance by way of tracking computer users by giving free software.

The book is a huge wake-up call for humans to remain skeptical of reports in the media (social or otherwise) that cite "studies" with conclusions based on flawed research processes, or simply misunderstanding of what the data actually says.

Data is always about the past and therefore cannot predict with certainty anything in the future. And even when we can see trends in data, we must still remain skeptical of how data is gathered, analyzed and reported. Too often data and studies are used to prove a desired truth (the Texas Sharpshooter Fallacy), and many are politically motivated.

[Deepak P · Rating 4 out of 5]

Wonderful book by a psychologist focusing on the 'smart world' around us. There are many interesting ways of looking at this book. It gives us insights on the sinister ways adopted by the tech industry to further their economic goals, such as the usage of a theory called 'intermittent reinforcement' by the major players in the attention economy. The second perspective tells us how the human being is much superior in a variety of tasks as compared to AI. For example, a young kid seeing one cat just needs that one example to get a concept of a cat, and can use that to identify cats accurately, whereas an AI may need hundreds of thousands of images to match that accuracy. One trivia which may be of interest is the underlying principle behind the design of the captchas; humans can solve it almost with no effort, whereas a machine would struggle to classify it accurately. Please read the book to get the information on the underlying principle :)

While the book isn't political in an explicit sense, there are significant politically aligned arguments which implicitly argue against the mindless usage of machine intelligence in sectors where human decision making is much more appropriate.

[Linh Nguyen · Rating 4 out of 5]

This book offers a fascinating perspective from a respected psychologist who specializes in heuristics and decision-making, exploring the impact of AI on human lives. I appreciate the author's thorough research on AI, which effectively explains different types of AI. As a psychologist, the author delves into the complexities of the human mind, emphasizing its irreplaceability by AI. I share the belief that the human brain is the most complex system in the world, unmatched by any AI.

Several key ideas stand out in the book:

- AI and machine learning excel in stable, predictable environments. Only the human mind can navigate instability and uncertainty.
- There is a distinction between psychological AI (teaching machines to understand human thought processes and decision-making) and machine learning (teaching machines to learn from big data).
- It's important not to fear AI but to understand its workings and impact to prepare for a future with it.

The book's downside is that the last third feels repetitive and could be more concise. Additionally, given the rapid advancements in AI, some content might become outdated in a few years. Despite these issues, the book provides an intriguing perspective, and I look forward to exploring more AI-related books on my reading list.

[Abdullah Cemil Akcam · Rating 4 out of 5]

An excellent book by decision making guru Gerd Gigerenzer. Very much like his other books, this work rigorously summarizes the dangers and misconceptions regarding technology, Internet and artificial intelligence. To be honest, the content of the book was unexpected to me because main topic of the book is not similar to author's main focus of study. Some takeaways from the book: life is uncertain and it is hard if not impossible to predict it, AI or big data dont help in highly uncertain circumstances such as predicting future behavior but simple heuristics can do a fair job, we need to develop critical thinking skills to not to be lost in current misinformation sea.
As a final note, I really enjoyed the repeated mention of author's academician friend who meets with someone online and goes for a date only to learn that the partner was posting fake pictures of other women instead of herself 🤓 guess the surname of the poor guy was Epstein

[Siim · Rating 5 out of 5]

A German has written this book, which makes it down-to-Earth and more critical, yet realistic about the possibilities and non-threats of AI. Because I am in the field of big data, the arguments felt like something I have thought myself but without formulating them as such and bringing the type of simple yet powerful examples that he brings. Especially crucial is the fact that AI can outperform humans in a stable and predictable environment. Humans do very poorly at prediction themselves, but somehow are more open to changing scenarios, even with limited learning data. The SMART thing to do is using AI within constraints and for uses where it excels. Examples are for creating stable environments where self-driving cars can drive safely (like dedicated highways), certain medical fields where the AI has a dedicated function (such as suturing). For most things, using simple decision trees usually performs as well if not better than a ML model - like criminal sentencing.

[Adam Zuwerink · Rating 4 out of 5]

This is a very hard book to rate. The concepts are great and very thought provoking, especially the discussion of the intuition of the human brain. But you will enjoy it a lot more if you go into it thinking of each of the 11 chapters as it's own self-contained unit, which is what finally made me to decide to rate the entire book 4 stars instead of 3.

There is no overarching theory or summary conclusion of how the 2 halves of the book tie together, often leading to contradictions between why the human brain is better than AI at intuition - and yet still gets manipulated so easy. For example, one chapter will claim social media newsfeeds alter our behavior, but another chapter will claim it has no meaningful change on our voting behavior. Which is it?

[Laurie · Rating 4 out of 5]

Author Gerd Gigerenzer has written this book based on the premise that human intelligence still beats the various options that alternatively exist such as algorithms, robots, and other machines powered by artificial intelligence. In a direct quote from the author, he tells us that ‘Although we shouldn’t trust smart technology unconditionally, we shouldn’t fear it unthinkingly either.” This is a very thought-provoking and original treatise on issues that will be coming up more and more in our society in the coming years. He espouses the ‘bounded rationality’ tradition of Herbert Simon in economics and decision making, and this book will certainly generate a lot of discussion amongst those who read it,

[Melinda Peters · Rating 3 out of 5]

Definitely not the type of book I was anticipating reading. From reading the title, I was really interested in reading a book about How To Stay Smart in A Smart World; however, I was rather disappointed while reading as I don't recall anything that pointed in the direction of your typical "How To" book. Instead, I feel that I was filled with a large amount of history about technology and the process that it has gone through with advancement. There were several pieces of information that I wasn't aware of, which was nice to see included in the book. I did find it rather frustrating though that I was anticipating one thing based on the title and read very little of the "How To" part. Next book I read, I won't be anticipating anything.

[James F. McGrath · Rating 5 out of 5]

This is a really important book that exposes how much of what people naively accept about AI is marketing hype disproven by the studies that have been made of what it is and does. The explanation of what AI does and does not do, the explanation of how the real world presents challenges that AI is not up to navigating in the form of fully autonomous vehicles, and the explanation of why so-called predictive AI is especially dubious, need to be read widely. The danger to humanity is not AI itself, but that humans who misjudge its capabilities will put it in charge of tasks it is inherently incapable of managing safely or accurately. Very clear and easy to understand even if you have no background with computer science or other relevant subjects. Highly recommended.

[Mikhail Filatov · Rating 2 out of 5]

While the author has a lot of valid points about the issues of modern technology, mostly AI and social networks, he is not objective so it’s difficult to trust him. He renamed old GOFAI (good old-fashioned AI) to “psychological AI” and tries to convince that it is not worse than “deep learning” and ML Unfortunately, he commits the same fallacies he himself describes. For example, he promotes “13 keys to the White House” even after acknowledging that originally it was about popular vote, so it did not predict the 2016 elections correctly as Clinton won popular vote. Typical “Texas sharpshooter” fallacy, but as it reminds him of his beloved expert systems the author just ignores it.

[Nick Lucarelli · Rating 3 out of 5]

Starts hot and peters out quickly.

The first half paints a fairly bleak picture about (pre ChatGPT) AI, implying that simple, transparent algorithms often do just as good a job as complex black box ones, and that AI in general will only thrive in stable, data rich spheres that often don't reflect (and do not exist in) real life (eg self driving cars on busy city roads).

The rest of the book veers into pop behavioural psych territory that is loosely tied to technology but which loses its initial focus on AI and its pitfalls.

Not bad overall. Made me more of an AI pessimist.