The Myth of Artificial Intelligence

Why Computers Can’t Think the Way We Do

De: Erik J. Larson

Narrado por: Perry Daniels

Duración: 10 h y 9 m

Resumen del Editor

Futurists insist that AI will soon eclipse the capacities of the most gifted human mind. What hope do we have against superintelligent machines? But we aren't really on the path to developing intelligent machines. In fact, we don't even know where that path might be.

Erik Larson takes us on a tour of the landscape of AI to show how far we are from superintelligence and what it would take to get there. Ever since Alan Turing, AI enthusiasts have equated artificial intelligence with human intelligence. This is a profound mistake. AI works on inductive reasoning, crunching data sets to predict outcomes. But humans don't correlate data sets: We make conjectures informed by context and experience. Human intelligence is a web of best guesses, given what we know about the world. We haven't a clue how to program this kind of intuitive reasoning, known as abduction. Yet it is the heart of common sense. That's why Alexa can't understand what you are asking and why AI can only take us so far.

Larson argues that AI hype is both bad science and bad for science. A culture of invention thrives on exploring unknowns, not overselling existing methods. Inductive AI will continue to improve at narrow tasks, but if we want to make real progress, we will need to start by more fully appreciating the only true intelligence we know - our own.

©2021 Erik J. Larson (P)2021 Tantor

Outstanding Counterpoint to the AI craze

If this guy is half right, AI will become the next dot bomb. And I can’t find evidence he’s wrong.

Must read on AI

Read this book if you want to cut through the hype on AI. It is insightful and well written.

A critique of contemporary AI

A too-short synopsis:

->(1) Ray Kurzweil, Nick Bostrom, Max Tegmark, etc., all erroneously believe that humanity is on track to develop human level, or above, AI.

->(2) We do not know how to automate novel inference, called 'abduction'. Although existing AI can solve various problems using deductive and inductive logic, it is really bad at inventing things out of the blue.

->(3) Charles Sanders Peirce coined the term abduction to mean the "logic" of coming up with novel hypotheses.

->(4) We don't know how to automate abduction, much less formally define it.

->(5) We can't go from mere deduction and induction, to abduction.

->(6) We are NOT on track to AGI. We are still at the whiteboard, guessing. We have no roadmap.

->(7) There is a mythology surrounding AI in the human social order. It is symbolic. Symbols also shape the direction society goes.

No one else has written a popular book that philosophically critiques contemporary AI as successfully as this one. Moreover, I agree with the approach taken in the book.

There were some aspects of the book that were slightly disappointing:

->(disappointment 1) I would like a deeper dive into the philosophical aspects... which is a tall order, especially in human language.

->(disappointment 2) I suspect that Larson could have better steel-manned some of the ideas he criticized. I do not appreciate appeals to common sense, as that seems like handwaving. Maybe I missed something, but I think his argument could have been further developed.

I guess I'll be reading Peirce, and reviewing automata theory and AIXI. But I am grateful for this book, and I think it deserves 5 out of 5. Or at least closer to a 5 than a 4. Because what counts here is the overall gist of the argument, which in this case is extremely on point, and does not conform to the contemporary narrative.

Intelligent and Convincing

This book presents a cogent and compelling counterargument to the popular current theories of impending AI dominance of the human race. Unlike those who warn of a coming AI apocalypse the arguments are well reasoned and sound. Humans still do not have much competition in the areas of creative intelligence and the apparently simple process of understanding.

I thoroughly enjoyed this book and strongly recommend it to all and especially Kurzweill, Musk and Bostrum.

Great information and explanation

Does a thorough and entertaining job at discussing the hype and mythology around AI and why it is hurtful in the long run (and even prevents progress in AI science).

Informative, Thought-provoking

Anyone interested in getting a real view of where we are in the development of artificial intelligence needs to read this book!

A path forward?

Larson gives us an excellent look at the current state of AI (prior to ChatGPT) and the myriad challenges and yes – even intractable problems — facing the field today. Though I disagree with his absolute pessimistic outlook, I thoroughly appreciate his highlighting of the areas where we need much further study and basic research. And he does so with a style and substance of writing that makes it a pleasure to read. 

Very Good

A sobering guide to a modern uncertainty. Definitely think it is more honest than Ray Kurzweil’s works.

great overall but not up to date

The author provides an excellent overview of the limits of AI and, more importantly, our own understanding of Intelligence.

spoiler alert: no one has a clear, programmable version of Intelligence to create.

However, several of his examples that "can't be done by current AI" are easily handled by ChatGPT 4.

Excellent

Good overview on the weaknesses of reaching general artificial intelligence. In summary, it is not going to happen unless Hume’s empirical theory of causality is shown to be wrong.