A Window on Intelligence: The Philosophy of People, Software, and Evolution – and Its Implications

by Dennis Hackethal

The human mind is the single most powerful entity in the universe. Yet we have made no progress in our efforts to simulate it as artificial general intelligence. Why is that? In this groundbreaking book, software engineer and philosopher Dennis Hackethal explains the mistakes intelligence researchers have been making – and how to avoid them. Based out of Silicon Valley, he proposes a research program for building truly intelligent machines. He argues for a fundamental unification of software engineering and reason generally that will aid greatly in our goal to simulate intelligence. Taking you on a journey through several fields, including the theory of evolution, epistemology, psychotherapy, and astronomy, Hackethal provides insight into the unlimited potential of artificial general intelligence that may one day take us to the stars. A Window on Intelligence is your field guide to the exciting world of your mind.

Genres: Philosophy

298 pages, Kindle Edition

Published March 13, 2020

Reviews

[Dan Elton · Rating 2 out of 5]

This book is an outgrowth of a much better book, "The Beginning of Infinity", by David Deutsch. Therefore, there is not much point reading this book until you've read that work.

The field of AI is closely related to epistemology. We won't be able to understand AI until we can agree on what knowledge is and how it may be acquired. This is a foundational issue and I agree with the author that Popper's epistemology is a strong foundation for thinking about AI which too few people are using. Thus, we need more works like this, describing Popper's ideas and how they can provide a foundation for understanding intelligence.

Alas, the author does not seem to have engaged with many other ideas outside of the narrow viewpoints of Popper and Deutsch. The density of errors and straw-men is very high in this book. I can't recommend this book generally. However, I think if you're interested in this particular area, it can help stimulate ideas because Hackenthal is engaging along intellectual tracts that are barely trodden, and often forging boldly in new directions of thinking.

Here's an example of a straw man: "This idea is known as instrumentalism, which is one way of denying reality and the systematic refusal to ask "why", that is, to seek explanations, and is false because there cannot be such a thing as an "explanationless prediction" -- wrong because instrumentalists due believe reality exists and we can understand it in the sense of predictions. Explanationless predictions can exist, see machine learning.

Some points are overstated. For instance, on pg. 26 - " Whether you are copying or innovating: both follow the same logic". Clearly a distinction can be made between the two.

Many highly questionable assertions made with great confidence but not well defended :
"dualism is not controversial among software engineers" (pg 96)
"I disagree with Popper .. that books literally contain knowledge" (pg 145)
"OpenAI's definition of AGI is wrong. People are not 'highly autonomous', they are entirely autonomous" (pg 161)
"curing mental ailments is a software engineering problem" (pg 236)
"AGI does not replace us in any capacity for nothing could replace us. It is one of us" (pg 204)
"There cannot be what Bostrom calls a singleton.." (pg 205)
"Intelligence is not a continuum - you either have it or you don't" (pg. 209 and other places, there are many ways to argue this is false which he doesn't address).
"Something which blindly follows instructions to make paperclips is not intelligent..." (pg 194)

The biggest errors in this book, in my view, are his redefinition of free will, his complete rejection of the biologically inspired AI (ala Jeff Hawkins) as a legitimate research direction, his dismissal of reductionism, and his theory of consciousness. The discussion on free will is very shallow -- determinism vs indeterminism is besides the point. "Free will" debates are mostly definitional debates, thus it is important to distinguish compatibalist vs libertarian notions of free will. That's a foundation that needs to be laid before you can argue for any one definition of free will or reject it outright (the view I hold). The discussion of reductionism skirts with promoting the mystical idea that emergent phenomena can "bend back" and influence lower level phenomena. There is no known mechanism by which that might occur. In fact, the entire notion of causality is on very shaky ground, an issue he doesn't engage with.

As with the discussion on free will, the discussion of consciousness is also very shallow. Apparently the author has never tried meditation - if he had he'd see clearly that there is such a thing as a "pure field of awareness" independent of problem solving or error correction. With meditation one also can see that problem solving happens unconsciously, too - thoughts and solutions just pop into the field of awareness.

While clearly I have much beef with this book, I still give it a solid 3 stars for exploring intelligence through a Popperian framework. Even if you don't agree with everything he says, reading this can help stimulate new directions of thought, at the very least through the exercise of thinking about how he's wrong.

[Christopher Elliott · Rating 4 out of 5]

Has some excellent ideas on how to fix some of the philosophical misconceptions stopping us from building AGI. Mapping some of the more precise language of David Deutsch into software engineering terms.

[Elliot · Rating 1 out of 5]

Plagiarized

This book contains extensive plagiarism of Elliot Temple (me) and David Deutsch (my colleague and mentor). Deutsch and I developed and published many of the ideas in the book which are presented as Dennis Hackethal’s own ideas.

Dennis learned not only from reading my work and watching my videos, but also from a personal mentoring relationship. I personally helped him learn about topics like universality, animal intelligence, Popper, Deutsch, and structural epistemology (the multiplication stuff). Then he plagiarized me on those topics and others. There are public records of many forum posts where I helped him learn this stuff over the 5 months leading up to around when he stopped speaking to me without explanation and started writing the book, around a year ago. I also helped teach Dennis in private chatrooms and on voice calls. My name does not appear in the book once.

Here is an example of the plagiarism. Me teaching Dennis: “X is a universal Y if it can do any Z that any other Y can do.” We had a lengthy email discussion about this idea and a voice call about it too. Dennis struggled to understand it.

Then the book says, “Criterion of universality – x is a universal y if it can do all the z’s all the other y’s can do”. No credit is given and it’s presented as Dennis’ own words, not as a quote.

Ideas plagiarized from Deutsch’s books, with zero credit, include: problems are soluble, problems are inevitable, the jump to universality, reach, and criteria for reality.

I initially contacted Dennis about the plagiarism privately, even though he didn’t even notify me the book existed even after publication. He responded indicating that he had gotten a lawyer and would not read my explanation of what he plagiarized. Then my website went down due to a DDOS attack.

[Connor · Rating 1 out of 5]

Like others have noted, this book is largely plagiarized ideas. Even aside from that it reads as if somebody copy and pasted a bunch of blogs into a book, which is exactly what I suspect happened here. Don't support this kind of low quality garbage.