Book review: Principles: Life and Work, by Ray Dalio.
Most popular books get that way by having an engaging style. Yet this book’s style is mundane, almost forgetable.
Some books become bestsellers by being controversial. Others become bestsellers by manipulating reader’s emotions, e.g. by being fun to read, or by getting the reader to overestimate how profound the book is. Principles definitely doesn’t fit those patterns.
Some books become bestsellers because the author became famous for reasons other than his writings (e.g. Stephen Hawking, Donald Trump, and Bill Gates). Principles fits this pattern somewhat well: if an obscure person had published it, nothing about it would have triggered a pattern of readers enthusiastically urging their friends to read it. I suspect the average book in this category is rather pathetic, but I also expect there’s a very large variance in the quality of books in this category.
Principles contains an unusual amount of wisdom. But it’s unclear whether that’s enough to make it a good book, because it’s unclear whether it will convince readers to follow the advice. Much of the advice sounds like ideas that most of us agree with already. The wisdom comes more in selecting the most underutilized ideas, without being particularly novel. The main benefit is likely to be that people who were already on the verge of adopting the book’s advice will get one more nudge from an authority, providing the social reassurance they need.
Some of why I trust the book’s advice is that it overlaps a good deal with other sources from which I’ve gotten value, e.g. CFAR.
Key ideas include:
- be honest with yourself
- be open-minded
- focus on identifying and fixing your most important weaknesses
Eric Drexler has published a book-length paper on AI risk, describing an approach that he calls Comprehensive AI Services (CAIS).
His primary goal seems to be reframing AI risk discussions to use a rather different paradigm than the one that Nick Bostrom and Eliezer Yudkowsky have been promoting. (There isn’t yet any paradigm that’s widely accepted, so this isn’t a Kuhnian paradigm shift; it’s better characterized as an amorphous field that is struggling to establish its first paradigm). Dueling paradigms seems to be the best that the AI safety field can manage to achieve for now.
I’ll start by mentioning some important claims that Drexler doesn’t dispute:
- an intelligence explosion might happen somewhat suddenly, in the fairly near future;
- it’s hard to reliably align an AI’s values with human values;
- recursive self-improvement, as imagined by Bostrom / Yudkowsky, would pose significant dangers.
Drexler likely disagrees about some of the claims made by Bostrom / Yudkowsky on those points, but he shares enough of their concerns about them that those disagreements don’t explain why Drexler approaches AI safety differently. (Drexler is more cautious than most writers about making any predictions concerning these three claims).
CAIS isn’t a full solution to AI risks. Instead, it’s better thought of as an attempt to reduce the risk of world conquest by the first AGI that reaches some threshold, preserve existing corrigibility somewhat past human-level AI, and postpone need for a permanent solution until we have more intelligence.
Descriptions of AI-relevant ontological crises typically choose examples where it seems moderately obvious how humans would want to resolve the crises. I describe here a scenario where I don’t know how I would want to resolve the crisis.
I will incidentally
ridicule express distate for some philosophical beliefs.
Suppose a powerful AI is programmed to have an ethical system with a version of the person-affecting view. A version which says only persons who exist are morally relevant, and “exist” only refers to the present time. [Note that the most sophisticated advocates of the person-affecting view are willing to treat future people as real, and only object to comparing those people to other possible futures where those people don’t exist.]
Suppose also that it is programmed by someone who thinks in Newtonian models. Then something happens which prevents the programmer from correcting any flaws in the AI. (For simplicity, I’ll say programmer dies, and the AI was programmed to only accept changes to its ethical system from the programmer).
What happens when the AI tries to make ethical decisions about people in distant galaxies (hereinafter “distant people”) using a model of the universe that works like relativity?
Book review: Artificial Intelligence Safety and Security, by Roman V. Yampolskiy.
This is a collection of papers, with highly varying topics, quality, and importance.
Many of the papers focus on risks that are specific to superintelligence, some assuming that a single AI will take over the world, and some assuming that there will be many AIs of roughly equal power. Others focus on problems that are associated with current AI programs.
I’ve tried to arrange my comments on individual papers in roughly descending order of how important the papers look for addressing the largest AI-related risks, while also sometimes putting similar topics in one group. The result feels a little more organized than the book, but I worry that the papers are too dissimilar to be usefully grouped. I’ve ignored some of the less important papers.
The book’s attempt at organizing the papers consists of dividing them into “Concerns of Luminaries” and “Responses of Scholars”. Alas, I see few signs that many of the authors are even aware of what the other authors have written, much less that the later papers are attempts at responding to the earlier papers. It looks like the papers are mainly arranged in order of when they were written. There’s a modest cluster of authors who agree enough with Bostrom to constitute a single scientific paradigm, but half the papers demonstrate about as much of a consensus on what topic they’re discussing as I would expect to get from asking medieval peasants about airplane safety.
Book(?) review: The Great Stagnation: How America Ate All The Low-Hanging Fruit of Modern History, Got Sick, and Will (Eventually) Feel Better, by Tyler Cowen.
Tyler Cowen wrote what looks like a couple of blog posts, and published them in book form.
The problem: US economic growth slowed in the early 1970s, and hasn’t recovered much. Median family income would be 50% higher if the growth of 1945-1970 had continued.
Book review: Where Is My Flying Car? A Memoir of Future Past, by J. Storrs Hall (aka Josh).
If you only read the first 3 chapters, you might imagine that this is the history of just one industry (or the mysterious lack of an industry).
But this book attributes the absence of that industry to a broad set of problems that are keeping us poor. He looks at the post-1970 slowdown in innovation that Cowen describes in The Great Stagnation. The two books agree on many symptoms, but describe the causes differently: where Cowen says we ate the low hanging fruit, Josh says it’s due to someone “spraying paraquat on the low-hanging fruit”.
The book is full of mostly good insights. It significantly changed my opinion of the Great Stagnation.
The book jumps back and forth between polemics about the Great Strangulation (with a bit too much outrage porn), and nerdy descriptions of engineering and piloting problems. I found those large shifts in tone to be somewhat disorienting – it’s like the author can’t decide whether he’s an autistic youth who is eagerly describing his latest obsession, or an angry old man complaining about how the world is going to hell (I’ve met the author at Foresight conferences, and got similar but milder impressions there).
Josh’s main explanation for the Great Strangulation is the rise of Green fundamentalism, but he also describes other cultural / political factors that seem related. But before looking at those, I’ll look in some depth at three industries that exemplify the Great Strangulation.
Book review: The Book of Why, by Judea Pearl and Dana MacKenzie.
This book aims to turn the ideas from Pearl’s seminal Causality into something that’s readable by a fairly wide audience.
It is somewhat successful. Most of the book is pretty readable, but parts of it still read like they were written for mathematicians.
History of science
A fair amount of the book covers the era (most of the 20th century) when statisticians and scientists mostly rejected causality as an appropriate subject for science. They mostly observed correlations, and carefully repeated the mantra “correlation does not imply causation”.
Scientists kept wanting to at least hint at causal implications of their research, but statisticians rejected most attempts to make rigorous claims about causes.
No, this isn’t about cutlery.
I’m proposing to fork science in the sense that Bitcoin was forked, into an adversarial science and a crowdsourced science.
As with Bitcoin, I have no expectation that the two branches will be equal.
These ideas could apply to most fields of science, but some fields need change more than others. P-values and p-hacking controversy are signs that a field needs change. Fields that don’t care much about p-values don’t need as much change, e.g. physics and computer science. I’ll focus mainly on medicine and psychology, and leave aside the harder-to-improve social sciences.
What do we mean by the word Science?
The term “science” has a range of meanings.
One extreme focuses on “perform experiments in order to test hypotheses”, as in The Scientist In The Crib. I’ll call this the personal knowledge version of science.
A different extreme includes formal institutions such as peer review, RCTs, etc. I’ll call this the authoritative knowledge version of science.
Both of these meanings of the word science are floating around, with little effort to distinguish them . I suspect that promotes confusion about what standards to apply to scientific claims. And I’m concerned that people will use the high status of authoritative science to encourage us to ignore knowledge that doesn’t fit within its paradigm.
Book review: Surfing Uncertainty: Prediction, Action, and the Embodied Mind, by Andy Clark.
Surfing Uncertainty describes minds as hierarchies of prediction engines. Most behavior involves interactions between a stream of information that uses low-level sensory data to adjust higher level predictive models of the world, and another stream of data coming from high-level models that guides low-level sensory processes to better guess the most likely interpretations of ambiguous sensory evidence.
Clark calls this a predictive processing (PP) model; others refer to is as predictive coding.
The book is full of good ideas, presented in a style that sapped my curiosity.
Jeff Hawkins has a more eloquent book about PP (On Intelligence), which focuses on how PP might be used to create artificial intelligence. The underwhelming progress of the company Hawkins started to capitalize on these ideas suggests it wasn’t the breakthrough that AI researchers were groping for. In contrast, Clark focuses on how PP helps us understand existing minds.
The PP model clearly has some value. The book was a bit more thorough than I wanted at demonstrating that. Since I didn’t find that particularly new or surprising, I’ll focus most of this review on a few loose threads that the book left dangling. So don’t treat this as a summary of the book (see Slate Star Codex if you want that, or if my review is too cryptic to understand), but rather as an exploration of the questions that the book provoked me to think about.
Scott at Slate Star Codex made some five year predictions.
I want to encourage a habit of people making long-term forecasts which include quantitative statements, so I’ll give probabilities for most of his predictions, comment on some of his qualitative predictions, then add some predictions of my own.