Summary

Deutsch argues that open-ended progress begins with good explanations – explanations with reach, hard-to-vary structure, and the power to solve problems beyond their original domain. The Enlightenment was the first sustained culture built around conjecture, criticism, error correction, and optimism. Knowledge grows when people create bold guesses and expose them to criticism.

The book ties epistemology, physics, computation, evolution, politics, beauty, culture, and AI into one claim: people are universal explainers and constructors. We are constrained by the laws of physics, but practically limited by knowledge. Problems are inevitable; problems are soluble. The future cannot be predicted because future knowledge has not yet been created, and that is why the beginning of infinity is a method – creativity + criticism – not a final destination.

“People who converge on the truth converge with each other.”

“If beauty is objective, then a new work of art - like a newly discovered law of nature or mathematical theorem - adds something irreducibly new to the world.”

“The essence of democratic decision making is not the choice made by the system at elections, but the ideas created between elections - and elections are one of the many such institutions whose function is to allow such ideas to be created, tested, modified, and rejected.”

“Although biologic and cultural evolution are described by the same underlying theory, the mechanisms of transmission, variation, and selection are all very different. That makes the resulting natural histories very different, too”

“Would we seem like insects to it? This can seem plausible if one forgets that there can only be one type of person: universal explainers and constructors”

“There is only one way of thinking that is capable of making progress, or of surviving in the long run, and that is the way of seeking good explanations through creativity and criticism.”

Chappy’s Review

One of the densest books I’ve ever read! Packed with fascinating interdisciplinary insight and a unique take on humans as universal explainers and constructors. Some criticism on his AI predictions.

Key Takeaways

Notes

Introduction

• “Progress that is both rapid enough to be noticed and stable enough to continue over many generations has been achieved only once in the history of our species”
  • Scientific revolution → now
• Central argument: all progress (practical + theoretical) has resulted from a single human activity: the quest for “good” explanations
• Progress has no end (infinity)

1: The Reach of Explanations

• Stars → supernovae → multiverse*
• Scientific theories are explanations
• Empiricism: belief that all theories are derived from our senses
• Actually, they’re just bold guesses
• Inductionism: Bayesian reasoning to repeated observations
• But 1) we’ve never observed a billion years and 2) things change in the future
• Senses are deceptive → misled by false interpretations
• Justificationism: how do we know → by what authority do we claim?
  • Confirmation bias
• Fallibilism: ↑ opposing: there are no authoritative sources of knowledge
• ↑ seeking to disprove and learn
• Needed to develop “a tradition of criticism”
• Predictions must be testable
• Instrumentalism: science doesn’t discover / “explain” reality, only predict outcomes
• Realism: true doctrine that the real world really exists and is accessible to rational inquiry
• Relativism: statements are not objectively true or false, just judged relative to arbitrary standard
• Occam’s razor:
• The best explanations are most constrained by existing knowledge
• Explanation: a statement about something that is there, what it does, and how and why
• Reach: the ability of some explanations to solve problems beyond those they were created to solve
• Creativity: the capacity to create new explanations
• Problem: created when a conflict of ideas is experienced
• The Enlightenment: the beginning of the way of pursuing knowledge with a tradition of criticism and seeking explanations instead of a reliance on authority
• Rational: attempting to solve problems by seeking good explanations
• The West: the political, moral, economic, and intellectual culture that has been growing around science, reason, and freedom

2: Closer to Reality

• Our observations are often deceptive (eg. telescopes)
• … (my notes got deleted?!)

3: The Spark

• Anthropic misconceptions
• Biosphere: earth is a spaceship life support system for humans
• Cultural evolution
• No evolutionary pressure for us to understand most things (too large or small)
• Eg. the brain is only so capable → we may cease to understand
• People: entities that can create explanatory knowledge
• People use knowledge to create automated physical transformations
• Need mass, energy, and evidence
• Must expand into the stars to ensure survival of species
• Problems are inevitable, soluble
• We can make knowledge just about anywhere
• Universal constructor: can cause any raw material to undergo any physically possible transformation given the right info

4: Creation

• Vestigial features disprove divine creation
• A good design is hard to vary
• Evolution creates knowledge
• Evolution based on selfish genes, not species wellbeing (The Selfish Gene)
  • Eg. Peacock tails
• Ideas / memes are replicators
• Though, memes are inexact and interpreted (vs genes exact code)
• Fine tuning → must be parallel universes
• Anthropic explanation: it is only in intelligent universes that people wonder why it is so
  • Predicts that the universe is only just capable of producing humans (lots of constants)
• Multiverse theory is a bad explanation
• “Science - and creative thought in general - creates is unpredictable creation ex Nilo. So does biological evolution. No other process does.”

5: The Reality of Abstractions

• Quasi-autonomous / nearly self-contained phenomena of given layer of abstraction
• Emergence: (levels of “”) can be explained well in terms of each other without realizing them as constituents
• Eg. explain why a copper atom is at the tip of Winson Churchill’s statue*
• Reductionism: must be explainable by fundamental components and thus no holistic explanations (false)
• Holism: everything is components of wholes (false)
• ToE: set of explanations that imply all other explanations
  • Doesn’t have to be atomic - can be high-level too
  • One solution to the fine tuning problem: if high-level explanations
• Possible example: universality of computation
• Abstractions / ideas / theories affect the physical world
• Abstract reasoning > physical reasoning in many cases
  • Eg. Hoffsteader’s dominos, computation / functions
• Moral philosophy: what sort of life to want
• “Moral philosophy is basically about the problem of what to do next, and more generally what sort of life to lead than what sort of world to want”
• Utilitarianism: defining morals based on their utility in meeting people’s preferences
• Universality: able to understand explainations

6: The Jump to Universality

• Pictography → phonetic alphabet (universality)
• Phoenician alphabet was first time ↑ was ever used solely / primarily
• “A small change in a system to meet a parochial purpose just happened to make the system universal as well”
• Roman numerals as universal in limited bound for arithmetic
• Universal printing system (Gutenberg) changed everything
• Analytic functions can all be approximated using addition and multiplication → universal math
• Universal classical computer: ↑ with control flow
• Computer / analytical engine ignored for 100 years
• Analog → digital: universal, faster, deterministic
• Error correction is the beginning of infinity
• Replicators: group selection → RNA → DNA
• ↑ DNA system itself flatlined, but took us from bacteria to human society
• Is DNA universal? If so, what was the jump?
• Universal classical computer made with DNA, but not unlocked naturally yet
• Humans are uniquely universal explainers, constructors

7: Artificial Creativity

• Turing test and AI progress (slow)
• Turing estimated only need 100mb memory and 10,000 ops/s
• Qualia: neither describable nor predictable
• David Deutsch: exciting that we haven’t learned to describe qualia; Daniel Dennett: there is no qualia - more like a myth we still have
• AI will probably be a jump to universality
• Behaviorism: if you can behave like you’re thinking, then you must be thinking
  • Instrumentalism applied to psychology
• Lamarkism: AI chatbots
• Genetic algos
• AGI stuck because we don’t know how to solve creativity
• We don’t understand the nature of the universality of DNA

8: A Window on Infinity

• Progress will be unbounded (potential infinity) or an infinite amount of progress will be made (realized infinity)
• Infinite: (mathematical) set that can be placed 1:1 correspondence with itself
• Infinite or bounded reason?
  • Implies supernatural
• Infinity is studied via infinite sets
• Requirement for self-consistency, not physical possibility
• Sizes of infinity
  • Eg. natural #s vs points on a line
• Singularity:
• Infinite regress: invalid move with infinity to avoid real explanation
  • Eg. turtles all the way down
• Because of infinite universes, cannot make Anthropic predictions
• Common mis-analogizing between abstract and physical concepts
• Non-Euclidean space
• Most functions are non-computable
• Proof that abstractions exist objectively*
• Laws of physics determine atomic logical operations like and, not, etc.
• Computation / proof: “a physical process in which objects such as computers or brains physically model or instantiate abstract entities like numbers or equations and mimic their properties”
• Proof theory is computer science, not mathematics
• P != NP “efficient computation”
• Even with “perfect knowledge” there are limits, eg:
  • Most ancient history details (entropy destroys info)
  • Laws of physics (speed of light)
  • Epistemology (conjecture + criticism)
  • Errors are inevitable
• 3 corollaries:
  • Insoluble questions are inherently uninteresting
  • What is interesting vs boring is objective fact in the long run
  • Every interesting problem is soluble

9: Optimism

• Future of civilization based on human choices - not random chance
• We can’t possibly predict future progress
• “What is the right philosophy of the unknown future?”
• “A blind pessimist argues that there is an inherent asymmetry between good and bad consequences… a single catastrophic consequence of an otherwise beneficial innovation could put an end to human progress forever”
• Precautionary principle: ↑ ↓
• Many past civilizations were destroyed by simple technologies - fire, sword, natural disaster
• Wealth: the repitour of physical transformations capable of causing
• We’re more limited by knowledge than raw materials
• “Would we seem like insects to it? This can seem plausible if one forgets that there can only be one type of person: universal explainers and constructors”
• “How can we hope to detect and eliminate error?”
• “Who should rule?” approach to political philosophy misguided
• The principle of optimism: all evils are caused by insufficient knowledge
• “We have a duty to be optimistic in general and about civilization in particular”
• Many pockets of Enlightenment thinking before finally taking grasp
• Blind optimism: reckless overconfidence - proceeding as if one new bad outcomes won’t happen
• Blind pessimism: avoiding everything not known to be safe

10: A Dream of Socrates

• Socrates + Hermes / Apollo
• Seeking to discover the truth about how the world is, why, and how it should be (objective knowledge)
• ↑ vs justified beliefs (infinite regress + error)
• Knowledge via justified beliefs is a chimera / self-deception / unattainable
• Reason / morals is more objective than self-observation since it’s abstract and independent of deceptive senses
• Culture of critical debate
• All knowledge comes from (self) persuasion
• “A city with a philosopher king would be truly great”
• The Socratic problem: all recounts of Socrates are secondhand, so may be misinterpreted or misattributed
• “People who converge on the truth converge with each other”

11: The Multiverse

• The many universes interpretation of quantum theory
• Properties of universe = conditions of the flow of information
• Universe = instances of the multiverse
• Dollars are abstract replicators
• Varying definitions of fungibility (for language)
• 3 reasons for unpredictability:
  1. Affected by 1 fundamentally random indeterministic variable (none in “real physics”)
  2. Factors are either unknown or too complex to account for (especially knowledge creation)
  3. Deterministic laws of physics + quantum = god plays dice (fungible → diverge)
• Sphere of differentiation - butterfly effects
  • Not infinite given “1 quantum” min excitation of electron etc.
• Unaware of event context within broader context (eg. watches stopping every minute globally → “unlikely yet certain”)
• Quantum interference re-unites parallel universes (evidence of it) (entanglement)
• Quantum decoherence
• Heisenberg uncertainty principle: for any fungible collection of instances of a physical object, some of attributes must be diverse
• ↑ + electromagnetic attraction = hydrogen atom
• Electron field + waves
• Quantum computation: flow of information is not confined to a single history
• “All fiction that does not violate the laws of physics is fact”
• “We are channels of information flow. So are histories, and so are all relatively autonomous objects within histories”
• Entanglement: info determining which multiversal objects can affect others
• Decoherence: the process of becoming infeasible to undo the effect of a wave of differentiation between universes
• Quantum interference: non-fungible objects becoming fungible
• “A universe is an emergent feature of the multiverse”

12: A Physicist's History of Bad Philosophy

• People are “an emergent quasi-autonomous flow of information in the multiverse”
• Whenever a measurement is made, all the histories but 1 cease to exist
• Bad philosophy: not only false but actively prevents the growth of competing theories
• Principle of complementarity: particle-wave duality means can only measure + talk about one attribute at a time
• Said that human consciousness caused transition - acting at a fundamental level of physics (Copenhagen interpretation)
• ↑ common stance: “if you think you’ve understood quantum mechanics, then you don’t” - definition of bad philosophy
• “Because I said so” - ↑ another example of bad philosophy
• Reinterprets truth seeking as justification
• Postmodernism / deconstructionism / structuralism: all ideas are conjectural, thus all narratives
• ↑ rejects all criticism as narrative
• Proxies + explanatory theories in science (eg. measuring happiness)
• Explanation-less science inhibits progress:
  • Suppresses philosophical debate by pretending solved by science (eg. animal suffering)
  • Amplifies errors
• With the Enlightenment, bad philosophy became much worse: empiricism → positivism → logical positivism → instrumentalism → linguistic philosophy → postmodernism

13: Choices

• Representative government: composition or opinions of legislature reflect those of the people
• Social choice theory: the study of how the will of society can be defined in terms of the wishes of its members and of what social institutions can enact its will, thus defined
• Apportionment problems of US representational seats
• Apportionment paradoxes
• Why does politics insist on democratic rule when science etc. would never work that way?
• Theorem: “every apportionment rule that stays within the quota suffers from the population paradox”
• ”Which option can be rationally defended as being the will of the people? The option that society wants?”
• “How should society organize its decision making so it does indeed choose the options that it wants?”
• ↑ led to game theory
• 5 axioms for “the will of the people”:
  1. Defined only in terms of that group’s members
  2. Must not just be the views of one member
  3. If group is unanimous, then must reflect group
  4. If some in group move to agreeing, must stay the group’s pick (else, population paradox)
  5. Should stay consistent when changing mind on unrelated choice
• ↑ proved that these are logically inconsistent (Nobel prize)
• ↑ Thus, no way for a society to make self-consistent choices
• Basically no group decision making system without paradoxes
• However, ↑ ignores the problem solving phase of decision making: creation of new options
• Zeno’s mistake: mistaking an abstract process for a real life process of the same name
  • ↑ in this case decision making
• “The essence of democratic decision making is not the choice made by the system at elections, but the ideas created between elections - and elections are one of the many such institutions whose function is to allow such ideas to be created, tested, modified, and rejected.”
• “Why don’t they just fix social choice theory by including creative processes such as explanation and persuasion in its mathematical model of decision making? Because it is not known how to model a creative process. Such a model would be a creative process - an AI.”
• Compromise policies are bad because the prevent accountability, learning
• Plurality > proportional representation*
• “Convergence in the broad consensus over time is made possible by the fact that all concerned are gradually eliminating errors in their positions and converging on objective truths”

14: Why are Flowers Beautiful?

• Deep truth is often beautiful + elegant
• Is it rhythm? Uniqueness?
• Is it all cultural relativism?
• Is it structure / intelligent design?
• “If beauty is objective, then a new work of art - like a newly discovered law of nature or mathematical theorem - adds something irreducibly new to the world”
• Attraction is learned
  • Eg. skydiving despite fear of falling
• What is profound is that flowers attract both insects (co-evolved) and humans
• “My guess is that the easiest way to signal such a gap [between species] is to use objective standards of beauty”
• Theory: human appearance satisfies standards of objective beauty as well species-specific
• Two types of beauty problems:
  • Pure: creating beauty for its own sake (eg. art standards)
  • Applied: signaling information (eg. UI/UX, propaganda)
• Aesthetics: the philosophy of beauty
• Elegance: the beauty in explanations of theories etc.

15: The Evolution of Culture

• Sets of ideas = culture
• Includes inexplicit (tacit)
• Memes: ideas that are replicators
• What is it about long-lived memes that gives them exceptional ability to resist change throughout many replications?
  • When they do change, under what conditions can they change for the better?
• “Although biologic and cultural evolution are described by the same underlying theory, the mechanisms of transmission, variation, and selection are all very different. That makes the resulting natural histories very different, too”
• Creativity is ↑ within a single brain
• Memeplex: a collection of ideas
  • Eg. quantum theory
• Memes have to be expressed as behavior (unlike genes)
• ↑ become simultaneously encoded in physical form as brain state and behavior
  • Potential 3rd: written
• Ideas must be persuasive, teachable
• Most successful memes are best as displacing others from the population (limited attention)
• Selection + evolution happens within a brain, too
• Memetic evolution far faster vs genetic
• Static society’s fear that any change is much more likely to do harm than good
• Creating knowledge is a natural human need + desire, so ↑ is unnatural suppression
• Useful truths are most likely to replicate
  • Eg. interesting, funny, elegant, easily remembered, morally right
• Rational meme: idea that relies on the recipients critical faculties to cause itself to be replicated
• Rational + anti-rational memes
  • Eg. gender roles are anti-rational
• Anti-rational memes discourage critical thinking about choices (low individuality / agency?)
• We’re in a transitional period from a static to fully dynamic society
• ↑ causes phenomena peculiar to ↑:
  • Anti-rational occasionally evolve against the grain, eg. autocratic → constitutional monarchy
  • Formation of anti-rational subcultures (eg. racism)
  • Education is still like fixed society predecessors
• Creatively: “emerge from individual minds as novel ideas, having acquired significant adaptations within those minds”

16: The Evolution of Creativity

• What were the original uses for creativity that resulted in evolutionary selection?
• How ↑ when there was so little sustained / compounding innovation?
• Sexual selection as cause unlikely
• Actually, enhanced creativity can help one to be more conformist
• In fact, even imitation / conformity requires incredible creativity
• Eg. Popper’s teaching “observe” - “scientific observation is impossible without preexisting knowledge about what to look at, what to look for, how to look, and how to interpret what one sees”
  • Theory has to come first
• We know the rules of our behavior + language largely Implicitly
• Parrots etc. use mirror neurons - but “cannot possibly” be a key to understanding human behavior
• Parrots are devoid of creativity, so must imitate sound naively without meaning
• Imitation = copying behavior; learning = copying knowledge causing behavior
• Apes leverage behavior parsing to learn rather than actual explanations
  • Basically just sequencing inborn behaviors
• Parroting language = Ape-ing behavior
• “Imitation is not at the heart of human meme replication” - eg. learning from lecture without being able to recite any sentences
• Lecture listeners seek good explanations + ideas
• Creativity = the ability to both create and replicate memes
• Capacity for universal explanation evolved simply to ↑ volume of memetic info we can acquire (+ speed, accuracy)
• Memes gradually came to dominate behavior
• Genes + memes co-evolution:
  • Genes evolving hardware to support more and better memes
  • Memes evolving to take over more previously genetic functions
• Blackmore’s brain as meme machine must be true
• People = creative universal explainers

17: Unsustainable

• Our civilization is unique in its history to make progress
• Progress often results in seemingly unsustainable trends (eg. population growth)
  • New problems necessitate progress
• “The problem is not making mistakes but more so failing to correct them”
• Guns, Germs and Steel over-indexes on geography, under on creativity + ideas
• A culture of optimism + criticism is the bedrock
• 1970s argument that we’re reaching numerous environmental carrying capacities
• → sustainability is the disease, people are the cure
• Should instead focus on our ability to put things right if they go wrong
• Focus on plans to ↓ temps + thrive at ↑ temp efficiently + cheaply, not force reductions in emissions at all costs

18: The Beginning

• Astronomy and technology have made the world feel small
• Believing in finite knowledge (in universe or for human utopia) is bad
• Optimism + criticism
• “Science claims neither infallibility nor finality”
• More then ever before we appreciate how little we know
• Future of the universe / physics*
• Given expansion of universe, eventually everything possible will be visible (in principle)
• Quantum suicide
• Simulation theory
• Doomsday argument: kind of like Lindy effect - 1/2 of all humans born
• Inevitable future of humanity is immortality (eg. via brain backup)
• Technological singularity
• Humans are already universal explainers and constructors
• There will be ever more efficient computer interfaces (→ BCIs)
• “Universality implies that in every important sense, humans and AIs will never be anything more than equal”
• Brain add-ons will help us keep up with pace of tech change
• Theory: there have been several singularities, eg. Industrial Revolution and the Enlightenment
• “We know people are of fundamental significance, but we do not know if we are among these people”
• “There is only one way of thinking that is capable of making progress, or of surviving in the long run, and that is the way of seeking good explanations through creativity and criticism.”