Lossy is one long argument, but it is built out of about 140 separate ideas, and you do not have to meet them in order. Every idea has its own short page: one plain sentence, an example, a few paragraphs, and links to the ideas next to it. This page is the map of all of them.
The 13 themes below run in roughly the order of the argument, so reading down the page follows the story. Inside a theme, start anywhere. When an idea is worth the long version, its page points to where it is argued in full.
Prefer the story from the top? Take the reading path: every chapter in order, each in a sentence or two. Short on time? The short version is the whole argument in a page.
1. How ideas travel
An idea reaches you through a chain of retellings, and two things happen at every link: it gets repackaged, and it gets filtered. Open this theme →
Ideas: Transport · Selection · Memetic fitness · Option space · Manufactured injection · The three-layer message
Examples: The telephone game · Astrology · ‘Scientists say X’ headline · The Voyager Golden Record
Open questions: Transport vs. selection: which is primary?
2. The medium sets the rules
The format an idea travels in decides which ideas can win, before anyone makes a choice. Open this theme →
Ideas: Technology vs. medium · Medium shapes want · Want as the prime mover · Algorithm as selection engine · Manipulation surface
Examples: Hammer of the Witches vs. Copernicus
Open questions: What does the medium do to the gates? · Does a medium forbid thoughts, or only price them?
3. Thinking inside frameworks
The tools you think with, from languages to slide decks, make some thoughts cheap and others expensive before you say a word. Open this theme →
Ideas: Framework · Framework pricing · Cost asymmetry · Defaults · Notation as a tool of thought · Linguistic relativity · Influence, not imprisonment · Framing effect · Conceptual metaphor · Theory-ladenness and paradigms · Einstellung effect · Law of the instrument · Inductive bias
Examples: The Columbia slide · Leibniz’s dy/dx and the British lag · Russian blues · Pirahã and exact number · The Asian disease problem · Crime as beast, crime as virus · Chess experts and the Einstellung effect · Prompt formatting and the 76-point swing
4. When the map moves the territory
Once people act on a simplified picture of the world, the world starts changing to fit the picture. Open this theme →
Ideas: The map is not the territory · Legibility · Classification torque · Framework feedback · Goodhart’s law · Performativity · Looping effects · Conway’s law · Path dependence and lock-in · Framework rotation
Examples: Scientific forestry and the Normalbaum · Apartheid race classification · The DSM and RDoC · AI benchmarks as milestones · Black-Scholes and the option market · The Gaussian copula and 2008 · QWERTY and the disputed lock-in story
5. The finite reader
You have a limited number of hours for hard ideas, which is why everything has to be shortened to reach you. Open this theme →
Ideas: Receiver budget · The abyss · Complex form · Compressed form · Complexity ceiling · Handle-ability · Transferable preconditions
Examples: matplotlib vs. seaborn · Latin Mass vs. vernacular liturgy · A Brief History of Time · Statutes vs. plain-language legal summaries · Bitcoin
Open questions: Are receiver budgets fixed or trainable? · Is ‘complexity’ doing two jobs?
6. What shortening does to truth
A shortened claim loses its reading instructions, so it can land as roughly right, as the opposite of what was found, or as a team badge. Open this theme →
Ideas: Truth-compression regimes · Interpretive latitude
Examples: Power posing · Arsenic-life · Stanford Prison Experiment correctives
Open questions: Where does truth value enter the model?
7. Truths made by agreement
Money, laws and nations are real because people agree they are, and for truths like that, spreading the word is part of what makes them true. Open this theme →
Ideas: Three realities · Truth generators · Constitutive transmission · Constitutive regimes · Reification
Open questions: What does the pipeline do to intersubjective truths?
8. What the filter rewards
What spreads is whatever the filter rewards, and the evidence on what that is turns out to be stranger than “people like lies”. Open this theme →
Ideas: Paradox of virality · Superspreader dynamics · Meme ecosystem competition · Neutral drift baseline · Replicator ontology · Bridge zone · Pseudo-context
Examples: COVID mask guidance · Nutrition science · Monetary policy in the financial press
9. How people actually believe things
People are harder to fool than the panic suggests, and the ways they do go wrong are mostly reasonable habits adding up badly. Open this theme →
Ideas: Epistemic vigilance · Open vigilance · Reflective vs. intuitive belief · Justification market · Credibility weighting · Conformity · False consensus among rational agents
10. Groups, splits and scale
As groups grow, their shared stories blur and they split along the blur, and simply reconnecting the pieces does not fix it. Open this theme →
Ideas: Myth dilution at scale · Network segmentation · Echo chambers vs. bubbles · Polarization via distrust · The Zollman effect · Inverted-U relationship
Open questions: Part IV: integration or fragmentation?
11. Who owns the filters
The filters that decide what you see have owners, and the owners are paid for your attention. Open this theme →
Ideas: Engagement-maximization as a captured equilibrium · The Orwell/Huxley axis · Cost-shifting from producers to consumers · Creation cheap / review expensive · Out-competition of institutional carriers in attention markets · Self-capture vs. external capture · Capture taxonomy · Consumer-key vs. surface capture · Capture asymmetry
12. AI in the chain
A language model puts every dial of the filter in one owner’s hands, which could finally help the finite reader or finish off the careful version. Open this theme →
Ideas: LLM as collapsed selection-design moment · Decompression on demand · Compressed form becoming authoritative · Corpus capture · Objective capture · Deployment capture · LLM as capability extender
13. Building bridges that last
The way through is people and institutions that can carry hard truths between groups that distrust each other, and survive long enough to matter. Open this theme →
Ideas: The curse of expertise · Versatile expertise · Preservation vs. training · Differential decay rates · Curation layer · Funding decoupling from attention markets · Survivable polarization · Survival, not victory · Work of generations
Examples: Common-law courts as a curation institution · Stack Overflow
Open questions: Transferable vs. specialized preconditions
Rather see it as a picture? Open the interactive map and pick “Idea map”.