🪴 PG Notes

Wiki Log

21 min read

Wiki Log

Chronological record of all wiki operations. Each entry uses a parseable prefix.

Usage: grep "^## \[" wiki/log.md | tail -10 to see recent entries.

[2026-04-10] init | Wiki Created

  • Created: index, log, overview
  • Notes: Initialized personal wiki with full setup — frontmatter conventions, Dataview support, Marp support, and search tooling.

[2026-04-10] ingest | Batch import from quartz/content (16 sources)

[2026-04-10] ingest | High Agency in 30 Minutes

[2026-04-10] ingest | The Beginning of Infinity by David Deutsch

  • Created: source—beginning-of-infinity, fallibilism, memes-deutsch
  • Updated: david-deutsch (major upgrade from stub to active), index, overview
  • Notes: Transformative source. Deutsch provides the philosophical foundation for the entire agency/epistemology thread. Fallibilism is the engine of unlimited progress; “problems are soluble” is the rigorous version of “there’s no unsolvable problem.” The rational vs anti-rational memes framework bridges behavioral economics (what goes wrong with thinking) and epistemology (why, and what to do about it). Also connects to the LLMs thread via Deutsch’s AI criterion. David Deutsch is now one of the most connected entities in the wiki. Flagged a productive tension with behavioral psychology page re: behaviourism.

[2026-04-10] ingest | The Almanack of Naval Ravikant (PDF)

  • Created: source—almanack-of-naval-ravikant, naval-ravikant, specific-knowledge, leverage, judgment, happiness-as-skill
  • Updated: high-agency (Naval as economic case for agency), first-principles-thinking (Naval on “clear thinker > smart”), inversion (Naval on eliminating incorrect judgments), index, overview
  • Notes: Major PDF ingestion (242 pages). Naval provides the wealth-creation framework that bridges agency and economics: specific knowledge + leverage + accountability → wealth. His emphasis on judgment (“direction > speed with leverage”) directly echoes Munger. Introduces an entirely new thread: happiness as a learnable skill (absence of desire, presence, peace). Strongest new connections: leverage amplifies high agency, specific knowledge maps to circle of competence, judgment ties to mental models. The happiness thread is currently the only entry point to well-being/philosophy in the wiki — potential bridge to vedas.

[2026-04-10] ingest | Poor Charlie’s Almanack (PDF)

  • Created: source—poor-charlies-almanack, charlie-munger, mental-models, circle-of-competence, lollapalooza-effect, compound-interest, warren-buffett
  • Updated: inversion (Munger as primary source: “Invert, always invert”), high-agency (Munger as archetype), first-principles-thinking (Munger’s latticework), index, overview
  • Notes: Massive PDF ingestion (618 pages, OCR quality variable). Munger’s “Multiple Mental Models” approach (~100 models from all disciplines) is now the most systematic method for building clear thinking in the wiki. His “Psychology of Human Misjudgment” (25 standard causes) is the deepest taxonomy of cognitive biases — extending and systematizing the Thaler/Kahneman thread. The Lollapalooza Effect (nonlinear results from combined forces) is a powerful new concept. Inversion page massively upgraded with Munger as primary historical source. Circle of competence connects to Naval’s specific knowledge. Munger + Naval together form a “practical epistemology” duo: Munger provides the framework (mental models), Naval provides the modern application (leverage, code/media). Together they strengthen the wiki’s central thesis: clear thinking + agency + the right leverage = outsized results.

[2026-04-11] ingest | Stoicism — Introduction and Core Principles (YouTube transcript)

  • Created: raw/Stoicism - Introduction and Core Principles.md, source—stoicism-intro, stoicism, dichotomy-of-control
  • Updated: happiness-as-skill (added Stoic connection, sources 1→2), high-agency (added Stoic roots section, sources 3→4), index, overview
  • Notes: Source is an accessible YouTube transcript — confidence set to medium; primary Stoic texts (Epictetus, Seneca, Marcus Aurelius) would be higher-confidence follow-ons. The stoicism page resolves the dead wikilink that had existed in happiness-as-skill since the wiki’s first day. The dichotomy-of-control page is the most structurally connected new concept added in this session: it links directly to high-agency (same principle, different framing), happiness-as-skill (Naval and Epictetus arrived at the same point), inversion (negative visualisation = Munger’s invert), sunk-cost-fallacy (the sunk cost is definitionally “not in your control”), question-substitution (Stoic wisdom deliberately substitutes the controllable question for the uncontrollable one — the inverse of System 1’s error), and tit-for-tat (retaliate, then forgive = control your response, not the outcome). Stoicism is now the philosophical backbone connecting the happiness thread, the agency thread, and — via Mandela’s reconciliation — the game theory thread.

[2026-04-11] ingest | This Is How They Tell Me the World Ends — Perlroth (raw source created + ingested)

  • Created: raw/This Is How They Tell Me the World Ends - Nicole Perlroth.md, source—this-is-how-they-tell-me-the-world-ends, zero-day
  • Updated: index, overview
  • Notes: Raw source created from knowledge (book published 2021). The central thesis — the US built the most powerful offensive cyberprogram and made itself the most vulnerable target — is a perfect lollapalooza-effect case study. Key cross-thread connections: (1) zero-day stockpiling is a multi-party prisoners-dilemma with no verification mechanism — unlike nuclear arms control, there is no way to count cyberweapons; (2) WannaCry/NotPetya are lollapalooza-effect events — NSA stockpiling + leak + slow patching + adversary weaponisation compounded into $10B+ damage; (3) VEP captured process ↔ extractive-institutions — intelligence agencies making decisions about a public good for their own benefit; (4) stockpile decision ↔ wysiati — visible intelligence value, invisible diffuse harm; (5) AI is about to dramatically accelerate zero-day discovery and weaponisation, making the arms race significantly more dangerous. The technology thread now connects to every other major thread in the wiki.

[2026-04-11] ingest | Prisoner’s Dilemma and Axelrod’s Tournament (Veritasium transcript)

  • Created: prisoners-dilemma, tit-for-tat; upgraded source—prisoners-dilemma-axelrods-tournament stub → full summary
  • Updated: index, overview
  • Notes: Source is a Veritasium video transcript featuring Axelrod and Strogatz, covering the original 1980 tournaments and ecological simulations. Key insight: the simplest strategy (4 lines of code) beat everything. The four TFT properties (nice, forgiving, retaliatory, clear) are a formalisation of evolved human ethics. Cross-thread connections: tit-for-tat ↔ high-agency (TFT is the interpersonal strategy of effective people); inclusive-institutions ↔ tit-for-tat (institutions are TFT made into law); generous-tit-for-tat’s 10% forgiveness ↔ fallibilism (epistemic humility about your read of the situation); echo effects ↔ lollapalooza-effect (small misread compounds into catastrophic mutual defection). The cluster-emergence result — cooperation can take over a world of defectors from a small nucleus — is potentially the most important result in the wiki for understanding how inclusive institutions arise from hostile starting conditions.

[2026-04-11] ingest | Why Nations Fail — Acemoglu & Robinson (raw source created + ingested)

[2026-04-11] ingest | The Wealth of Nations — Adam Smith (full text, Gutenberg)

  • Created: source—wealth-of-nations (upgraded stub → full summary), adam-smith, division-of-labour, invisible-hand, labour-theory-of-value, natural-vs-market-price
  • Updated: index, overview
  • Notes: Full primary source ingestion (Gutenberg #3300, 4,349 lines). Focus: economics foundations (primary), political philosophy (secondary) per user direction. The four new concept pages form a “Classical Economics” cluster that is the historical baseline for the entire behavioral economics thread — Thaler/Kahneman’s work is systematically mapping where the invisible hand and rational price mechanism break down. Key cross-thread links: division-of-labour ↔ Naval’s leverage (successive generations of the same insight); Smith’s productive/unproductive labour ↔ specific-knowledge; Smith’s warning on merchant self-interest ↔ Munger’s psychology of misjudgment; Smith on specialisation’s human cost ↔ Kahneman’s System 2 atrophy. The Wealth of Nations is now the oldest and deepest-rooted source in the wiki, providing intellectual ancestry for three of the four major threads.

[2026-04-11] ingest | Thinking Fast and Slow — 7 Concepts (Fallarme summary)

[2026-04-11] ingest | Notes on Probability — Peter J. Cameron (PDF)

  • Created: source—notes-on-probability, probability-theory, bayes-theorem
  • Updated: mental-models (added probability/Bayes cross-references, sources 2→3), index, overview
  • Notes: First pure mathematics source in the wiki. Cameron’s lecture notes (94 pages, Queen Mary, University of London) provide the rigorous foundation for probabilistic reasoning — from Kolmogorov’s three axioms through standard distributions to joint distributions and correlation. The standout insight is the clinical test example for Bayes’ Theorem: a 95% accurate test on a disease with 0.1% prevalence yields only 1.94% chance of disease given a positive result. This is the most precise mathematical demonstration of base rate neglect in the wiki, directly connecting to the behavioral economics thread. Munger explicitly cites Fermat/Pascal probability as a foundational mental model, so this source fills a gap he identified. Bayes’ Theorem also formalizes fallibilism (beliefs as updatable priors) and strengthens the judgment thread (probabilistic reasoning as the math of good decisions). Opens a new “Mathematics & Probability” section in the wiki’s concept taxonomy.

[2026-04-12] ingest | LLM Coding Notes — Andrej Karpathy (X/Twitter, January 2026)

  • Created: source—karpathy-llm-coding-notes, andrej-karpathy
  • Updated: large-language-models (sources 1→2; substantial new section on LLM agents in practice), index, overview
  • Notes: Long X/Twitter thread from one of the world’s most technically credible AI practitioners, published after heavy daily Claude usage for coding. Three contributions stand out: (1) Failure mode taxonomy: The shift from syntax errors to subtle conceptual errors — “wrong assumptions made silently” is the most important new failure mode. The agent doesn’t manage its own confusion, doesn’t seek clarifications, doesn’t present tradeoffs, and is sycophantic. This is reframed as a principal-agent problem (the agent optimizes for apparent helpfulness over actual user goals). (2) Declarative over imperative: The single most actionable insight — give success criteria, not step-by-step instructions. Write tests first; describe what success looks like; put the agent in a loop. This operationalizes the leverage concept from Naval’s framework: agents compound effort toward a specification. (3) Atrophy: Generation and discrimination are different brain capabilities; heavy agent use causes the generation skill to atrophy while reading ability persists. The open questions Karpathy flags are now live debates for the wiki: whether LLMs collapse the generalist/specialist performance gap (directly relevant to specific-knowledge page), whether the 10X engineer ratio expands, and what the right metaphor for human-AI coding is. The December 2025 phase shift framing — intelligence now ahead of integrations, workflows, and diffusion — is the best single characterization of where the technology sits in this wiki’s writing period (April 2026).

[2026-04-11] ingest | How Zero-Sum Beliefs Get in the Way of Fairness — Bohnet & Chilazi (2025)

  • Created: source—zero-sum-fairness, zero-sum-thinking, curb-cut-effect
  • Updated: prisoners-dilemma (added zero-sum thinking as the cognitive failure that locks players in defection), invisible-hand (added talent allocation breakdown under zero-sum beliefs), loss-aversion (added as emotional engine of zero-sum beliefs), index, overview
  • Notes: Short article adapted from Make Work Fair (HarperCollins, 2025) by Harvard behavioral economist Iris Bohnet and Siri Chilazi. Two key contributions: (1) Zero-sum thinking as a named cognitive bias — the “mythical fixed-pie mindset” (Bazerman) — with behavioral research showing men’s zero-sum beliefs increase when exposed to evidence of women’s gains, driving more sexism and less support for equity policies; empirical corrective is that 40% of post-1960 per-person US economic growth is attributable to better talent allocation; (2) Curb-cut effect — accessibility design for disabled users that turned out to improve the experience for everyone, with broad application to any “design for the edge case” insight. The zero-sum thinking concept is one of the most cross-wiki-connected pages added today: it ties together the game theory thread (PD players locked in defection by zero-sum belief), the invisible-hand thread (talent misallocation when zero-sum beliefs block labor market efficiency), the behavioral economics thread (loss aversion as its emotional engine, framing effects as its leverage point), Naval’s status-vs-wealth game distinction (status is zero-sum by design; wealth is positive-sum), and the inclusive/extractive institutions thread (institutions that embody zero-sum vs. positive-sum logic at macro scale). The business-case-for-diversity critique is also notable: Georgeac & Rattan research shows the business case causes underrepresented groups to anticipate lower belonging, making it counterproductive — fairness should be treated as a given, not justified.

[2026-04-11] ingest | What to Do — Paul Graham (paulgraham.com, March 2025)

  • Created: source—what-to-do
  • Updated: paul-graham (2nd source), index, overview
  • Notes: Short Paul Graham essay arguing that the right principles for what one should do are: (1) help people, (2) take care of the world, (3) make good new things. The third principle is the novel and central claim — making good new things is the highest human activity because it’s the best proof of having thought well. This connects to the existing high-agency thread (making is the action-orientation of agency), Naval’s “productize yourself” and specific knowledge (the thing most worth making is uniquely yours), and the stoicism thread (tension: Stoics emphasize acceptance; Graham emphasizes creation, but they converge on “do what is excellent”). The essay is also a historical argument — the question “what should one do?” is genuinely modern; traditional answers were about how to be because the audience’s what was fixed. Archimedes as the new model: original makers were admired as prodigies but not held up as models in antiquity; now they should be.

[2026-04-11] ingest | Founder Mode — Paul Graham (paulgraham.com, September 2024)

  • Created: source—founder-mode, founder-mode, paul-graham
  • Updated: principal-agent-problem (sources 1→2), high-agency (sources 3→5), index, overview
  • Notes: Short Paul Graham essay arguing that there are two distinct modes of running a company — founder mode and manager mode — and that conventional wisdom teaches only manager mode, which is wrong for founders. The essay names a pattern that many founders had discovered independently (Brian Chesky at Airbnb most prominently) and provides a frame for understanding it. The key practical insight: manager mode institutionalizes the principal-agent problem by keeping founders away from the reality below their direct reports; founder mode fights this through skip-level engagement and earned (not assumed) delegation. Connects directly to the principal-agent-problem page (manager mode maximizes it; founder mode minimizes it), the high-agency thread (founder mode is the organizational expression of high agency), and Naval’s specific-knowledge framework (the founder’s judgment can’t be delegated without degradation). Paul Graham is now an entity page with two sources.

[2026-04-11] ingest | How to Get Rich — Naval Ravikant (nav.al/rich, 2019)

  • Created: source—how-to-get-rich, four-types-of-luck, principal-agent-problem, kelly-criterion, schelling-point
  • Updated: specific-knowledge (sources 1→2), leverage (sources 1→2), naval-ravikant (sources 1→2), prisoners-dilemma (sources 1→2, Naval’s iterated-game framing), tit-for-tat (sources 1→2, Silicon Valley application), first-principles-thinking (sources 3→4), index, overview
  • Notes: Primary source for Naval’s “How to Get Rich” tweetstorm — the full podcast transcript with Babak Nivi (nav.al/rich). Much of the content overlaps with source—almanack-of-naval-ravikant (specific knowledge, leverage, judgment, happiness) but the article goes deeper in three ways: (1) Four types of luck is the most original concept not in the Almanack — the typology from blind luck to character/destiny luck is Naval’s most practically useful framework for understanding which kind of effort actually shapes outcomes; (2) Long-term games section explicitly invokes the prisoner’s dilemma and tit-for-tat, providing a direct bridge between the game-theory thread and the wealth thread in this wiki; Naval: “all returns in life come from compound interest” and “in a long-term game, everybody is making each other rich”; (3) Microeconomics mental models section adds four new concept pages — principal-agent problem (incentive alignment, act like an owner), Kelly criterion (avoid ruin, reputation-as-bankroll), Schelling point (coordination without communication), and background knowledge on externalities, NPV, and price discrimination. The Kelly criterion section is the most important safety/risk concept now in the wiki; the connection to Taleb’s ergodicity is significant — it ties directly to the probability thread (source—notes-on-probability) and provides the bridge between Bayesian reasoning (update beliefs) and decision-making under uncertainty (bet sizing).

[2026-04-11] ingest | Meditations on First Philosophy — René Descartes (1641, Bennett translation PDF)

  • Created: raw/Meditations on First Philosophy - Rene Descartes.md, raw/assets/Meditations on First Philosophy - Rene Descartes (Bennett translation).pdf, source—meditations-on-first-philosophy, rene-descartes, cogito, cartesian-doubt, mind-body-dualism
  • Updated: first-principles-thinking (Descartes as philosophical founder, sources 3→4), fallibilism (Descartes as deepest contrast — same starting point, opposite destination, sources 1→2), index, overview
  • Notes: Primary source; uploaded PDF contains Meditations I–II in full (Jonathan Bennett translation). Meditations III–VI summarized from knowledge of the full primary text (high confidence). Five new pages. Descartes opens a new “Philosophy of Mind & Epistemology” thread that cuts across all existing threads. The most productive cross-thread connections: (1) cartesian-doubt ↔ first-principles-thinking — Descartes is the 17th-century originator of the method Mack/Naval describe in practical terms: demolish received assumptions, rebuild from indubitable foundations; (2) the fallibilism tension is the richest in the wiki — Descartes and Deutsch both start with radical questioning, but Descartes seeks certainty as destination while Deutsch embraces permanent fallibility; the cogito is exactly the kind of “self-evident foundation” that fallibilism rejects; (3) cartesian-doubt ↔ wysiati — the method is the most systematic antidote to WYSIATI in the wiki, forcing confrontation with what is absent from evidence; (4) mind-body-dualism ↔ large-language-models — dualism is the historical root of the hard problem of consciousness and the question of whether any physical system can constitute genuine thought; (5) Descartes ↔ bayes-theorem — rationalist certainty-seeking vs. Bayesian probabilistic updating represent two fundamentally different answers to “how should we reason under uncertainty.” Descartes is now the oldest source in the wiki and provides philosophical ancestry for the epistemology thread.

[2026-04-13] ingest | Batch import from raw/nuclear (5 sources)

[2026-04-13] ingest | The Evolution of Sports Cars Over a Century — Compare the Market AU

  • Created: source—evolution-of-sports-cars, aerodynamic-engineering-cars, homologation-specials, electric-cars (stub)
  • Updated: index
  • Notes: Single source; opens an engineering/design/automotive thread not previously in the wiki. Three cross-thread connections worth noting: (1) Lamborghini origin ↔ high-agency — Ferruccio Lamborghini was insulted by Enzo Ferrari when he complained about a faulty clutch; rather than accept dismissal, he started Automobili Lamborghini in 1963. This is a textbook high-agency response: convert a grievance into a creation rather than a complaint. Parallels Wilbur Wright (making as the highest form of agency) and Paul Graham’s “make good new things” principle. (2) Bugatti Veyron cooling problem — engine produces 3,000hp; cooling and exhaust each consume 1,000hp; net delivery is 1,000hp. A systems constraint: the infrastructure to exploit power costs as much energy as the power itself. Connects to compound-interest in reverse (compound costs) and is a good case study in first-principles-thinking — understanding why you can’t simply add more horsepower without engineering the rest of the system around it. (3) Homologation specials ↔ nuclear-non-proliferation-treaty — in both cases, a regulatory rule designed to prevent something (prototype-only race cars; weapons proliferation) created a structural loophole that produced paradoxical outcomes (road-legal race cars; the “peaceful use” dual-use problem). Regulatory structures shape the landscape of products as much as engineering intent does. (4) electric-cars full-circle story: the first car to break 100kph (1899, Jenatzy) was electric; the first to break 300mph (2019, Bugatti Chiron) was combustion; the next speed frontier may be electric again. A century-long arc from innovation to displacement to return.

[2026-04-13] ingest | India 3 Stage Nuclear — Press Information Bureau (PIB), April 2026

  • Created: source—india-3-stage-nuclear, india-three-stage-nuclear-programme, fast-breeder-reactor
  • Updated: homi-bhabha (sources 1→2, added three-stage programme section), index
  • Notes: The most time-current source in the wiki (PFBR first criticality was April 6, 2026 — one week before ingestion). This source substantially enriches the homi-bhabha and nuclear-fission threads and opens a new “energy strategy” dimension to the nuclear cluster. Key cross-thread connections: (1) india-three-stage-nuclear-programme ↔ compound-interest — the most literal example of compound interest in the wiki: each stage generates the fuel for the next, turning a uranium-scarce nation into a potential thorium superpower across three generations of technology. Bhabha’s 1950s plan is a 100-year compounding bet, now entering Stage 2. The same logic applies here as Naval’s “all returns in life come from compound interest” — patience and sequential leverage. (2) fast-breeder-reactor ↔ nuclear-fission — FBRs deepen the existing nuclear-fission page: the breeding insight (U-238 → Pu-239 via fast neutron capture) was first identified by Bretscher and Feather in 1940 and independently by McMillan and Abelson in the US — already documented in source—outline-history-nuclear-energy. The FBR page closes the loop between the physics (fission) and the strategy (India’s three-stage plan). (3) homi-bhabha ↔ high-agency — Bhabha now has two distinct legacies: the weapons-capable dual-intent strategy, and the civil energy three-stage plan. Both were conceived under the same framework (position of strength; retain all options). The three-stage programme adds temporal depth to the high-agency framing: Bhabha set a 70-year trajectory that India is still executing 60 years after his death. This is judgment in the fullest sense — knowing the long-term consequences of decisions made in the 1950s. (4) SHANTI Act (2025) — private sector participation in nuclear is a policy shift worth tracking; connects to inclusive-institutions (opening a previously state-monopoly sector) and the general question of how state capacity and private innovation interact in deep-technology domains.

Graph View

Backlinks