Spokes.wiki

mathematics as a subject in its own right — the bodies of theory that sit as infrastructure beneath the hub's applied spokes and are the subject of none of them, taught from first principles in textbooks and course notes, together with the pedagogy that shapes them (level of rigour, exercise design, the computational tool assumed). Five families: analysis and calculus — vector & multivariable calculus, real and complex analysis, matrix calculus and calculus on general vector spaces, Fourier / transform theory and signal analysis, and the infinitesimal alternative to ε-δ; algebra — linear algebra, abstract algebra and group theory, and general (universal) algebra with the category theory it runs on; probability and statistics — probability theory and its combinatorial foundations, stochastic processes and stochastic calculus, and mathematical statistics and inference; discrete and combinatorial — graph theory; theories of structure applied to decisions and signals — game theory, information and coding theory, and measurement theory

149 pages · 31 tiered · 0 orphans · 0 stale

sourcing 98 · connection 100 · freshness 100 · depth 100

optimization algorithms — how a search for a best solution is actually carried out, and how competing methods are compared. Four corners: metaheuristic & population-based global optimizers (evolutionary, swarm, and nature/behavior-inspired, and their relatives) — the founding and deepest corner: how they explore search spaces, their mechanisms & parameters; classical, exact and convex optimization — convex programming, linear and integer/mixed-integer programming with the solvers and benchmark suites that rank them (HiGHS, MIPLIB, the Mittelmann benchmarks); gradient-based and stochastic optimizers — gradient descent, SGD, Adam, and the local/derivative-free methods beside them (Nelder–Mead, simulated annealing, Bayesian optimization); multi-objective optimization — Pareto-front methods and the indicators that score them (NSGA-II, hypervolume); plus the comparative benchmarking apparatus that cuts across all four — rankings, test functions, dimensionality, convergence, and the No Free Lunch result that limits what any ranking can mean

47 pages · 41 tiered · 0 orphans · 0 stale

sourcing 88 · connection 100 · freshness 100 · depth 98

AI governance, policy & regulation — how AI is governed, regulated, and strategically steered across organizations and nations: risk/compliance frameworks & standards (EU AI Act, ISO/IEC 42001, NIST AI RMF, OECD), national AI strategies & regulatory regimes, responsible/trustworthy AI, and AI assurance, auditing & certification

54 pages · 28 tiered · 0 orphans · 0 stale

sourcing 81 · connection 100 · freshness 100 · depth 100

transactional / operational (OLTP) databases and the tooling around them — row-oriented and distributed-SQL engines across the embedded↔distributed spectrum (SQLite-compatible embedded engines like Turso at one end; sharded / distributed SQL like Vitess/PlanetScale at the other), the sharding, replication, and consistency mechanisms that scale them, and the cross-cutting tools for working with databases. The operational sibling of analytical-databases-wiki — distinct from analytical / OLAP columnar stores

74 pages · 31 tiered · 0 orphans · 3 stale

sourcing 93 · connection 100 · freshness 81 · depth 100

how software and ML engineers learn and are assessed — the large self-directed curricula, meta-curricula and roadmaps that teach the disciplines (CS fundamentals, system design, AI/ML engineering), the interview-preparation canon they mostly serve, and the pedagogy they encode (build-from-scratch vs curated pointers, capstones, spaced study plans)

43 pages · 20 tiered · 0 orphans · 0 stale

sourcing 86 · connection 100 · freshness 100 · depth 90

the mechanics of LLM inference and serving — how a trained model turns a prompt into tokens and does it efficiently at scale: logits/softmax & sampling, prefill/decode, the KV cache, and continuous batching

55 pages · 31 tiered · 0 orphans · 0 stale

sourcing 80 · connection 100 · freshness 100 · depth 100

quantitative & algorithmic trading — how markets are traded systematically, and how such claims are checked. Four corners: the systems (founding and deepest) — the software and machine learning that trades markets: financial time-series / market-prediction foundation models, reinforcement-learning trading frameworks, and event-driven algorithmic-trading & backtesting engines, spanning the prediction, strategy and execution layers, plus the hyperparameter optimization that tunes them; derivatives pricing and options market structure — Black–Scholes and option pricing, the Greeks, dealer gamma exposure and gamma fragility, and what option volume reveals about the underlying; the statistics of evaluation and research integrity — backtest overfitting, the deflated Sharpe ratio, multiple-testing across thousands of alpha candidates, and the standing argument that much published backtest research does not survive its own correction; the industry itself — the firms, the exchanges and data vendors, and the interview canon that gates entry to them

80 pages · 25 tiered · 0 orphans · 0 stale

sourcing 79 · connection 100 · freshness 100 · depth 100

open-source intelligence (OSINT) — gathering and analyzing publicly available information for intelligence and reconnaissance, and especially the AI/LLM agents now automating it: autonomous and multi-agent OSINT tools, people/profile-building, username/IP/DNS reconnaissance, and the social-engineering reconnaissance and resources around the human element of security

53 pages · 30 tiered · 0 orphans · 2 stale

sourcing 82 · connection 100 · freshness 85 · depth 100

business & conversational messaging platforms and their developer APIs — the WhatsApp Business Platform as the founding, deepest corner (Cloud API, WhatsApp Business Accounts, message templates, webhooks, per-message pricing, and the BSP / Solution-Partner program), plus the sibling channels — RCS Business Messaging (Google/carriers) and Apple Messages for Business — and the client SDKs and intermediaries built over all of them: the messaging-transport layer beneath chat bots

29 pages · 14 tiered · 0 orphans · 0 stale

sourcing 95 · connection 100 · freshness 100 · depth 60

tools for building and running LLM agents — agent skills & the open skills standard, agentic coding harnesses, orchestration/multi-agent coordination, and the products/frameworks that package them (Claude Code, ADK, Gemini CLI, and the wider ecosystem)

355 pages · 218 tiered · 0 orphans · 23 stale

sourcing 81 · connection 100 · freshness 80 · depth 100

analytical / OLAP databases and the data-engineering stack around them — columnar query engines across the embedded↔distributed spectrum (DuckDB, ClickHouse), data warehouses, real-time analytics platforms, and the managed services that package them (Tinybird). How analytical data is stored, scaled, and queried — distinct from transactional (OLTP) databases

25 pages · 13 tiered · 0 orphans · 0 stale

sourcing 95 · connection 100 · freshness 100 · depth 52

the LLM model & provider landscape — the labs/providers and their models, API access & pricing, open-weight vs proprietary & licensing, and capability/cost/speed benchmarks (OpenAI, Anthropic, Google, DeepSeek, Meta/Llama, Mistral, Qwen, Moonshot, xAI, and the cloud resellers like Azure/Bedrock)

119 pages · 68 tiered · 0 orphans · 6 stale

sourcing 75 · connection 100 · freshness 88 · depth 100

the history of philosophy — major philosophers and their systems of thought across metaphysics, epistemology, ethics, and phenomenology, and the arguments and lineages connecting them (rationalism, the mind–body problem, substance, nihilism, phenomenology, Stoicism; the nature of the self, being, and world), in any tradition — Western and non-Western alike (Neo-Confucian li and

48 pages · 14 tiered · 0 orphans · 0 stale

sourcing 76 · connection 100 · freshness — · depth 100

psychology and the mind sciences — how minds judge, perceive, suffer, learn and create. Six corners: judgment & decision-making, the heuristics-and-biases tradition (Kahneman/Tversky, System 1/2, the focusing illusion); depth / analytical psychology and theories of creativity (Jung, the unconscious); perception and learning, studied empirically (psychophysics, perceptual world-models), including learning differences (dyslexia, dysgraphia) and the study of how people take in and retain material; clinical psychology and psychotherapy — trauma, grief, burnout, addiction and the named modalities that treat them (CBT, ACT, existential psychotherapy); behavioural neuroscience and the free-will question — what biology does to agency (Sapolsky); the discipline's own reliability — the replication crisis, famous results that did not survive re-examination, and findings that are routinely misread; plus psycholinguistics where language and thought meet (Sapir–Whorf). The empirical and clinical study of mind, distinct from philosophy of mind

131 pages · 50 tiered · 0 orphans · 0 stale

sourcing 69 · connection 100 · freshness 100 · depth 100

web performance, page-weight minimalism, and front-end byte budgets

30 pages · 18 tiered · 0 orphans · 1 stale

sourcing 98 · connection 100 · freshness 86 · depth 63

defending systems and software against attack — host & OS hardening against compliance benchmarks (CIS/STIG/NIST/PCI-DSS, OpenSCAP), security operations (SOC alert triage, detection engineering, threat hunting, incident response), and coordinated vulnerability response across the open-source supply chain (CVD, SIRT, disclosure governance) — the blue-team side of security, including the AI now being applied to all three

69 pages · 27 tiered · 0 orphans · 4 stale

sourcing 84 · connection 100 · freshness 67 · depth 100

developer tooling and programming infrastructure — the systems-level tools and libraries used to build, version, and ship software: compilers and language toolchains as build infrastructure (their implementation, performance and internals), version-control systems, serialization & data-interchange formats, and build/package tooling

35 pages · 21 tiered · 0 orphans · 1 stale

sourcing 87 · connection 100 · freshness 91 · depth 73

game engines and the real-time creation landscape, across 2D and 3D — their technology, tooling, community, licensing/business models, and the AI tooling reshaping development. General-purpose 3D engines are the founding/deepest corner (Godot, Unreal, Unity, O3DE); 2D and framework-based engines are in scope too (e.g. Flame on Flutter)

83 pages · 35 tiered · 0 orphans · 4 stale

sourcing 76 · connection 100 · freshness 83 · depth 100

production platform engineering, SRE & observability — running and operating cloud-native distributed systems in production (Kubernetes/CNCF platform ops, SRE / AIOps & incident response, observability & service topology), including the AI agents now being applied to operations

60 pages · 31 tiered · 0 orphans · 2 stale

sourcing 75 · connection 100 · freshness 82 · depth 100

static site generators and the Jamstack architecture — the tools, comparisons, and the shifting web-publishing / CMS landscape they are displacing (Astro, Eleventy, Hugo, Jekyll, Docusaurus; the WordPress decline)

32 pages · 23 tiered · 0 orphans · 0 stale

sourcing 85 · connection 100 · freshness 100 · depth 67

cryptography as a constructive discipline — the schemes themselves and what it costs to run them: succinct proof systems (SNARKs, polynomial commitments, sumcheck/IOP machinery) as the founding, deepest corner, alongside lattice-based constructions and program obfuscation, the finite fields and error-correcting codes they are built over, the hardness assumptions they rest on, and above all the concrete-efficiency frontier — how a construction travels from theoretically possible to actually runnable, and what it trades to get there

14 pages · 6 tiered · 0 orphans · 0 stale

sourcing 98 · connection 100 · freshness 100 · depth 29

free and open-source software that end users run themselves — the applications and environments people install on machines they control (Android apps, desktop window managers, Windows-compatibility runtimes) and the services they host themselves instead of renting (self-hosted media and business platforms), together with the directories that catalogue open-source substitutes for proprietary products

30 pages · 22 tiered · 0 orphans · 7 stale

sourcing 97 · connection 100 · freshness 67 · depth 63

music technology for electronic sound creation & production — synthesizers and sound-source instruments (hardware, DIY/open-source, and repurposed lab/test equipment), audio effect & processing plugins (delay, filter, dynamics/OTT, saturation), and the gear, software, and techniques of electronic music-making. The modular / patchable signal-routing lineage and the analog-vs-digital sound argument are the founding through-lines

129 pages · 47 tiered · 0 orphans · 3 stale

sourcing 66 · connection 100 · freshness 82 · depth 100

programming languages as subjects in their own right — their design and the paradigms behind it (memory and ownership models, type systems, concurrency models, immutability, code-as-data), the runtimes and hosts they target, the governance and evolution processes that change them, and the idiomatic program design their features make possible

70 pages · 52 tiered · 0 orphans · 14 stale

sourcing 88 · connection 100 · freshness 46 · depth 100

speech & audio AI models and the providers behind them — text-to-speech (TTS) / speech synthesis, speech-to-text (STT / ASR), and audio / music generation — their architectures (autoregressive / neural-codec / diffusion / state-space / flow), voice & audio quality, latency & error-rate benchmarks, licensing, and the open-weight-vs-proprietary market

87 pages · 52 tiered · 0 orphans · 6 stale

sourcing 71 · connection 100 · freshness 75 · depth 100

web browsers and the shifting browser market — the Chrome/Safari incumbents and the wave of alternatives challenging them, with the AI/agentic transformation of the browser as the founding, deepest corner (AI browsers, in-browser assistants, and agentic "act-on-your-behalf" browsing), alongside the privacy-first and productivity/niche browser movements

40 pages · 32 tiered · 0 orphans · 2 stale

sourcing 71 · connection 100 · freshness 92 · depth 83

machine learning as an engineering discipline — how models are built, trained, adapted, shipped, and understood. Five corners: classical ML algorithms and their internals (gradient descent, trees, boosting, clustering); model training & adaptation (fine-tuning, PEFT/LoRA/QLoRA, RL post-training) and the toolchains that run it; the design and productionization of ML systems (industry system-design case studies, the demo→production gap); training data and annotation as an engineering problem — label quality and label error, annotator agreement, who does the annotating, and the data cascades that follow when this is treated as someone else's job; the probabilistic and information-theoretic foundations the above rests on (Bayesian inference, information theory, Monte Carlo methods), together with interpretability and model science — what a trained network is actually doing inside (induction heads, phase transitions in attention, capability emergence), which is a different activity from building or shipping one

91 pages · 37 tiered · 0 orphans · 4 stale

sourcing 73 · connection 100 · freshness 69 · depth 100

search, SEO & performance marketing — how anything gets found, and who pays for it. Three corners: the AI transformation of search — AI overviews, generative and answer-engine optimization, agentic commerce, no-AI alternatives, and the market shifts behind them; traditional and performance search marketing — organic SEO operations, paid PPC, affiliate and lead generation: strategy, visibility and measurement; and the crawl economy — the payment and licensing layer now forming between publishers and AI crawlers: pay-per-crawl, machine-readable licensing standards, crawl-to-refer ratios, collective licensing bodies, and content monetization as an alternative to search traffic

91 pages · 54 tiered · 0 orphans · 2 stale

sourcing 66 · connection 100 · freshness 78 · depth 100

bit manipulation and low-level bitwise programming techniques — bit twiddling hacks, population count (popcount) / Hamming weight, branchless integer tricks (swap, abs, sign, min/max without branching), SWAR, and the canonical references that collect them (Sean Anderson's Bit Twiddling Hacks, Stephan Brumme's bits, Hacker's Delight)

21 pages · 13 tiered · 0 orphans · 0 stale

sourcing 83 · connection 100 · freshness 100 · depth 44

machines operating over file-based markdown to manage and mechanize thought — the knowledge-management / tools-for-thought lineage (memex, Zettelkasten, PKM) and formal methods / theorem proving — and the dynamics by which ideas & technologies diffuse and get adopted (diffusion of innovation, technology-adoption curves, trend analysis). (Agentic LLM products, skills & harnesses were split out to agentic-tooling-wiki on 2026-06-01; the two remain joined by bridge nodes — agent-skills, gbrain, compound-engineering, MCP.) Domain broadened 2026-06-09 (user directive) to add the diffusion/adoption-of-ideas cluster (F)

168 pages · 81 tiered · 0 orphans · 10 stale

sourcing 78 · connection 100 · freshness 50 · depth 100

loudspeaker & electroacoustic design — how moving-coil drivers and their enclosures are engineered: Thiele/Small parameters (Fs, Qts, Vas, Bl, Xmax…), driver electromechanics, sealed vs. vented/bass-reflex enclosure alignments, frequency/impedance/phase response, crossovers, the materials and measurements involved, and the calculators DIY and professional speaker builders use

32 pages · 16 tiered · 0 orphans · 0 stale

sourcing 66 · connection 100 · freshness 100 · depth 67

UI frameworks, component libraries & design systems — the software toolkits for building user interfaces across surfaces (terminal and web), their rendering/reconciler internals (native cores, React/Solid reconcilers, design tokens & theming), and the shift toward agent-buildable UI. Founding corpus: a terminal-UI core (OpenTUI) and a web design system (Astryx)

21 pages · 11 tiered · 0 orphans · 0 stale

sourcing 78 · connection 100 · freshness 100 · depth 44

cloud-hosting providers and infrastructure offerings

16 pages · 16 tiered · 0 orphans · 1 stale

sourcing 84 · connection 100 · freshness 90 · depth 33

life sciences — biology and biomedicine studied at the molecular and cellular level, and the AI now being applied to them: molecular neuroscience and genomics (single-cell epigenomics, protein-interaction maps, ageing tissue), disease biology and drug discovery, and the biological foundation models and bio-AI infrastructure being built to learn from that data. The organism and its data — distinct from psychology-wiki (mind and behaviour) and machine-learning-wiki (the model as a model)

15 pages · 7 tiered · 0 orphans · 0 stale

sourcing 90 · connection 100 · freshness — · depth 31

embedded systems and the Internet of Things — computing at the device and edge layer: microcontrollers and mixed-signal silicon, single-board computers and edge-compute boards, the firmware / RTOS stack that runs on them, the IoT connectivity protocols and platforms that connect them, and the open-hardware / DIY culture that builds with them

33 pages · 18 tiered · 0 orphans · 5 stale

sourcing 81 · connection 100 · freshness 50 · depth 69

application architecture for client-facing software — how a large frontend or full-stack application is divided into parts, who owns each part, and what decides the structure. Three registers: boundaries inside one codebase (feature-first layouts, layer separation, state ownership), composition across teams (micro-frontends, app shells, routing and communication between independently deployed units), and the frameworks and templates that make the decisions for you (full-stack meta-frameworks, declarative app compilers, SaaS boilerplates). The layer above the UI toolkit and below production operations

16 pages · 9 tiered · 0 orphans · 2 stale

sourcing 80 · connection 100 · freshness 60 · depth 33

formal knowledge representation and the semantic web — describing meaning so machines can reason over it: the RDF data model and its serializations, schema/ontology languages (RDFS, OWL, SHACL), query (SPARQL), shared vocabularies & structured data (schema.org), linked data and knowledge bases (Wikidata, DBpedia), and the tools for authoring ontologies

18 pages · 10 tiered · 0 orphans · 2 stale

sourcing 68 · connection 100 · freshness 33 · depth 38