KNOWLEDGE NOTESKnowledge Notes
Explore curated knowledge, course notes and methods, with their existing content and sources preserved.
This index lists complete English editions. Additional Chinese notes and bilingual updates remain available in the Chinese index.
350knowledge notesAll AI techniques for the Deliver phase of Will's "AI-Assisted UX Workflow". The real goal of this phase is not to hand off specs to developers, but to distill the project into a presentable portfolio page plus interview talking points. Therefore, all AI techniques center on two things: "rearranging information into a story-telling visual/layout" and "clearly explaining the human-AI collaboration process". Coverage includes: ① Generating multiple layout options (1:1 text extraction from old layout → content direction exploration → direction selection → final image output) ② Optimizing layouts / migrating visual styles (style transfer from reference images, only changing the skin without altering the bones) ③ Layout-to-code & Figma mockup (single-page image → restoration prompt → HTML → reverse-engineer editable source file via Figma Make) ④ Infographic semantic extraction + visual redraw (strip visuals to retain only logic → 5 styling directions → wireframe → apply style) ⑤ Metric-driven design review using the will-ux-metric-teardown skill (reverse-engineer business model/North Star metric, identify issues by Tier 1-3 priority) ⑥ Showcasing AI capabilities and value in projects (for each UX phase, output "what AI did / interview talking points / what evidence to put in PPT"). Each section provides reusable prompt skeletons + Why reasoning + counterexamples. Interpersonal collaboration rules for design review meetings (roles / receiving feedback / turning feedback into actions) are covered in [[google-ux-testing-iteration]]; this article focuses specifically on prompts and workflows for using AI to run layout generation/code conversion/metric review/value articulation.
This is the prompt and workflow handbook for the first phase (Discover & Research) of Will’s "AI-Assisted UX Workflow". It covers 7 tasks in the discovery and research phase that can be accelerated by AI: ① Distill scattered project background into a professional "Project Background (PB) statement" ② Derive business goals × experience goals × plain-language driving logic from the background ③ Locate experience problems on old interfaces and quantify their metric impact ④ Rewrite vague problems into actionable problem statements + problem overview diagrams ⑤ Use MVP hypothesis/metric scripts for interview verbal delivery ⑥ Generate a fully structured product concept user research report (including simulated data + risk and objection signals) ⑦ Convert content into "chart prompts" for professional infographics. Each section provides reusable prompt skeletons + Why explanations + counterexamples + sample outputs. It also includes the underlying mental models (business outcome = traffic × conversion × cost, experience only acts on the behavioral layer, MVP dual hypotheses, four-dimensional demand scoring). It complements google-ux-research: that piece covers UX research methods themselves (how to conduct interviews/draw empathy maps/run usability tests), while this piece explains how to use AI to quickly turn these research outputs into deliverables suitable for reporting or inclusion in a portfolio. It is oriented toward non-technical PMs, with jargon explained in parentheses.
Will’s prompt manual and methodology for the Ideate phase of the "AI-Assisted UX Workflow". It covers five categories of ideation work: ① Three types of project decision report templates (multi-version optimization / MVP / exploratory innovation) — reverse-engineer an already shipped design into a complete decision chain of "attribution → direction → candidate solution comparison → decision table → PPT → interview talking points", whose core value is "proving you are making judgments rather than executing"; ② AI competitive research and analysis report — the "variable validation method" that breaks down variables, runs frequency statistics, and draws probability-level conclusions, to make near-objective judgments when you cannot access competitors’ real data; ③ Generate innovative products from a one-sentence requirement (Peter Thiel-style PRD prompt); ④ A 5-step chain for optimizing solutions through usability testing (test report → data table → dashboard → revision prompt → new interface); ⑤ Use AI + PRD to generate user flow diagrams (4-node syntax) and wireframes (multi-structure divergence). Each section provides reusable prompt skeletons, Why explanations, counterexamples, and interview talking point formulas. Division of labor with [[google-ux-wireframe-prototype]]: that article covers "hand-drawing methods and symbol standards for user flows/wireframes/IA" (pure manual design language), while this article covers the prompt workflow of "using AI to turn these tasks into reports and decision documents" (the AI acceleration layer), as well as the unique cognition of "design strategy" and "decision-making ability expression" in the ideation phase.
Will’s complete pipeline for non-technical PMs to go from “idea to runnable prototype”, breaking vibe coding into 7 copyable steps: ① Use 3 tiers of templates to turn a one-sentence product idea into structured input; ② Have AI act as a “requirements interrogator” to ask follow-up questions in sequence and force out a Grill Summary; ③ Use a “top-tier product architect” role prompt to expand the 5 answers into an investment-grade PRD; ④ Extract feature points from the PRD to generate 3 non-overlapping interaction DRD options, then convert the chosen option into a coding prompt that can directly produce a product; ⑤ Feed it to Figma Make (to produce an interactive demo) or Claude (to produce a real launch-ready product) to generate a prototype; ⑥ Feed screenshots back in to let AI spot errors and iterate on its own; ⑦ Publish to mobile (Claude outputs HTML → Netlify Drop drag-and-drop hosting → add to home screen as an App) + one-click reskinning with ReferoStyle. Each step provides reusable prompt skeletons + Why + examples + pitfalls. Three parallel methods (Method 1 Document-driven mainline / Method 2 Document-driven HTML output / Method 3 GPT+Figma natural language UI building). The audience is non-technical product managers who care about “when to use which step, how to copy the prompts” rather than code. Division of labor with [[newtype-ai-coding]]: that article covers tool stack selection and mindset (how to choose Cursor/Claude Code/Codex), while this article covers the concrete action chain and prompt templates for non-technical PMs to run through a prototype from scratch.
Design highlights from 11 quarterly product Releases of Airbnb from Summer 2021 to Summer 2026, breaking down "why Airbnb designs this way" from a product manager perspective. Two main threads: ① Timeline — 2021 Flexible search + 10-step host onboarding / 2022 Categories system (50+ themes) + Split Stays + free AirCover protection / 2023 Host Passport + total price transparency + Guest Favorite tag + AI photo tours + smart locks / 2024 Icons celebrity immersive experiences + shared wishlists + Co-host marketplace / 2025 Services value-added services + Experiences local experiences + Connections social relationships + task-based AI customer service / 2026 lifestyle services (groceries/transfers/luggage storage/car rentals) + boutique hotels + AI review summaries and comparisons + travel maps. ② Transferable insights — how Airbnb does "dual-track narrative Releases" (every feature communicates both UX value and business value at the same time), how trust is designed into the product (AirCover/identity verification/Guest Favorite/Host Passport), how belonging evolves from "staying in someone else's home" to emotional connection (passport/social/shared itineraries), how the platform strategy evolves from a "booking tool" to an "end-to-end travel and lifestyle platform", and the three overarching threads of flexibility, personalization, and AI. Audience = non-technical PMs; jargon is explained within the industry; each section includes Why and counterexamples; instead of restating feature lists, it distills design motivations.
Founders often ask whether a model upgrade will erase their feature. Albert moves the question earlier: can you deliver defined value to a defined audience now, and identify which change improved the result? If product quality remains difficult to judge, discussing permanent defenses can lead a team into waiting without feedback.
A business trip can appear to require only an itinerary. Across cities, days and meetings, however, hotels, transport and timing constrain one another. An early choice can make the eleventh step impossible. Bill uses this example to explain why an agent needs to consider consequences, search alternatives and recover when a path fails.
How can a team outside its target market know whether it is building something people need? Zhang Yang begins by asking whether the team obtains real, firsthand information continuously and connects that understanding to product, payment and acquisition.
Themed condensation of 13 official 2026 blog posts. The core is Claude Managed Agents — a set of APIs for "renting Anthropic's cloud infrastructure to run long-horizon agents": memory (cross-session memory) + dreaming (sleep-time self-improvement) + outcomes (scored self-correction) + multiagent (lead-worker collaboration) + self-hosted sandboxes (bring-your-own sandbox) + MCP tunnels (private network tunnels), underpinned by a decoupled meta-harness design of three components: "brain (Claude + harness) / hands (sandbox) / session logs". Additional content covers the Claude Code desktop app rebuild (parallel agents) + harness design philosophy (use tools Claude already knows / keep asking "what can I turn off") + enterprise deployment data + containment security architecture for three products + a security checklist for AI-accelerated offense and defense + chart visualization / frontend Skills / lifestyle connectors.
Hub of 11 entries from the Anthropic research section: Constitutional AI (2022 foundational) + Automated Alignment Researchers (2026 scalable oversight) + Petri open source (donating-open-source-petri) + deprecation commitments (models not deleted) + end-subset-conversations (Claude welfare) + Teaching Claude Why (follow-up to agentic misalignment) + Project Vend Phase 2 (failed Claudius red team experiment) + 2028 AI Leadership (US-China) + Anthropic Institute 4 major research agendas + Core Views on AI Safety + full ASL3 overview. The 5 pillars of Anthropic's safety stance: Constitutional + RSP/ASL / Transparency / Welfare / Red Team
Anthropic's "AI Economic Impact Measurement" research series. From its launch in early 2025 to May 2026, it has released 5 major reports plus 1 monthly survey project. Coverage: **the real distribution of what people use Claude for** (coding is not the only use case) / **uneven geographic and enterprise distribution** (40% of US workers use AI; China, India, Brazil, Africa) / **economic primitives** (new metrics: scope / breadth / depth) / **81K user survey** (occupational exposure vs concern + largest productivity gains at both the high- and low-wage ends) / **AI's impact on programming** (the industry chain has already changed) / **economic futures UK+Europe** (policy project). This wiki compresses the 5 reports + 3 news items into 1 map of "AI economic impact timeline + 5-dimension framework + Anthropic measurement methodology".
Anthropic enterprise product lines + strategic partnerships + geographic expansion: a hub of 9 pieces. **Vertical industry products**: Claude for Financial Services + finance-agents (10 ready-to-run templates) / Claude for Creative Work / Claude for Small Business / Claude Design (Anthropic Labs). **Strategic partnerships**: PwC expansion / Gates Foundation $200M / new company with Blackstone+H&F+Goldman / SpaceX compute / Stainless acquisition (SDK + MCP server tools). **Geographic expansion**: Australia + NZ GM appointment + Sydney office. **Model**: Anthropic is evolving from a pure model company into a four-dimensional parallel play of "industry verticals + tool acquisitions + strategic partnerships + geographic coverage".
5 hub articles from the Anthropic research section: Natural Language Autoencoders (decoding Claude's internal activations into natural language, a major breakthrough in May 2026) + How people ask Claude for personal guidance (sampling of 1 million conversations / 6% are seeking life advice) + BioMysteryBench (bioinformatics evaluation) + Why do AI models hallucinate? + What is sycophancy? + 4 research team pages (Alignment / Economic Research / Interpretability / Societal Impacts). Covers three areas of research: "how AI thinks internally + how users use it + what bugs it is prone to"
The entry wiki for the Chinese-English bilingual notes on the Apple Human Interface Guidelines (HIG). It explains the 4 categories of tags in Apple design (philosophy/concept/principle/strategy) + the three-dimensional methodology of atomic design, multi-platform, and hierarchy model + a quick reference for device characteristics and design tradeoffs across the 7 major platforms (iOS/iPadOS/macOS/watchOS/tvOS/visionOS/games) + a reverse lookup map for all 14 apple-hig-* wikis (which wiki to check when you want to look up X). Start here to locate information for any Apple design question.
The 'make way for everyone' layer of Apple design. 6 dimensions of accessibility (Vision / Hearing / Mobility / Speech / Cognition / visionOS) + 9 core accessibility principles (Contrast / Dark Mode / Larger Text / Don’t Rely on Color Alone / Reduce Motion / Voice Control / VoiceOver / Captions / Audio Descriptions) + App Store Accessibility Nutrition Labels evaluation method + Mac VoiceOver acceleration (contain / combine / ignore + rotor + actions) + Inclusive design (diversity factors / inclusive language / avoiding stereotypes) + Privacy (ask for permissions only when needed / purpose strings / location button / data protection) + Writing (tone / action orientation / error messages / empty states) + Right-to-Left RTL (text alignment / numbers do not flip / controls flip / icon localization). Each entry includes real-world test cases from the Landmarks App.
A decision manual for "when to use which" covering all 63 UI components in Apple HIG, organized by 8 functional categories. Menus and Actions (Buttons/Menus/Context Menus/Toolbars/Pop-up Buttons/Pull-down Buttons/Activity Views/Ornaments) · Selection and Input (Text Fields/Pickers/Segmented Controls/Sliders/Steppers/Toggles/Color Wells/Image Wells/Combo Boxes/Virtual Keyboards) · Presentation (Alerts/Action Sheets/Sheets/Popovers/Panels/Windows/Scroll Views/Page Controls) · Layout and Organization (Lists and Tables/Collections/Outline Views/Split Views/Tab Views/Sidebars/Boxes/Labels/Column Views/Lockups/Disclosure Controls) · Navigation and Search (Tab Bars/Sidebars/Search Fields/Path Controls/Token Fields) · Content Containers (Charts/Images/Text/Web Views) · Status Indicators (Progress Views/Activity Rings/Gauges/Ratings) · System Experiences (Widgets/Live Activities/Notifications/Controls/Complications/Shortcuts/Menu Bars/Status Bars/Top Shelf/Watch Faces). Each component includes a definition + when to use it + key do/don't guidance + cross-platform support. Includes anatomy diagrams for Primary/Secondary buttons.
The most 'methodology-focused' layer of Apple's design system — it does not talk about which component to use, but about why. Design philosophy (why Human Interface is called Human · design serves people · 6 qualities of great design) + 12 core design principles (Wayfinding / Feedback / Visibility / Consistency / Mental Model / Proximity / Grouping / Mapping / Affordance / Progressive Disclosure / 80-20 / Symmetry) + 5 elements of Intentional Design (radical simplification / deep understanding / extreme focus / personal connection / direct communication) + Fluid Interfaces (dynamic motion vs timed animation · elastic physics · projected momentum · 6 sub-principles) + complete workflow from concept to interface (structure → navigation → content → visuals, with a vinyl record App hands-on example) + multisensory design (sound design · 3 haptic principles · transient/continuous · technical constraints) + multilingual, intelligence and ML experiences (data/metrics · system intelligence journey) + 60-minute rapid prototyping.
Design guidelines for 13 input methods in Apple HIG. Gestures (consistency of standard gestures · use custom gestures sparingly) / Keyboards (Full Keyboard Access · standard shortcuts · modifier key order) / Pointing devices (three iPadOS content effects: highlight/lift/hover · snapping · macOS pointer style catalog) / Apple Pencil & Scribble (tilt, pressure, azimuth · hover preview · double tap/squeeze/barrel roll · PencilKit) / Digital Crown (primary navigation on watchOS · immersion adjustment on visionOS) / Eyes (visionOS gaze + hover effect · privacy: gaze point not exposed) / Focus and selection (five tvOS focus states · iPadOS focus groups) / Game controls (touch virtual buttons · physical controller ABXY mapping · single-key keyboard controls) / Action button / Camera Control / Nearby Interaction / Gyroscope and accelerometer / Remotes. Each entry includes applicable platforms + best practices + anti-patterns.
Apple HIG 25 cross-component 'interaction flow' design patterns. Onboarding trio (Launching · Loading · Onboarding) / Data and input (Entering data · Drag and drop · Undo and redo · Searching · Charting) / Feedback and interruption (Feedback · Modality · Managing notifications) / Account settings (Managing accounts · Settings · Ratings and reviews) / Content and media (Audio · Video · Haptics · Full screen · Live-viewing) / System collaboration (Collaboration and sharing · File management · Multitasking · Printing · Workouts). Each pattern provides core principles + best practices + anti-patterns + platform differences. For component-level usage, see components; this section covers 'stringing components into flows'.
In-depth design special topic for iPad and Mac platforms (Will’s reorganized deep read of iOS/macOS). Essential differences between Mac and iPad (touch vs. keyboard & mouse · 77% font scaling · neutral colors · gesture mapping) + three major frameworks · information layout / iPadOS 26 new window system (drag handle · window controls inside toolbar · one window per document · descriptive naming) / pointer (1:1 tracking · liquid glass highlight disc) / 3 adaptive cursor effects (Highlight / Lift / Hover) + adaptive vs. traditional pointer (adaptive precision · magnetism · realignment) / material form design / navigation (sidebar ↔ tab bar adaptation · flat vs. hierarchical · Push vs. Modal · navigation design derived backward from core value) / UX copywriting (PACE framework · VOICE & TONE · 4-step tone modulation · 4 improvement tips: delete filler words / avoid repetition / start with why / build a glossary · Great Alerts · Tips · teach users to use the app) / chart system (Marks / Axes / Descriptions / Interaction / Color · static vs. interactive · macro-meso-micro three layers · differentiation across multiple charts · fixed vs. dynamic range) / subscription design (Effortless / Transparent / Engaging · 3 steps 61% vs. 9 steps 7%) / modal sheets / safe area / Live Activities · Shortcuts · App Clips · Intents · Interactive Snippets · custom visual effects (iOS 18 SwiftUI).
