Footnotes & References
MX: The Introduction
Source References
- [ai-internals]
-
For deeper exploration of AI statistical foundations and linguistic bias, see my blog posts.
- The stripped-down truth: how AI actually works
- Does AI mean algorithmic interpolation
- The digital language caste system
- The mathematical heartbeat of AI
- The tokenization trap: processing German
- The no-elephants problem
- When a five-year-old beats an AI
- A framework for evaluating AI confidence
- Bystander effect (Simply Psychology)
- Diffusion of responsibility (The Decision Lab)
- [devops-qa]
-
DevOps implementations demonstrate the balance between shared responsibility and specialized roles. Whilst everyone takes responsibility for quality, dedicated QA engineers focus on defining standards, designing test frameworks, and preventing bugs.
- [organizational]
-
Research shows 82% of respondents have limited ability to hold others accountable, and only 14% of employees feel their performance is managed in ways that inspire responsibility. Clear ownership structures address this gap.
- [cmscritic]
-
Tom Cranstoun, "A CMS Consultant's Takeaways from CMS Kickoff 2024," CMS Critic, February 2024. The article argued that AI's real value lies in consuming content, not generating it, and proposed treating "Machine" as a fourth device type alongside mobile, tablet, and desktop.
- [cowork]
-
Anthropic launched Claude Cowork in January 2026, marking a fundamental shift from chatbot to autonomous digital colleague. The system manages local file systems, orchestrates workflows, and executes complex tasks through multi-agent architecture.
- [sovereignty]
-
For a fuller treatment of data sovereignty's two dimensions, jurisdictional control and ownership control, and how MX practices address both.
- [principles]
-
For a practitioner's account of how these principles reshape day-to-day building practice.
- [cms-workflow]
-
For the full argument that the CMS is shifting from destination to invisible workflow.
- [gathering]
-
The Gathering, open standards for Machine Experience. Community-governed, MIT licensed, practitioner-led.
Industry and Market References
- [adobe-holiday]
-
Adobe Holiday 2025 data: AI referrals to retail surged 693% year-over-year, travel 539%. AI-referred visitors converted 31% higher than other traffic and were 33% less likely to bounce. Based on over 1 trillion visits to U.S. retail sites.
- [platform-launches]
-
Four machine commerce platforms launched within eight days in January 2026: Amazon Alexa+ at CES (5 January), Microsoft Copilot Checkout at NRF with PayPal and Shopify (8 January), Google Universal Commerce Protocol at NRF with Target and Walmart (11 January), and Anthropic Claude Cowork (12 January).
- [adamuz]
-
The 2026 Adamuz train derailments killed 46 and injured 292. Spain's worst rail disaster since Santiago de Compostela (2013). Iberia capped fares at €99 and added flights. Spain's national ombudsman investigated transport companies. Consumer association FACUA called for legal reform.
- [commoncrawl]
-
Common Crawl language distribution statistics. Language identification via Compact Language Detector 2 (CLD2) across monthly crawl archives. English consistently represents approximately 44–46% of crawled content.
- [webmcp]
-
WebMCP (Web Model Context Protocol), W3C Draft Community Group Report, 10 February 2026. Created by engineers at Google and Microsoft under the W3C Web Machine Learning Community Group. Early preview in Chrome 146 Canary.
- [cloaking]
-
Google defines cloaking as "presenting different content to users than to search engines" and classifies it as a violation of spam policies. Sites caught cloaking face manual actions including removal from search results.
- [llmoptimizer]
-
Adobe LLM Optimizer, edge-based deployment that detects agentic traffic and serves AI-friendly content modifications at the CDN layer. Targets only agentic requests; does not affect human users or SEO bots.
- [huggingface]
-
Hugging Face model growth data: model tracking began in March 2022. The first million models took over 1,000 days; the second million arrived just 335 days later. By August 2025, models added that year had already surpassed the entire 2024 total.
- [iso4217]
-
ISO 4217 defines three-letter currency codes (EUR, INR, GBP, USD) that eliminate currency identification ambiguity. Schema.org's price specification requires numeric values to use a period as the decimal separator with no thousands grouping, whether expressed as a Microdata
contentattribute or a JSON-LD "price" property. This convention, aligned with ISO 80000-1 (Quantities and units), eliminates all cultural formatting ambiguity at the data layer. -
This chapter was co-authored with Claude (Anthropic). The frontmatter metadata, the structured arguments, and the MX patterns described in the text are all implemented in the chapter itself, the medium demonstrates the message.
Online Appendices
Full appendices and additional materials are maintained online. These resources include practical implementation guides, quick references, and real-world case studies.