Is it weird that AI coding assistance is not giving... 🔗 **Source:** [Is it weird that AI coding assistance is not giving...](https://twitter.com/esrtweet/status/2023978360351682848/?rw_tt_thread=True) > [!note]+ Document Notes > tags: [AI, programming, identity] > author: Eric S. Raymond > source: Twitter > reference: [https://twitter.com/esrtweet/status/2023978360351682848/?rw_tt_thread=True] > date created: 2026-02-20 14:30 > > # 🏷️ Title: Is it weird that AI coding assistance is not giving... > <!-- deep-read:sections start --> ## 🧭 Table of Contents - [📝 Summary](#-summary) - [📖 Main Content](#-main-content) - [🗒️ Notes](#️-notes) - [📚 References](#-references) - [🗣️ Paraphrase](#️-paraphrase) - [📰 Newsletter](#-newsletter) - [✅ Action Items](#-action-items) - [🧪 Hypotheses](#-hypotheses) - [📊 Methodology](#-methodology) - [❓ Questions](#-questions) - [🔮 Future Research](#-future-research) - [💡 Implications](#-implications) - [🧠 Learning Styles Assessment](#-learning-styles-assessment) - [📊 Accessibility Assessment](#-accessibility-assessment) - [🌍 DEIB Assessment](#-deib-assessment) - [✅ Logical Validity](#-logical-validity) - [🛡️ DARVO Analysis](#-darvo-analysis) - [🎯 Conclusion](#-conclusion) ### 🔗 References - Eric S. Raymond, “Is it weird that AI coding assistance is not giving me identity fracture?” Twitter thread, 2026-02-18. --- ## 📝 Summary - What happened. - Raymond reflects on why AI coding tools don’t threaten his identity as a programmer, unlike many peers. - Why it matters. - It reframes programming identity from “hand-coding” to “directing power” via evolving tools. - Key stakeholders. - Software developers, AI tool builders, tech leaders. - Risks and tradeoffs. - Risk of alienating those grieving change; tradeoff between craftsmanship and leverage. - Bottom line. - See yourself as the wizard, not the spellbook: your value is the mindset that bends reality, regardless of whether code is typed or spoken to AI. --- ## 📖 Main Content - Raymond notes many developers feel “disoriented and threatened” as “programming by hand is clearly going the way of the buggy whip and the hand-cranked auger” ([wizard metaphor section](https://read.readwise.io/read/01khsvaa6dghrrnq0vst3ewxdj?endLexposition=00000_00005_00000-002n&startLexposition=00000_00000&emphasizeText=Programming+by+hand+is+clearly+going+the+way+of+the+buggy+whip)). - He contrasts this with his own reaction: he doesn’t miss hand-coding, just as he didn’t miss assembler after compilers “ate the world.” - Core idea: your identity should not be “person who types code” but “wizard” who “weave[s] complex patterns of logic that make things happen in the world. Writing code is a way to manifest my will.” - He extends the metaphor: languages are “arcane knowledge” and AI tools are “powerful familiars” and “machine spirits.” These come and go, but “you are the magic-wielder. Without you, none of it happens” ([closing passage](https://read.readwise.io/read/01khsvaa6dghrrnq0vst3ewxdj?endLexposition=00000_0000A_00000-000y&startLexposition=00000_00006&emphasizeText=you%3F+You+are+the+magic-wielder.+Without+you%2C+none+of+it+happens.)). - Thus, AI doesn’t fracture his identity; it upgrades his “spells” and assistants. --- ### 🗒️ Notes - Many developers feel identity threat as AI reduces the need for manual coding craft. - Author’s stance: tools evolving (assembler → compilers → AI) is normal; attachment to old tools is misplaced. - Wizard metaphor: - Programmer = wizard (agent of change, will, and understanding). - Code/languages = spells, “runes of power.” - AI tools = “powerful familiars,” “obedient machine spirit.” - Skill redefined: - Not “knowing 8086 assembler” but “assuming the kind of mental states required to bend reality with the application of spells.” - The enduring asset is problem-solving, systems thinking, and clarity of intent. - Key identity reframing move: - Does it matter if spells are “painstakingly scribed” or “spoken to an obedient machine spirit”? - Answer: “It’s all one; it’s all the manifestation of will.” - Historical analogy: - Compilers made low-level coding rare; AI may do the same to much hand-typing of higher-level code. - Emotional offer: - He acknowledges his lack of disorientation may “disqualify” him from speaking, but suggests others “emulate my mental stance and be completely unbothered.” --- ## 📚 References - Eric S. Raymond on AI coding assistance and programmer identity (Twitter thread): https://twitter.com/esrtweet/status/2023978360351682848/?rw_tt_thread=True - Historical context on high-level languages superseding assembly (general CS history, e.g., early compiler development). --- ## 🗣️ Paraphrase - Paragraph A paraphrase. - Many programmers feel threatened as AI makes manual coding less central; their sense of self is tightly tied to the craft of hand-writing code. - Paragraph B paraphrase. - Raymond argues that the real essence of being a programmer is the ability to think like a “wizard” who uses evolving languages and tools—now including AI—to impose their will on reality, regardless of how the code is actually produced. --- ## 📰 Newsletter In this week’s WordPress-adjacent musings, Eric S. Raymond offers a timely frame for AI-era development: you’re not your hand-written code; you’re the wizard directing it. As AI copilots seep into plugin shops, agencies, and core workflows, the durable skill is clear intent and architectural thinking, not keystrokes. Curious how this plays out for WordPress builders? - Experiment with AI-assisted scaffolding for plugins and themes, then review the generated code like a senior engineer. - Use tools like Code Snippets, Playground, and WP-CLI scripts as your “familiars,” freeing attention for UX, performance, and security. - For learning, pair AI with docs: ask it to explain hooks and filters in your own project context, then validate against the Codex and DevHub. The takeaway: treat AI as an upgrade to your spellbook, not a rival wizard. --- ## ✅ Action Items #### DevRel - Developer advocacy - Publish a talk or blog series on “AI as familiar, not replacement” for WordPress and PHP devs, with practical demos. - Education - Design workshops where attendees refactor AI-generated WordPress code for readability, security, and performance. - Developer experience - Provide example prompts and recipes for using AI with WP-CLI, REST API, and plugin boilerplates. - Developer marketing - Run campaigns around “build more, type less” emphasizing human design and architecture. - Community management - Facilitate discussions acknowledging anxiety about AI while steering toward the “wizard” framing. - Internal/external/cross-project - Gather feedback from agencies and hosts on AI-assisted workflows and feed it into docs and tooling. #### OSPO - Policy & Governance - Draft guidance on acceptable AI-assisted contributions (e.g., disclosure, license compatibility checks). - License Compliance - Integrate AI-aware scanning into CI to detect problematic snippets. - Security Management - Require review of AI-generated code for common vulns (XSS, SQLi, CSRF) before merging. - Software Inventory - Track where AI-generated code lives in repos for future audits. - Developer Support - Offer office hours on responsibly integrating AI into open source workflows. - Community Engagement - Support upstream tools that help validate and explain AI-generated patches. - Strategic Alignment - Add AI-impact KPIs (e.g., reduced toil, review load, defect rates) to OSPO roadmaps. #### Community - Maintainers - Host “review clinics” focusing on AI-generated PRs and setting expectations. - Contributors - Create contributor guides that show how to use AI for learning, not bypassing understanding. - Sponsors - Encourage funding for review, tooling, and education around AI-assisted contributions. - Ecosystem - Coordinate cross-project incident drills for AI-induced bugs or license issues. #### WordPress - Core - Clarify stance on AI-generated contributions and ensure review capacity. - Core Performance - Benchmark AI-generated patterns against performance budgets. - Core AI - Define “responsible AI” principles and evaluation sets aligning with the wizard/assistant framing. - Core Program Team - Align release processes with new kinds of contributions and tools. - Mobile - Explore AI-assisted scaffolding for mobile UI around the REST API and Gutenberg. - Playground - Provide labs showing AI-powered prototyping of plugins/themes in-browser. - Community - Add AI topics to meetups and WordCamps with a focus on human judgment and design. - Training - Build learning paths where students use AI as a tutor while still writing and reviewing code. - Marketing - Emphasize builders’ creativity and problem-solving as the enduring value. - Support - Document how support teams can leverage AI while avoiding hallucinations. - Design - Encourage AI as a generator of options, with humans curating and refining. - Documentation - Add guidance on reading and validating AI-generated examples. - Photos / TV / Polyglots / Accessibility / Meta / CLI / Hosting / Tide / Openverse / Themes / Plugins - Integrate AI cautiously as helper tools (translation assistance, caption drafts, code scaffolding, scoring suggestions), always keeping human review and governance in place. #### The WPOCC - Programs - Offer microgrants for experiments on AI-assisted maintenance and documentation. - Projects - Run governance clinics on how AI affects contributor roles and responsibilities. - Funding experiments - Pilot matched grants for projects that demonstrably improve safety and quality of AI-assisted contributions. ### FAIR - Technical independence - Prefer AI tools that allow local or controllable models and clear provenance. - Policies and covenants - Require transparency around AI assistance in funded or governed projects. - Productized governance artifacts - Ship ready-to-adapt charters and operator agreements covering AI usage norms. --- ## 🧪 Hypotheses - H1: Programmers who adopt the “wizard” identity will experience less anxiety about AI tools than those who identify primarily as hand-coders. - H2: Exposure to historical tooling shifts (e.g., assembler → compilers) reduces perceived threat from AI coding tools. - H3: Framing AI as “familiar” rather than “replacement” increases willingness to experiment with AI in workflows. - H4: Teams that formalize review practices for AI-generated code will see fewer AI-induced defects. --- ### 📊 Methodology - Study designs - Surveys and interviews of developers about identity, tool attitudes, and stress levels. - Controlled experiments where one group gets “wizard” framing and historical context, another does not. - Data sources - Developer communities (WordPress, PHP, JS, etc.), commit histories, PR review comments. - Analyses - Compare stress/attitude scores across framings; correlate experience with past tool shifts and AI comfort. - Track defect rates and review comments for AI vs non-AI code over time. --- ### ❓ Questions - How much of programmer identity is tied to specific tools vs underlying problem-solving skills? - Which metaphors (wizard, architect, craftsperson) best reduce AI-related anxiety? - How can education help new devs avoid over-reliance on AI while still benefiting from it? - What governance patterns help open source projects safely absorb AI-assisted contributions? --- ### 🔮 Future Research - Longitudinal studies on how developer identity evolves as AI tooling matures. - Comparative research across professions (writers, designers, analysts) facing similar AI shifts. - Best-practice playbooks for code review and licensing in the AI era. - Exploration of how AI changes on-ramps for junior developers. --- ### 💡 Implications - Practice - Focus training on architecture, debugging, and domain understanding, not syntax memorization. - Policy - Update contribution guidelines and licensing checks for AI-generated content. - Security - Treat AI code as untrusted until reviewed; add automated scanning. - Education - Teach AI literacy and critical reading of generated code. - Funding models - Support maintenance, review, and governance for AI-heavy ecosystems, not just raw feature output. --- ### 🧠 Learning Styles Assessment - Reading level estimate - Roughly high school to early college (Flesch-Kincaid around grade 10–12): metaphor-heavy but short and conversational. #### 👀 Visual learners - Present - Strong imagery via wizard, spells, familiars. - Improve - Add diagrams showing historical tool evolution and where AI fits. #### 🔊 Aural learners - Present - None explicitly. - Improve - Provide a short audio reading of the thread plus commentary. #### 🗣️ Verbal learners - Present - Vivid metaphors and analogies (wizard, runes, machine spirits). - Improve - Add explicit definitions of “identity fracture,” “languages of invocation,” etc. #### 🧩 Kinesthetic learners - Present - None. - Improve - Hands-on exercises: replace manual coding in a task with AI-assisted coding and reflect. #### 🧠 Logical learners - Present - Implicit causal chain (tools evolve → some feel threatened → reframe identity → anxiety drops). - Improve - Make the argument more explicit with step-by-step reasoning and examples. #### 🧑‍🤝‍🧑 Social learners - Present - Addressed to a community of programmers. - Improve - Add prompts for group discussion and peer mentoring around AI adoption. #### 🧘‍♀️ Solitary learners - Present - Reflective tone invites self-examination. - Improve - Include journaling questions on personal identity as a developer. #### 🌿 Naturalistic learners - Present - None, aside from oblique references to evolution/change. - Improve - Use ecosystem metaphors (toolchains as evolving environments). --- ### 📊 Accessibility Assessment #### Screen reader semantics - Present - As a tweet thread, basic text; headings depend on hosting context. - Improve - When archived, ensure proper headings, language attributes, and descriptive titles. #### Keyboard navigation - Present - Standard Twitter navigation. - Improve - When republished, ensure logical tab order and skip links. #### Contrast and readability - Present - Twitter default (generally acceptable). - Improve - When reusing, maintain good contrast and readable font sizes. #### Text resizing and reflow - Present - Twitter responsive layout. - Improve - Check reflow at 200% zoom when embedding. #### Media and captions - Present - Pure text; no media. - Improve - Optional audio with transcripts. #### Forms and interactions - Present - Likes/replies/retweets; accessibility depends on platform. - Improve - When embedding, ensure controls are reachable and labeled. #### Standards reference - Map improvements to WCAG 2.2 AA success criteria (1.4.x, 2.1.x, 3.3.x, etc.). --- ### 🌍 DEIB Assessment #### Cultural breadth - Present - General tech culture; references like 8086 and 1983 may be US/Western-centric. - Improve - Broaden examples beyond specific hardware histories. #### Inclusive language - Present - Mostly neutral; casual “you youngsters” aside. - Improve - Avoid age-related framing; define jargon for non-experts. #### Neurodiversity support - Present - Short, chunked thread; consistent metaphor. - Improve - More explicit structure and summaries for readers who struggle with metaphor-heavy prose. #### Representation - Present - Single author perspective, implicitly experienced male engineer. - Improve - Include voices from varied backgrounds on AI and identity. #### Psychological safety - Present - Validates that disorientation exists; offers a nonjudgmental alternative stance. - Improve - Explicitly acknowledge grief and loss some may feel and suggest support resources. --- ### ✅ Logical Validity #### Strong reasoning - Type: analogical and inductive. - Analogy: assembler → compilers → AI tools; tools evolve, core skills persist. This is strong because we have historical evidence that programmers adapted and gained productivity. - Analogy: programmer as wizard; tools as spells/familiars. Strong at conveying that agency resides in the human, not the tool. #### Fallacies - Potential overgeneralization - Implies the wizard framing might broadly solve disorientation but offers no data; it’s more inspirational than evidential. - Fix: note that some roles may indeed shrink or change dramatically despite mindset shifts. #### Tone of voice - Tone - Conversational, slightly wry, confident, and gently mentoring. - Examples: “for you youngsters,” “Same as it ever was. Same is it ever was. And so mote it be.” --- ### 🛡️ DARVO Analysis - Deny - No denial of others’ feelings; he explicitly notes many feel disoriented. - Attack - No attacks on those who feel threatened; no undermining of credibility. - Reverse victim/offender - He does not cast himself or AI as a victim; instead he offers a coping frame. - Overall - No DARVO pattern present; the rhetoric is reframing and invitational rather than defensive. --- ## 🎯 Conclusion Raymond argues that AI coding assistance need not threaten a programmer’s identity if that identity is rooted in problem-solving and intentionality rather than typing code. Tools—languages, compilers, AI “familiars”—come and go, but the “wizard” remains the source of magic. Practically, developers and communities can focus on strengthening human judgment, design, and review, while embracing AI as another powerful tool in the spellbook. --- ## 🧱 Refactor Appendix (append after 🎯 Conclusion) ### 📚 Literature Split Pad - Purpose: Stage Konik CEE units that become per-source Literature notes. - Split level: **H3**. - Include verbatim quotes under 25 words and a [view highlight] link. ### [H1] 5–10-word present-tense claim > [!quote] [{Title or entity}]({ARTICLE_URL}) by [[{Author or [MISSING]}]] via [[{Site or [MISSING]}]]. Published {DATE_OR_[MISSING]}. Accessed . > “Short quote, verbatim, ≤25 words.” Two lines: context and use. Cite data if relevant. [view highlight]({ARTICLE_URL}#H1) #xref #articleseed ID:rw_[MISSING] ### [H2] 5–10-word present-tense claim > [!quote] [{Title or entity}]({ARTICLE_URL}) > “Short quote, verbatim, ≤25 words.” Two lines: context and use. [view highlight]({ARTICLE_URL}#H2) #xref ID:rw_[MISSING] --- ### 🔖 Atomic Split Pad (Zettel Candidates) - Purpose: Promote reusable claims into Atomic notes. - Split level: **H4**. - Each item: one idea, present tense, specific, testable. #### [Z1] 5–10-word present-tense claim - Scope/condition in one line. - Pointer to page/time or dataset. - Links: [[Parent-Concept]] [[Related-Idea]] #### [Z2] 5–10-word present-tense claim - Scope/condition in one line. - Pointer to page/time or dataset. - Links: [[Parent-Concept]] [[Related-Idea]] --- ### ✏️ Fleeting Queue (Extract by Selection) - Purpose: Capture quick actions or sparks for Fleeting notes. - Extraction: **Select one line** → “Refactor: Extract selection to new note (content only).” - Keep each line short and self-contained. - [F1] Short action — context/{ARTICLE_URL} #fleeting #inbox - [F2] Short action — context/{ARTICLE_URL} #fleeting #inbox - [F3] Short action — context/{ARTICLE_URL} #fleeting #inbox --- <!-- deep-read:sections end -->