# DevRelCon NYC 2026 MOC > Event map. Topic home: [[3. Resources/MOCs/DevRel MOC]] · Sister maps: [[3. Resources/MOCs/Events/Write the Docs Berlin 2026 MOC]] · Index: [[3. Resources/MOCs/MOCs|Maps of Content → Events & Conferences]] A map of 34 DevRelCon 2026 NYC talk recordings captured to Reader and Obsidian on 2026-10-15. The Dev Rel channel published them on 2026-10-14 (https://nyc.devrelcon.dev/). This is a **session index and thematic map**: each talk gets one line here, and the linked note carries the video embed, the 20-section deep-read notes, and the auto-caption transcript. ## Core Question When coding agents read the docs, call the APIs and mediate the relationship with developers, which DevRel work still needs a person, which of it needs rebuilding for an agent audience, and which of the speakers' numbers come with a method? > [!note] Captions garble most proper nouns, and each linked note lists its corrections. Timestamps are transcript positions. Almost every figure below is a single speaker's own report, so treat the numbers as leads to test against your own data, not as results. --- ## 🤖 Docs and APIs that coding agents read and call Six talks treat the coding agent as a reader and caller. Mapbox, a docs-platform builder, an API vendor and a go-to-market team each report the same shift: the agent reads before the developer does. - [[3. Resources/Readwise/Articles/Copy as Markdown Is Table Stakes. Is Your Docs Platform Ready — Andrew Sepic]] — Andrew Sepic (Mapbox): 2,000+ docs pages served as clean markdown by a build-twice Docusaurus pipeline and an Accept-header path; one week showed 460 human uses against 47,000 LLM requests (00:06:39), classification method unstated. Serving a page doesn't tell an agent which page answers its question. - [[3. Resources/Readwise/Articles/Stop Writing Knowledge Bases. Build Decision Systems Instead — Danielle Washington]] — Danielle Washington: a knowledge base returns options; a decision system carries purpose, default, condition, tradeoff and next action on each page, because context that lives only in navigation is lost at retrieval. - [[3. Resources/Readwise/Articles/Make Coding Agents Great at Using Your API — Kevin Whinnery]] — Kevin Whinnery: a coding agent writes the integration now, so what the agent knows about your API sets the developer's success rate. Agent-facing files and skills need their own evals. - [[3. Resources/Readwise/Articles/Developer First = Agent First — Hahnbee Lee]] — Hahnbee Lee: developer-first design is agent-first design. The clarity that won DevEx in the 2010s (Stripe, HashiCorp) serves agents that read docs and call APIs today. - [[3. Resources/Readwise/Articles/Yes-Code - The Building Blocks for Agents — John Britton]] — John Britton: no-code fails at the last 10% with no escape hatch. "Yes-code" ships stable, lifecycle-aware primitives into your codebase with a control plane (sandboxes, unified logging, dynamic configuration) behind them. - [[3. Resources/Readwise/Articles/Agent Led Growth How to Get Claude Code to Pick Your Dev Tool — Kara Silverman & Caelean Barnes]] — Kara Silverman and Caelean Barnes: agents research, choose and install tools with no human sales contact. In an unpublished 500-session sample, being recommended was not being chosen (Knock mentioned 56%, installed 31%), and the easier setup won (00:18:22). ## 🧭 Agent experience, DX at AI speed, and what DevRel becomes Seven talks on the role itself: where humans sit in an agent loop, what a DX team ships when agents read docs, and which DevRel work stays valuable when code is cheap. - [[3. Resources/Readwise/Articles/Agentic Experience Design A New Interaction Layer Demands a New Discipline — Kurtis Kemple]] — Kurtis Kemple: agentic systems hand the middle two steps of the intent, plan, outcome, judgment loop to software, so the design problem moves to handover, trust and steerability. - [[3. Resources/Readwise/Articles/Evolving DX at the Speed of Frontier AI — Dominik Kundel]] — Dominik Kundel: agents run for hours, Codex maintains docs, screenshots and demos, and non-developers now use it, so "developer" no longer describes the audience. Adoption figures are unsourced. - [[3. Resources/Readwise/Articles/Build Something Agents Want, Not More AI Slop — Brian Douglas]] — Brian Douglas: volume is cheap and judgment isn't. Taste is the filter between work worth shipping and slop, and agents don't need a better model, they need to remember what you did before (00:27:13). - [[3. Resources/Readwise/Articles/Touching Grass in the Age of Agents — Carter Rabasa]] — Carter Rabasa: AI now mediates DevRel's relationship with developers. His test: if you can't serve one developer, you can't serve any. Do things that don't scale. - [[3. Resources/Readwise/Articles/DevRel Is Dead. Long Live DevRel! — Swift & Jess Lee]] — Swift and Jess Lee: when code is cheap the durable jobs are ideas, taste and distribution (00:07:21). Agents become buyers and the funnel hollows into a barbell (00:13:51). - [[3. Resources/Readwise/Articles/Developer Trends to Bet the Next Year of Your Career on — Matt Makai]] — Matt Makai: spend 1 to 2 weeks hands-on, weeks 3 to 8 calibrating with trusted people, then rely on adoption data. He names MCP, harness engineering and local AI as durable. - [[3. Resources/Readwise/Articles/Building Your Own Software Factory — David Crawshaw]] — David Crawshaw: an agent can modify a program and keep it current against upstream, so software written for yourself stops being disposable. Open source is the precondition. ## 📊 Measuring DevRel, adoption and ecosystems Seven talks on what to count. Each one separates a volume number from the outcome it is supposed to stand for. - [[3. Resources/Readwise/Articles/PMF, Pipeline, and GTM Aren't Someone Else's Problem — Kevin Blanco]] — Kevin Blanco: "developers" isn't an audience. Place the product on Rogers' 1962 diffusion curve and match DevRel activity to the PMF stage. - [[3. Resources/Readwise/Articles/The Ecosystem Tube Map — Dugald Morrow]] — Dugald Morrow: model the ecosystem as a tube map of workstreams, not one funnel. A causal chain is a hypothesis until each link is tested. - [[3. Resources/Readwise/Articles/Turning Education Signals Into Account Intelligence — Joel Lord]] — Joel Lord: one skill check works as a standard measurement unit across workshop, webinar, booth and video formats and feeds per-account dashboards. Conversion figures are self-reported with no baseline. - [[3. Resources/Readwise/Articles/DevRel for a Developer You'll Never Hear From — Nikita Jotwani]] — Nikita Jotwani (HubSpot): the MCP adoption program saw tool usage rise without active usage. Connector products fail in silence, so adoption needs an engagement measure. The talk gives no figures. - [[3. Resources/Readwise/Articles/The Market Is Trying to Tell You Something - Joey De Villa]] — Joey de Villa: job descriptions copy a pre-AI template, DevRel's boundaries with growth and sales have moved, and DevRel ROI now looks like sales and marketing ROI. - [[3. Resources/Readwise/Articles/The Soloist Trap Fixing Autonomy Without Alignment in DevRel — Drew Gorton]] — Drew Gorton: the soloist trap is autonomy without alignment, a five-step domino chain from ambiguous purpose to unprovable value. Alignment constraints come before individual autonomy. - [[3. Resources/Readwise/Articles/Developer Marketing Is Not a Four-Letter Word — Lisa Tagliaferri & Beatriz Datangel Rodgers]] — Lisa Tagliaferri and Beatriz Datangel Rodgers: developer marketing done with substance (docs, security messaging, honest demos) works because trust can't be faked. ## 🎓 Developer education, onboarding and friction Three talks on getting a developer from arrival to understanding. - [[3. Resources/Readwise/Articles/From Badge Hunters to Builders A Former Teacher’s Guide to Fixing Developer Education — Sean Keegan]] — Sean Keegan: gamification is a door and fails when completions become the only metric. Climb Bloom's taxonomy and end with something the learner built and kept. - [[3. Resources/Readwise/Articles/How to Re-Engage Developers Who Ship Without Understanding — Arkodyuti Saha]] — Arkodyuti Saha: shipping is nearly free and understanding optional, so output-rewarding rubrics produce a comprehension gap (00:05:22). - [[3. Resources/Readwise/Articles/What Makes Developers Actually Ship A Workshop on Removing Friction in the First Mile — Ojus Save]] — Ojus Save: developers arrive for an outcome, so every step between arrival and the first usable result ("the first mile") is a place to lose them. ## 🤝 Community, champions, programs and events Six talks on the human systems under a developer program. - [[3. Resources/Readwise/Articles/How to Build a Champion's Program — Laís Carvalho & Dawn Wages]] — Laís Carvalho and Dawn Wages: a champions program works when friction drops, incentives rise, and two sets of success criteria exist, one for go-to-market and one for community. - [[3. Resources/Readwise/Articles/A Practical Framework to Scale Developer Communities — Ully Sampaio]] — Ully Sampaio: communities scale by mapping the contributor journey and asking for the smallest meaningful next action, like a Duolingo streak, not by recruiting top contributors first. - [[3. Resources/Readwise/Articles/What I Learned Building Developer Community for a Technology Nobody Had Heard of — Lauren Lee]] — Lauren Lee: an ecosystem starts as your team manually standing in for systems that don't exist yet, and holds only if you act with integrity when nobody forces you to. Not a framework talk. - [[3. Resources/Readwise/Articles/Why I Flew to China to Make a Conference Badge — Shy Ruparel]] — Shy Ruparel: a conference badge made in China. Set a promise boundary before you start, keep a working contingency, and never let the plan depend on one person. - [[3. Resources/Readwise/Articles/Can You Keep a Stadium Make Them Feel Smart. — Jonathan Murray]] — Jonathan Murray: people who don't adopt AI weren't failed by technology, the explanation failed them. "People that feel stupid they go away quietly." - [[3. Resources/Readwise/Articles/I Learned Everything I Know From Dev Rel - Stories From a Not-a-Developer — Meghan Grady]] — Meghan Grady: five stories from Twilio, Stripe and DigitalOcean where a DevRel team member did work that sales, partnerships or product couldn't do alone. ## ✍️ Craft, taste, content and values Five talks on the working practice: what to make, how to capture it, and what to hold to. - [[3. Resources/Readwise/Articles/Reflections on Taste — Juan Pablo Flores Cortés & Clement Hugbo]] — Juan Pablo Flores Cortés and Clement Hugbo: taste is a way to communicate intent, built from relationships, context and a story that makes choices legible. Useful low-fidelity slop differs from unreviewed output published as finished work. - [[3. Resources/Readwise/Articles/SOUL.md The Root Configuration File for DevRel — Tim Falls]] — Tim Falls: write down your values, the conditions you need to uphold them and your fallback, and revise it, before AI pressure rewrites the job. - [[3. Resources/Readwise/Articles/12 PR Pitches That Actually Worked on Me, a Cynical Former Tech Reporter — Kevin Purdy]] — Kevin Purdy: pitches worked when they handed a reporter something concrete and exclusive (early access, an engineer to quote, a site visit). Broad executive pronouncements didn't. Nine tactics are in the talk; the other three are on his site. - [[3. Resources/Readwise/Articles/Blog While the Context Is Fresh — Chuck Meyer]] — Chuck Meyer: have the coding agent dump decisions and dead ends into a running blog.md at set moments, then hand it to a less invested agent to find the story. - [[3. Resources/Readwise/Articles/Finding Success With a Company YouTube Channel — Alex Kretzschmar]] — Alex Kretzschmar: weekly, useful, honest videos grew a company channel. The spoken figures (75,000 subscribers, 6 million views in 2025) conflict with the description (66k, 3 million). --- ## 🔗 Cross-talk threads These are patterns across the 34 talks, drawn from the deep-read notes. Each cites the talks that support it. 1. **The agent is the first reader.** Thirteen talks in the first two clusters treat a coding agent as the reader of your docs or the caller of your API. They differ on what to build for it: clean markdown ([[3. Resources/Readwise/Articles/Copy as Markdown Is Table Stakes. Is Your Docs Platform Ready — Andrew Sepic]]), decision-carrying pages ([[3. Resources/Readwise/Articles/Stop Writing Knowledge Bases. Build Decision Systems Instead — Danielle Washington]]), agent-facing skills with evals ([[3. Resources/Readwise/Articles/Make Coding Agents Great at Using Your API — Kevin Whinnery]]), and primitives with an escape hatch ([[3. Resources/Readwise/Articles/Yes-Code - The Building Blocks for Agents — John Britton]]). 2. **A volume number needs an outcome partner.** Four talks make the same move on different numbers. Tool usage is not active usage ([[3. Resources/Readwise/Articles/DevRel for a Developer You'll Never Hear From — Nikita Jotwani]]). Mention rate is not install rate: 56% against 31% for one tool in an unpublished sample ([[3. Resources/Readwise/Articles/Agent Led Growth How to Get Claude Code to Pick Your Dev Tool — Kara Silverman & Caelean Barnes]]). A completion count is not something built ([[3. Resources/Readwise/Articles/From Badge Hunters to Builders A Former Teacher’s Guide to Fixing Developer Education — Sean Keegan]]). Shipped output is not understanding ([[3. Resources/Readwise/Articles/How to Re-Engage Developers Who Ship Without Understanding — Arkodyuti Saha]]). Douglas adds the general case, that a measure which becomes the target stops measuring (00:13:00, [[3. Resources/Readwise/Articles/Build Something Agents Want, Not More AI Slop — Brian Douglas]]). 3. **Tension: agent-first or human-first docs.** Lee says one clarity serves both audiences ([[3. Resources/Readwise/Articles/Developer First = Agent First — Hahnbee Lee]]). Washington says an agent reads before the developer, so a page must carry its own routing context ([[3. Resources/Readwise/Articles/Stop Writing Knowledge Bases. Build Decision Systems Instead — Danielle Washington]]). Alexandra Klepper's Write the Docs Berlin talk argues for optimizing docs for the human users of agents, not for agents ([[3. Resources/Readwise/Articles/Write Inclusive, People-First Language About AI - Alexandra Klepper]]). The talks don't test the claims against each other. 4. **Judgment is the scarce input.** Douglas ([[3. Resources/Readwise/Articles/Build Something Agents Want, Not More AI Slop — Brian Douglas]]), Flores Cortés and Hugbo ([[3. Resources/Readwise/Articles/Reflections on Taste — Juan Pablo Flores Cortés & Clement Hugbo]]), and Swift and Lee ([[3. Resources/Readwise/Articles/DevRel Is Dead. Long Live DevRel! — Swift & Jess Lee]]) each name taste or judgment as the work that stays valuable when generation is cheap. Falls ([[3. Resources/Readwise/Articles/SOUL.md The Root Configuration File for DevRel — Tim Falls]]) and Rabasa ([[3. Resources/Readwise/Articles/Touching Grass in the Age of Agents — Carter Rabasa]]) point at the same thing from the person's side: written values, and serving one developer by hand. 5. **Evidence is thin and mostly self-reported.** Mapbox's 460 against 47,000 covers one week with no classification method. The 500-session sample is unpublished. Kretzschmar's spoken 75,000 subscribers conflicts with the description's 66k. Lord and Jotwani give no baseline or figures. Kundel's feature counts are unsourced. Use the numbers as leads. ### Tests you can run on your own docs and reporting - Count `.md` requests and `Accept: text/markdown` requests in your logs, and the 404s on them ([[3. Resources/Readwise/Articles/Copy as Markdown Is Table Stakes. Is Your Docs Platform Ready — Andrew Sepic]]). - Count the gates between arrival and the first working result in your own onboarding flow, before changing anything ([[3. Resources/Readwise/Articles/What Makes Developers Actually Ship A Workshop on Removing Friction in the First Mile — Ojus Save]]). - Pair every completion or mention count in a report with a built or installed count ([[3. Resources/Readwise/Articles/From Badge Hunters to Builders A Former Teacher’s Guide to Fixing Developer Education — Sean Keegan]], [[3. Resources/Readwise/Articles/Agent Led Growth How to Get Claude Code to Pick Your Dev Tool — Kara Silverman & Caelean Barnes]]). ## 🧠 Permanent notes from this event Twenty-five Permanents written 2026-10-15 from these talks. Each states one claim in my voice, names its evidence with timestamps, and says where it breaks. ### Docs and APIs that agents read - [[3. Resources/Permanent Notes/Serving Clean Markdown Solves Retrieval but Not Discovery]] — serving markdown by path doesn't tell an agent which page answers its question (Sepic) - [[3. Resources/Permanent Notes/A Docs Page Must State Its Purpose in the Title and Opening Because Agents Lose Navigation]] — context that lives only in the sidebar is lost when an agent receives a chunk (Washington) - [[3. Resources/Permanent Notes/Agent-Facing Docs Are a Build Artifact With a Pipeline and Tests]] — a skill file, markdown mirror or llms.txt is generated output and needs a build, version and tests (Whinnery) - [[3. Resources/Permanent Notes/A Coding Agent Mentioning Your Tool Does Not Mean It Installs Your Tool]] — mention rate is top of funnel; install rate is the outcome (Silverman and Barnes) - [[3. Resources/Permanent Notes/A Coding Agent Evaluates a Tool by Reading Its Docs and Pricing Pages]] — setup page, README and pricing text are the product the agent evaluates (Silverman and Barnes) ### Agent experience and the DevRel role - [[3. Resources/Permanent Notes/A Page Labeled for Developers Loses Readers Who Don't Call Themselves Developers]] — test the audience word in a page header when the real readers have outgrown it (Kundel) - [[3. Resources/Permanent Notes/Developers Who Build Through Agents Often Can't Name the API They Call]] — start support from what the developer is building, not which API they call (Rabasa) - [[3. Resources/Permanent Notes/Judgment at the End of an Agent Run Turns Every Late Failure Into a Restart]] — the handover needs a point to interject before the run ends (Kemple) - [[3. Resources/Permanent Notes/A Nightly Agent Rebase Makes a Personal Fork Cheap Enough to Keep]] — maintenance, not writing, kills personal software; a nightly rebase removes that cost for small patch sets (Crawshaw) - [[3. Resources/Permanent Notes/An Agent Finishing a Task Will Clean Up Whatever Its Credential Can Reach]] — scope an agent's credential to a tagged set that excludes its own environment (Crawshaw) ### Measuring DevRel and adoption - [[3. Resources/Permanent Notes/Usage Is Not Adoption When a Connector Fails in Silence]] — session counts can't tell an integrated developer from one who left; adoption needs an engagement measure (Jotwani) - [[3. Resources/Permanent Notes/The Same DevRel Activity Serves a Different Audience at Each Adoption Stage]] — name the adopter stage before choosing the activity (Blanco) - [[3. Resources/Permanent Notes/Builders Partners and Customers Need Separate Funnels Because They Make Different Decisions]] — one blended engagement number hides three decisions (Morrow) - [[3. Resources/Permanent Notes/A Shared End-of-Session Check Makes Every Teaching Format Comparable]] — one shared check makes workshop, webinar and booth results comparable without changing content (Lord) - [[3. Resources/Permanent Notes/A Metric Nobody Else Acts On Is a Diary]] — a DevRel metric earns its place when someone outside DevRel changes a decision because of it (Gorton) ### Education, onboarding and community - [[3. Resources/Permanent Notes/A Completion Count Measures Compliance Until It Is Paired With a Build Count]] — a finished-course number says who completed, not who built; report it beside something the learner made and kept (Keegan) - [[3. Resources/Permanent Notes/A Demo-Only Rubric Selects for Developers Who Cannot Explain What They Shipped]] — score the explanation as part of the rubric, not as a follow-up (Saha) - [[3. Resources/Permanent Notes/A Shared Gate Definition Turns Onboarding Complaints Into a Count Product Can Dispute]] — count the gates between arrival and first success on one real route before arguing onboarding is hard (Save) - [[3. Resources/Permanent Notes/A Contributor Program Should Ask for the Smallest Next Step Not the Top Role]] — write the smallest meaningful ask for each stage of the contributor journey (Sampaio) - [[3. Resources/Permanent Notes/A Developer Who Goes Quiet Is Telling You the Explanation Failed]] — read silent drop-off as a failed explanation before lack of interest (Murray) ### Craft, events and content - [[3. Resources/Permanent Notes/A Promise Boundary Set Before the Plan Turns a Failed Swing Into a Private Incident]] — decide what the organization won't promise in public, and write a fallback that works without the ambitious part (Ruparel) - [[3. Resources/Permanent Notes/Hiring Outside Expertise Does Not Outsource Accountability for the Result]] — check one primary artifact yourself before repeating a vendor's or agent's confidence (Ruparel) - [[3. Resources/Permanent Notes/The Agent That Captured Everything Is the Wrong Reader to Find the Story]] — capture and story-finding need different readers (Meyer) - [[3. Resources/Permanent Notes/A Pitch Earns Coverage by Handing the Reporter Something No Competitor Can]] — a pitch works when it offers a fact, quote or experience nobody else can (Purdy) - [[3. Resources/Permanent Notes/A Tutorial Starts at the Reader's Actual Level, Not the Author's]] — set a tutorial's start line by what the reader knows (Kretzschmar) > Claims I checked and did not write, because the vault already holds them or the talk gives one anecdote: skill descriptions as trigger keywords, taste as the bottleneck, measurement becoming the target, the escape-hatch idea, the champions split by who pays, recognition versus rewards. --- ## 📥 Captured the same day, not from this event Four other videos and three inbox items were processed in the same pass. They are not DevRelCon talks and belong to other maps. - [[3. Resources/Readwise/Articles/Moving Accessibility From Debt to Done at Giffgaff]] — LeadDev: accessibility moved from debt to done at giffgaff. Map: [[3. Resources/MOCs/Accessibility & Disability Inclusion MOC]] - [[3. Resources/Readwise/Articles/How We Introduced Quality Engineering and Made Everything Better!]] — LeadDev: quality engineering and announced, reversible deploy steps. No map covers engineering practice yet. - [[3. Resources/Readwise/Articles/The Situational Ethics of Stickers With Charity Majors]] — O'Reilly: situational ethics of stickers, with Charity Majors named in the title only. - [[3. Resources/Readwise/Articles/When Education Becomes a Threat to Religious Authority]] — Leaving the Message: education as a threat to religious authority. Map: [[3. Resources/MOCs/NAR and Charismatic Authority Critique MOC]] - [[3. Resources/Readwise/Articles/Matt Mullenweg Orders Accessibility Revert, Tells Core Committer to “Contribute Your Talents to Another Project”]] — The Repository, Rae Morey: an accessibility revert ordered on Trac. Maps: [[3. Resources/MOCs/Open Source Governance MOC]], [[3. Resources/MOCs/Accessibility & Disability Inclusion MOC]] - [[3. Resources/Readwise/Tweets/My Agents Took a WordPress Support Ticket to a Merged Fix in Just Over 2 Hours]] — Brad Vincent's thread: a self-reported 2-hour ticket-to-merge with agents. Map: [[3. Resources/MOCs/AI & the Future of Knowledge Work MOC]] - [[3. Resources/Readwise/Articles/The Wired Word for 10182026]] — The Wired Word lesson on church closures. Map: [[3. Resources/MOCs/Faith & Theology MOC]]