This post captures some of my reflections on attending the 2025 Write the Docs conference in Portland. Some themes I discuss include the paradox of AI fatigue, the delight and difficulty of unconference sessions, why lightning talk formats are so challenging, and more.
In this Q&A with Fabrice Lacroix, founder of Fluid Topics, I ask him questions about his recent tcworld article in which he argues for an innovative, advanced model for Enterprise Knowledge Platforms (EKPs) acting as a central AI-powered brain for all company content, delivered via APIs. Fabrice also outlines a future where tech writers become information architects, governing vast knowledge ecosystems and coaching diverse content contributors.
In this post, I share my enthusiasm for API quick reference diagrams, which have significantly improved user comprehension and findability in our API documentation. I also explore how these diagrams serve as informative, low-token context for AI tools, enhancing their understanding of API structures (especially when accompanied with reference documentation) and how they counter hallucination in AI outputs. Finally, I end with a short tutorial on how to create these diagrams using AI.
In this post, I argue that technical writers should actively challenge ideas they find problematic, drawing inspiration from Jonathan Rauch's The Constitution of Knowledge. Rauch argues that truth arises from social debate and critique. Taking it a step further, we should listen to internal red flags or intuition when something feels amiss, even if the reasons aren't immediately clear. When direct confrontation is challenging, use open-ended, clarifying questions to investigate concerns and collaboratively explore issues.
Today was one of those days where I felt like I've seen the future. In about a day and a half, I used AI to create 8 different tree diagrams for APIs in an SDK I support. Each tree diagram has varying numbers of elements (from 50 to 350+). The tree diagrams visually depict the API structure and hierarchy, showing the data type, required/optional status, and sometimes other details. Each element links to its specific section in the reference documentation.
I recently added Document360 for API docs to my Chapter 4: OpenAPI spec and generated reference docs in my API doc site. Document360 lets you publish API reference documentation from OpenAPI specification files. Importantly, it lets you integrate this API reference content smoothly with your broader knowledge base and regular documentation, all within a single portal. The platform supports various import methods for OpenAPI files (versions 3.1, 3.0, 2.0) and Postman Collections, focusing on the publishing aspect rather than OpenAPI authoring (which I feel is a smart move to avoid unnecessary UI complexity).
This post has notes and questions for discussion for More Than Words: How to Think About Writing in the Age of AI, published in February 2025 by Jonathan Warner. Warner's book, which explores what we lose when we outsource writing to AI, is the first book in the AI Book Club: A Human in the Loop.
Jonathan Warner's book More Than Words: How to Think about Writing in the Age of AI is a spirited defense about the value and humanity of writing without AI at a time when AI promises to replace many writing activities. Warner argues that writing involves thinking and feeling, and as we grapple with ways to identify, express, and articulate our ideas in writing, it's an experience that changes who we are.
Crossings: How Road Ecology is Shaping the Future of Our Planet, by Ben Goldfarb, is a richly descriptive work of investigative journalism exploring the topic of road ecology, which looks at how roads impact their surrounding environment. More specifically, road ecology is "'the study of how 'life change[s] for plants and animals with a road and traffic nearby'" (6).
I'm starting an book club called AI Book Club: A Human in the Loop. As you might expect, the book club involves reading and discussing books about AI on a regular basis. The book club will meet online, currently planned for the third Sunday of each month at 10 am Pacific Time (Seattle's time zone). There are some async chat options through Slack as well, and we'll record the meetings. Check it out at AI Book Club: A Human in the Loop. If you've been looking to read more (or get back into reading) while also increasing your understanding of AI, this club could be a good fit.
While AGI refers to performing tasks at a human level, superintelligence refers to performing tasks that exceed human capabilities. If tech writers want to survive the AI apocalypse, we'll have to go beyond mere AGI levels of competence and tread water within the superintelligent space.Reason being, AI will eventually replace most of what we do, making it such that when AGI is reached, job displacement for tech writers will be more common because AGI will perform the same tasks, only cheaper. But the likelihood of AGI progressing to Superintelligence seems less likely to me (in the same way that moving from assisted driving to fully autonomous driving is so much harder than anyone anticipated). Striving for superintelligent docs seems like the most logical counter-move against AI's encroachment on tech writer territory.
This post is a compilation of various thoughts and responses to articles I find interesting. I decided to compile these into a single post, similar to a News and notes style rather than separating them into individual articles.
Although code samples have long been a staple in API documentation, I'm not sure users need them that much. Many developers now use AI tools that can generate the same basic code samples that are commonly provided in documentation. If these same developers pass in either the source files or reference documentation, AI tools can generate the code samples they need in the language they want, and better yet, tailored to their project and business context.
This post includes a mix of various thoughts on AI, including fixing bugs without thinking, competitive pressures to adopt AI workflows, risks of atrophied critical thinking, recursive self improvement, and the shift toward more complex tech comm tasks. There's not necessarily an argument throughline here, just various thoughts and perspectives on AI topics in my tech comm world.
Divided Highways: Building the Interstate Highways, Transforming American Life, by Tom Lewis (2013), tells the history of how the US Interstate Highway System (IHS) was built. This monumental project, initiated in 1956 under President Eisenhower, spanned decades and multiple administrations before finally reaching completion in 1992. The project encompassed over 46,000 miles of roadways, a network longer than the circumference of the Earth. The final cost, adjusted for inflation, was more than $600 billion.