Rule-Breakers

Edd Wilder-James joins us in this episode for a wide-ranging conversation that begins in the early XML and SGML communities and stretches all the way to today’s AI-driven world. Along the way, we revisit the rise of pivotal standards such as DocBook, HTML and DITA. Edd and Sean discuss the unique culture of the structured document markup conferences that brought together technologists, publishers, academics, and information architects from radically different disciplines ranging from hi-tech engineering to the liberal arts.Edd reflects on the optimism and intellectual energy of that era, including debates around standards, schemas, hypermedia, and the Semantic Web, while also acknowledging some of the orthodoxy and “tech tribalism” that shaped those communities. The discussion explores how ideas from SGML, HyTime, RDF, and structured publishing continue to resurface in modern systems, often under different names.The conversation also looks at the Big Data (Hadoop) era in the formative years of cloud computing, the evolution of open-source ecosystems at Google and Microsoft including his current work at GitHub around Copilot and large-scale agentic AI systems.Throughout the episode, Edd offers a thoughtful perspective on recurring patterns in computing: the tension between elegant models and practical systems, the repeated reinvention of timeless old ideas, and the enduring importance of context when interpreting information.

Simon St.Laurent offers an insider's perspective on the evolution of structured content and markup technologies at O'Reilly and throughout his career, emphasizing how tools like DocBook, Framemaker, and SGML laid the groundwork for today's componentized publishing models. In this podcast, Simon discusses the enduring value of structured data formats and how early practices are proving essential in the age of AI.

In this episode, Sean and John talk about 25 years of technological change, focusing on AI's impact on legislation and regulation. They highlight the shift to natural language interfaces and the need for responsible AI adoption, emphasizing data quality and human oversight. While excited about AI's potential, they caution against its unbridled use and the dangers of treating it as a human persona.