Thinking in public on intelligence, scale, and the hard parts.

Notes on AI-native platforms, regulated markets, scaling engineering orgs, and operational reliability.

Ask the archive

All writing

  1. Altdown: RSS for agents* New Reader mode for agents, made by the publisher. AI-native
  2. Plan before you prompt How to steer coding agents with four small documents (architecture, ADR, plan, tasks) so your team makes the design decisions instead of the agent. AI-native
  3. Guardrails for AI-generated code A four-layer guardrails model for AI-generated code (scope, structural, behavioral, human verification), plus notes on a shrinking architectural half-life. AI-native · Reliability
  4. Death to Scrum? If the process of software development is changing this dramatically, what happens to Scrum? Scaling orgs · AI-native
  5. The New SDLC AI is compressing the development cycle and changing how product, engineering, design, QA, and security teams operate. AI-native · Scaling orgs
  6. From Reading Changelogs to Finding Zero-Days: LLMs and the Future of AppSec The old model of security patches cannot survive contact with systems that operate at machine speed. AI-native · Regulated
  7. Spec First, Then Code LLMs have no context. You have to give it to them. AI-native
  8. Something Has Changed With AI Coding For most of the past year my view on AI coding tools was old-fashioned. Something has changed. AI-native
  9. OAuth Is Not Enough: Authorization Challenges for Autonomous AI Agents OAuth gives us delegated access, but by itself it can't ensure that autonomous agents stay inside the guardrails. AI-native · Regulated
  10. Managing by Metrics How to select, use, and think about engineering metrics. The difference between good metrics and the ones that just make you feel good. Scaling orgs · Reliability
  11. Engineering Culture and Principles Culture is the default way a team operates when no one is watching. Scaling orgs
  12. Giving Negative Feedback Most managers either put off negative feedback or reach for the sandwich method. Scaling orgs
  13. Parsing Text Files in Shell I spent years reaching for awk before landing on something simpler, built right into how the shell reads data. Reliability
  14. 9 Years of Scrum, and All I Got Is a Lousy T-shirt (Sizing) Nine years of scrum, with varying degrees of rigor. Scaling orgs