Founder / builder / Wellington

I build software companies for the long game.

I started coding at ten and never quite stopped. What changed was the size of the system: from programs, to products, to companies built to endure. Today that means Raygun and Autohive—where product taste, engineering discipline, and clear thinking meet.

Software since childhood Long-term companies Systems over theatre
Read the writing Full archive
50B+ errors processed for Raygun customers
100,000+ developers using Raygun across 120+ countries
Two decades founding and building global technology businesses

The recurring questions

The work I keep coming back to

Software quality

I care about the full chain between code, telemetry, diagnosis, and customer trust. The interesting work is rarely just finding the bug; it’s shortening the path from signal to resolution.

AI systems for real teams

Useful AI is orchestration, context, and judgment, not just model output. I’m interested in building systems teams can trust enough to use in their actual operating rhythm.

Operating leverage

The systems underneath the company matter: decision quality, engineering culture, product taste, and the loops that let a team move faster without getting sloppier.

How I work

The computer is a tool. The point is what it lets people build.

I’ve learned to value uninterrupted thinking as much as frantic output. The best leverage usually comes from a clearer decision, a shorter feedback loop, or a system that lets the team do better work without heroics.

Read the original note

Current work

Two companies, one bias toward useful systems

Raygun Software intelligence

Monitoring, diagnostics, and AI-assisted debugging for serious software teams.

Raygun helps thousands of teams gain real-time visibility into errors, crashes, and performance issues before customers feel the impact. The platform spans Crash Reporting, Real User Monitoring, APM, and the newer AI Error Resolution workflow.

  • Processes more than 2.5 billion monthly errors for brands including Domino’s, Electrolux, Virgin Atlantic, and iFit.
  • AI Error Resolution launched in April 2024 and won Innovation Gold at the 2024 Wellington Gold Awards.
  • Built for developers, tech leads, and product teams who want fast diagnosis without losing customer context.
Visit Raygun See AI Error Resolution
Autohive Collaborative AI

AI teammates, workflows, and automation built for everyday teams.

Autohive is designed to make AI agents practical: describe a repetitive task, create an agent in seconds, collaborate with it, and automate the routine work without needing to code.

  • Supports use cases across 100+ industries, from marketing and analytics to customer support and operations.
  • Combines agents, shared workspace context, scheduled jobs, workflows, and integrations in one system.
  • Security is a first-class feature: data is never used for AI training, with enterprise-grade controls, CASA certification, and a zero-knowledge architecture approach.
Visit Autohive See security details

Writing

Notes, experiments, and opinions—mostly learned the expensive way

See the full archive

05 Jan 2017

Starting with Hugo

I just setup Hugo on my pc. Update: I’m working on getting this automated with S3, hopefully then will update more.

Read article