A collection of thoughts, a digital garden if you will, about agentic development.
Who am I?
I’m a harness engineering practitioner. I focus my efforts on migrating my own day-to-day away from traditional software development and into the new era of agentic development. I make agents work for me in situations where it is difficult.
I’ve been in the industry for about 25 years. I’ve been acting technical lead, team lead, architect, principal engineer, and consultant. I’ve built greenfield projects and maintained brownfield projects.
My focus is application architecture within SaaS platforms. I like to write software that is easy to change.
My favorite language is TypeScript, though I am eager to embrace modern languages like Golang and Rust.
- Professional LinkedIn
- Retired Twitch Partner & Factorio speedrunner
- Currently employed as a Senior Staff SE at Direct Supply
What is this, really?
I wanted to construct a living, ever-changing digital garden that shares my understanding of the industry as it changes.
It’s not designed like a traditional blog post, it’s designed to be skimmable, it’s designed to allow you to take what you want, quickly, and leave the rest for another day. From time to time, it’ll be fuel for an essay.
How this writing affects me
Writing helps me articulate and better understand the concepts that are on my mind. I have an obsession with developer experience and AI assisted workflows. I think about them a lot throughout the day. It’s not easy to think about something else. Writing helps. Once I articulate my thoughts, my brain no longer wants to ruminate on them to refine them and better understand them. It frees me to focus on other areas of my life, like my family, friends, or what is for dinner.
Writing this also helps me learn a lot about myself and how I come across. I write what I think. After that, I can use LLMs to get constructive feedback on that writing. I’ve learned a lot about my tone and how I state things. It’s helping me become more balanced and humble about my shortcomings. We have to write out what we think to get insights from an LLM on how we can improve our thoughts.
Harness engineering
Philosophy
How I guide my writing and why I choose to view harness engineering in this way. These are not truths, nor tightly held. My own philosophies have changed since 2025. As they change, I’ll update them here.
Theory / concepts
Insights and thoughts on the theory of harness engineering that can shape the way we build.
- Agent-managed tooling
- Agent verification
- testing loop
- Application architect
- Application architects and harness engineering
- Application, harness, workflow, & workspace
- Average content of the internet
- Digital garden
- Exemplars
- Forward deployed engineering
- Harness components
- Harness engineer
- High level thinking drift
- Iterative agents
- Iterative planning
- Lift
- Local ephemeral environments for agents
- Overloaded terms
- Progressive disclosure
- Readability = skimmability
- Team of the new era
- Traditional feedback loops
- Walk, run, wheels, rocket into space!
Application architecture & harness engineering
How the quality of your codebase affects how LLMs work within it.
Practical
Applying harness engineering in a non-idyllic environment. Harness engineering is not just for greenfield applications.