Observation log - 29 entries

Writing.

Engineering leadership, architecture, team practices - and the occasional build log.

2026-05-22Positivity Is a Leadership Tool, Not a Personality TraitA small amount of warmth at the start of a meeting can lower social threat, improve openness, and make direct work easier. Used badly, it becomes noise.2026-03-24OpenCode Is My AI Coding Home BaseI use OpenCode because it gives me a clean terminal workflow, strong extensibility, and the freedom to switch models without rebuilding how I work.2026-03-09Otto 2/x: Building the Thing I Actually WantedA personal build update on Otto: what changed since part 1, why it is built around OpenCode, what already works, and why the architecture became much more practical.2026-01-29Otto 1/x: Vision, Boundaries, and the BuildOtto is a personal assistant that runs on your network, with explicit permissions and a sandboxed execution model. This is the vision and why it starts with trust boundaries, not hype.2026-01-23We're Building on Rented Ground (And the Lease Is Expiring)AI tools feel like magic and we're building around them. But what happens when you can't code without them, and they cost $3,000 a month?2026-01-05I built AI Agents to fix my job Search. It worked.Job titles are meaningless. 'Engineering Manager' can mean building teams or managing HVAC systems. I built five AI agents that search by role substance, not keywords. Five weeks later, I had an offer from my top choice.2025-12-28Designing for What SurvivesWe design for anticipated change and get blindsided by everything else. Residuality Theory offers a different lens: stop predicting what will change, start understanding how your system falls apart.2025-10-27Choose Boring Technology (Your Future Self Will Thank You)The tech that wins isn't the newest or most exciting. It's the one that stays running while you build features.2025-10-21The Modern Monolith 4/4: Operating the Monolith-Simplicity as StrategyThe operational story matters more than the architectural one. A well-structured monolith is straightforward to run, which means you spend time on features instead of infrastructure.2025-10-20The Modern Monolith 3/4: When to Stay, When to SplitThe modular monolith isn't a temporary solution. For most teams, it's where you should stay. But sometimes you do need to split.2025-10-18The Modern Monolith 2/4: Building a Modular MonolithStructure without distribution. Clear boundaries enforced at dev time. Here's how to build a monolith that won't trap you later.2025-10-16The Modern Monolith 1/4: Why Small Teams Should Start HereA monolith isn't spaghetti code. It's a deployment decision that makes sense for most teams. Microservices solve problems you probably don't have.2025-10-15LLMs as Semantic Middleware 3/3 - Between Humans and SystemsHow the same architectural logic that makes LLMs useful in software also applies to how people communicate and coordinate.2025-10-14LLMs as Semantic Middleware 2/3 - Where They Belong in the StackPlacing LLMs inside real systems: how semantic middleware fits between intent, execution, and feedback.2025-10-13LLMs as Semantic Middleware 1/3 - Translating IntentThinking of LLMs not as oracles but as middleware that sits between human ambiguity and system execution.2025-10-12Transparency 6/6: Measuring Transparency - Finding the BalanceTransparency isn’t a feeling; it’s a system property. You can measure how information moves, how safe people feel using it, and how fast learning travels.2025-10-11Transparency 5/6: Leadership Transparency - Sharing Context at ScaleLeaders should share almost everything. The challenge isn’t how much to reveal, but how to make radical openness useful, humane, and scalable.2025-10-10Transparency 4/6: Work Transparency - Showing Progress Without MicromanagementTransparency shouldn’t feel like surveillance. Here’s how to make work visible in ways that build trust, not control.2025-10-09Transparency 3/6: Decision Transparency - Making the Why VisibleGood decisions age well when their reasoning is visible. This is how to make the why discoverable without drowning people in paperwork.2025-10-08Transparency 2/6: Psychological Safety - The Precondition for TransparencyTransparency fails without safety. Fear turns openness into performance, while trust turns it into learning.2025-10-07Transparency 1/6: Why Transparency Matters (and Where It Breaks)Transparency isn’t a moral good; it’s a system property. What happens when organizations claim openness but operate in the dark?2025-10-04Mob Programming 3/3 - Beyond Code: Where Else Can We Mob?Exploring how mob programming applies outside of coding - design, incidents, onboarding, and more.2025-10-03Mob Programming 2/3 - Can You Mob Without a Room?Exploring how mob programming works when everyone is remote - what changes, what stays the same, and how to set it up for success.2025-10-02Mob Programming 1/3 - One Task, One Keyboard, Full Team AttentionAn introduction to mob programming as part of my learn-in-public series - how it works, why it matters, and when it pays off.2025-09-27Enterprise setups that work: Opencode with demos and hosted modelsHere is how I run CLI-first AI coding in real teams without drama. One realistic enterprise pattern, plus a simple demo setup anyone can try.2025-09-26Make AI Coding CLIs Reliable: Practices That Actually Hold Up In TeamsHere's how I run AI from the terminal day to day. It's opinionated, boring on purpose, and it keeps repos clean.2025-09-25AI Coding CLIs in 2025: A Practical GuideA hands-on guide to the best AI coding CLIs in 2025, including setup, configuration, and a practical comparison to help you choose the right tool for your workflow.2025-09-22Feedback as your teams core value: How to Build a Culture That LearnsExplore why feedback is a critical tool for growth, how most feedback cultures fail, and a step-by-step guide to building a culture of continuous improvement.2025-09-01Domain Thinking: Your Superpower in the AI EraAI has undeniably changed the day-to-day of software engineering - but not always in the ways the headlines suggest.

Let's build something that lasts.

engineering leadership · system architecture · or just a chat