One might think computer vision models are supposed to be easy to put into production. There are whole companies built on that promise: label a few images, click train, click deploy, done. In practice, it’s messier. Most of us working with these models aren’t ML experts, and moving fast to keep up w...

I’ve been working with AI long enough to be past the “please write an epic poem about my dog” phase and into something that actually moves my work. The real shift for me, in the last couple of months, has been using tools like Claude Cowork and Cursor. It’s not because they generate better prose […]...

I’ve been approached more times than I can count by organizations on LinkedIn, all of them claiming to be impressed by my background and eager to feature my story in their publication. After doing my research, the pattern became clear. Your name gets listed in a booklet for free, but your photo, you...

Anyone who has spent time navigating downtown Raleigh, N.C., knows the routine well. You have a clear destination in mind, you know the most direct route, and with any luck, the lights will cooperate. But Raleigh is a city that never stops evolving. Road construction, lane closures, and unexpected d...

Pair programming has always had a bit of a reputation problem. It’s easy to look at two people working on the same task and see inefficiency. More coordination, more talking, less obvious forward progress. Even on teams that value it, there’s usually some quiet pressure to just go knock something ou...

There’s an assumption floating around right now that working with AI is supposed to feel effortless. When you prompt AI, you describe what you want, iterate a bit, and eventually land on something usable. Sometimes that works. More often, it leads to a long chain of almost-correct outputs, missed ed...

Recently, my software development team needed to upgrade our Expo version (it was at 52). Doing so in a monorepo containing multiple web and mobile apps was a big job—when we upgraded, we’d need to verify that the React and React Native versions in all the apps in the monorepo were compatible with t...

Recently, I was onboarded to a React Native project on a new developer laptop. I realized that the initial setup to launch the app had a few tricks and small gotchas that weren’t obvious at first. Here’s a simple breakdown that will hopefully help if you need it! Setting Up Developer Tools Before co...

I don’t write code, I don’t build models, and a year ago, I was the person Googling, “What is a prompt” while pretending I already knew the answer. Now, I give a talk called “Claude for Normies” to rooms full of professionals who feel the same way I did then. I’ve given it enough times […] The post ...

97% of executives say they’ve deployed AI in the past year. Only 29% say they’re seeing real ROI. (Both from Writer’s 2026 Enterprise AI Adoption survey — 2,400 global leaders) I’ve been sitting with that gap for weeks now — not because the numbers surprise me, but because I keep watching it play ou...

One important practice we employ at Atomic is developer rotations. This provides benefits for the longer-term success of the project and keeps our developers at the top of their game. While you might assume that long-running project teams, living with the project through its conception, implementati...

My software development team had a client who wanted to represent activity on a stage, not in the usual top-down layout that shows up all over the internet. A 3D stage? Reasonable request, slightly unreasonable visual problem. My wacky idea: what if this was mostly CSS? Maybe a little JavaScript. Bu...

A few years back, I created portable-color for adding color to shell scripts. Then I deprecated it in favor of a new library, dye, that fixed a number of things that bothered me about portable-color. dye eventually added its own built-in templating, which meant users could just print a string full o...

Previously, I showed how you can use JSDoc and jsconfig.json as a replacement for TypeScript style type checking — without a build step, no less. At the end of that post I teased an ESLint follow-up. And here we are, I’m happy you’re here. First Things First Before jumping in, I wanted to address a ...

I’ve never really been great at side questing in video games. In most games I play, I tend to lock in on the main storyline. The objective is clear, the path is defined, and progress feels measurable. Side quests exist but they get pushed off with the assumption that I’ll come back to them later. […...

Over the past 14 months on my current project with an AI startup, I’ve typed out an average novel length’s worth of meeting notes. Here’s a quick summary: Metric Value Total files (1 file per meeting) 132 Total words 77,967 Total characters 429,329 Average words per file ~590 Average characters per ...

I don’t know about most developers, but there is a very specific kind of sinking feeling you get when you join a project, clone a repo, and realize you’ve just inherited a total disaster. I’m talking about the kind of codebase where you don’t even know left from right: unhelpful generic variable nam...

In previous posts, I examined claims about “building a product in one day” and discussed my attempt to take a game from wireframe to working prototype. Here’s what happened next. By the time the game was fully built, tested, and stable, there was one step left that felt harder than anything technica...

In a previous post, I wrote about the idea of “launching a product in one day.” Today, let’s talk about how I took the initial idea and turned it into a working prototype. In the first phase of building my puzzle game, I wasn’t thinking about deployment or databases. I was thinking about mechanics, ...

If you spend any time on YouTube right now, you’ve probably seen the headlines: “I built a SaaS in 24 hours.” “AI built this app for me.” “From idea to launch overnight.” I genuinely enjoy those AI product development videos. They’re motivating. They make building feel accessible. They show what’s p...

My client team recently updated its code review policy: an approval from an AI tool is now all that’s needed to merge. In this post, I’ll reflect on this change and what it means for the way our team works together. My Personal Relationship with Code Review Reviewing PRs is one of my very favorite [...

When I joined Atomic Object nearly 16 years ago, I was drawn in by many factors: people who were generous with their time, interesting development tools and clients, personal connections, and wanting to work with the smartest people I could find. Threaded through all of that was a deep curiosity tha...

At Atomic Object, we build great software, and we partner with our clients to help them figure out what the right thing to be building is in the first place. As software consultants, we show up as strategic thought partners who bring perspective on direction, tradeoffs, and timing. Getting that righ...

When ChatGPT went mainstream, the narrative was clear: anyone can build an app now. Software engineering, as we know it, is over. My Bias I’ll be honest about my bias. I run a software consultancy with about 100 software engineers. I have skin in this game. But from my front-row seat, what I’ve watc...

Running and deploying with Kubernetes may sound excessive and intimidating, but it’s not nearly as hard as it once was. The immense popularity of Kubernetes over the past decade means tools can quickly set up and deploy a cluster in almost any environment. And what once was a technology platform req...

I have long been a fan of OpenBSD, over two and a half decades after I first started using it. It’s a complete operating system maintained with a strong and smart design ethos. It makes sense to me in a way that makes every other operating system feel frankly kind of cobbled together. Many folks […]...

AI tools have made it easier than ever to build and ship software. On one of my recent projects, I’ve seen firsthand how quickly a team can generate features, prototype ideas, and release new functionality. But I’ve also noticed something else. Just because we can build faster doesn’t mean we’re lea...

In a time of pressing adaptation to AI, it’s tempting to get to the result with fast prompts. But I’ve found that with some planning before prompting, I get better results that scale well. Here’s what I ask before I get started. What is the end goal? Am I solving a bug ticket? If it’s […] The post Q...

In previous posts, I have written about my Proxmox Backup client helper scripts and an Ansible role to provision them. The main limitation of the prior version of the scripts is that they could only specify one backup target file or directory. Now they can target multiple. Backup scripts improvement...

Picture Monday morning at a growing virtual diabetes clinic. Over the weekend, 800 patients called or messaged about refills, scheduling, and portal issues. Fifteen support reps are already behind before they log in. Your board wants higher self‑service and lower unit costs. Clinical leaders want le...