Start With Why

💡 “It’s got to start with a project, I think, in many ways. That’s what this site is. It’s a project, an experiment, a way for me to build in the open, exploring more deeply what AI can do, where it works, where it doesn’t. It is all over the place right now… …many people are frustrated with it but see the potential, and that’s where I am.”

This is Dev Diary #0 where I’m going to take a step back because you always want to start with why. And this is the “why” memo.

I think in order to answer that question, you need to step back a little bit. When I was laid off, I found myself a little bit rudderless—I don’t think that’s an uncommon situation. I got to a point where you ask yourself: what’s next? I’m not beholden to anyone in this moment, at least from my career standpoint. I have the opportunity to pause a bit and not just reflect, but really dig a little deeper into what comes next and what could be an interesting next gig.

In many scenarios—almost all the time I’ve been between jobs—I’ve left a gig and told myself I would spend a fair amount of time really digging into what I wanted to do next, but never really did because I always got pulled into something. I’d talk to one or two people and realize there was something interesting, and it was immediately off to the races. That’s worked well—I’m not complaining about it—but I think there’s something to be said about really being deliberate, doing a deeper dive into what could be a really good opportunity, something really energizing that speaks to me on a deep level.

That’s coupled with this world we live in now as software engineers that is changing—it feels like every hour. They always say that tech moves fast, but in my 20 or so years doing this stuff, it has been nothing like it. I’m in a scenario where I need to be moving very quickly but also making good decisions, which is the tricky part. It’s also incredibly exciting. The future seems to be hyper coalescing before our eyes, how can one not be stoked?

So you look forward by looking backwards. In the past, what roles really gave me the most energy? It’s always been working as part of team of engineers jamming on cool software, where I contributed by making the engineer next me more productive. First job in startups was as a server engineer—building systems for my teammates to deliver on. I see it as the same for leadership roles. When you’re an EM what you’re doing is creating a culture for your engineers to deliver at a world class level. Same impacts with different modalities.

DevEx is something that has always really spoken to me. How do you create a development experience that really does multiply the value that organizations and teams can deliver? AI is now turbocharging all that. Or at least, that’s the promise. Everything’s so chaotic right now, quantum states that only resolve when we look at them. So I’m looking at them with this project, this website, this “blog”.

I feel like a lot of the utility right now, at least from my perspective, is focused mainly on the solo single dev scale. The individual developer starts from zero, they learn how to move through the tooling on their own. Even if there is documentation, it’s very light, mostly non-existent. A friend of mine described it as “everyone just starts from zero.” They need to do so much just to get to cruising altitude. So how do you solve that, building teams that have a head start? That’s the core of what’s driving me right now.

It’s got to start with a project, I think, in many ways. That’s what this site is. It’s a project, an experiment, a way for me to build in the open, exploring more deeply what AI can do, where it works, where it doesn’t. It is all over the place right now. It almost depends on who you ask whether or not they are excited about it. Everyone’s got some experience—many people are frustrated with it but see the potential, and that’s where I am.

So in many ways, it’s just a weekend personal hackathon project where I explore some tech I’ve been interested in for a while (hugo site gen; fly.io). Where things get interesting for the project is how the content is created, as well as how the infrastructure glue comes together.

The longer form posts will always start as voice memos, which then get fed into Claude for cleanup, with light editing for final publication. The prompt directs for minimal cleanup to keep my narrative voice—not my actual voice, though maybe I’ll play those sometimes too. Really just taking the ideas, dumping them into the voice memo, lowering the friction of going from idea to blog post. You don’t want it to be sterile. AIs have their own voice, so the trick is saying “clean this up” in a way that lets me still come through even as the AI translates the transcript.

It’s all pretty manual right now. So where the development comes in is to build features into the codebase that will make this loop—(voice → LLM → edit → publish)—much more automated, much tighter. So that’s the project, just me and Claude (for now), building something real in the open, and building understanding of how to use—and how not to use—AI to augment the whole thing through a little elbow grease.

Then finally on top of all that, as I start to look for my next job, I’m hopeful that people will find value in this kind of working in the open. I’ve always been a fan of that idea. It’s important to allow people to see your work—you can’t always do that when you work for a company and your work is not public. Having something I can point people to might give them a better understanding of how I do things and maybe help them get a head start on deciding if there’s a fit.