I ended my last post — twelve years ago this month — with a joke:
If the next time you see me and all my hair has turned grey and fallen out you will know why.
Some of the hairs went grey. Some of it left entirely. I want the record to show that I called it.
What I did not call was the twelve-year gap between that post and this one. I never sat down and decided to stop writing. There was just always something more urgent than the blog, and then one day I looked up and the “something more urgent” had lasted over a decade. That is how these things go. Nobody announces a hiatus. You just get busy, and the getting-busy compounds.
So here is the missing decade, roughly in order.
The Move
That last post was about leaving Philadelphia for Seattle to join Amazon. I described the logistics as “2 young kids, 2 cats, and my wife,” plus a house to sell in a market that owed me nothing.
All of it happened more or less as advertised. We landed in Seattle in the summer of 2014. The kids started pre-school. Those two young kids are teenagers now, and we had a 3rd who will be a tween this year. One of them is closer to leaving for college than to the age she was when we moved. I have no clever observation about this. It is just the single most disorienting fact of the last twelve years, and it dwarfs everything else in this post.
Seattle: Learning What Scale Actually Means
The job I moved for was Software Development Manager for the Windows and Mac desktop clients for Amazon Cloud Drive. In the 2014 post I wrote that the alluring part “wasn’t working at Amazon.com, it was working on a product that operated at the scale of Amazon.”
That turned out to be true, but not in the way I meant it. I thought scale was a number of users. Scale is actually a set of problems that only appear past a certain size, and most of them are boring. Uploads that fail on hotel Wi-Fi. Files that a customer swears they backed up. Release trains that slip because six teams share one binary.
We shipped five major versions of that desktop client. Version 1.0 was, honestly, a downloader. Then an upload client, then a background uploader, then sync, then automated backup. The automated backup release is the one I am still proud of, because it moved a metric that mattered, an 82% increase in the volume of data customers actually uploaded, somewhere north of a billion files a month. The client itself grew 60% in customers in a single year, from 328,000 to 526,000.
In late 2016 I picked up the Amazon Drive business as a whole and added the iOS and Android clients. In 2017 the web and iOS teams for Prime Photos. By 2018 I was a Senior Manager running the platform teams for Amazon Photos, mobile, web, and desktop.
Coming Back East
In 2021 we moved again, Seattle to the Washington, DC area, to Amazon’s HQ2 in Arlington.
Moving cross-country once is an adventure. Twice is a decision. We did it to be closer to family on the East Coast, which is the reason most people eventually give for going back where they came from. I do not regret Seattle for a second, and I am also very glad to be a short 3-hour car ride from Philadelphia again.
Alexa
In 2019 I moved to Alexa, and I have been there ever since.
The first stretch was Alexa Sharing, Messaging, and Announcements, about 40 engineers, working on how you connect to the people you care about through a device that has no screen and no keyboard. I wrote a three-year technical strategy for Alexa Sharing during that time. Two things happened to it that I did not expect: it survived a change of leadership over the program, and years later it became a real component of Alexa+. The underlying invention is now granted as US Patent 12,088,543.
Since January 2021 I have been working on the Alexa app, the web experience, the backend services behind Alexa endpoints, and, since 2025, the connected accessories: Echo Buds, Echo Frames, Echo Auto.
And then generative AI arrived and rearranged the furniture…
Some Things I Believe Now That I Did Not in 2014
The 2014 version of me was a Microsoft MVP who wrote a book about ASP.NET MVC and blogged mostly about frameworks. That person would find some of this unrecognizable.
Boring reliability beats clever architecture. I have never once been thanked for an elegant abstraction. I have been thanked for uploads that work.
Instrumentation is a feature. Every serious quality problem I have fixed in twelve years started with making the problem visible. Every one.
The constraint is rarely effort. When a team is missing its dates, look for the mechanism that is making it hard, not the motivation that is supposedly missing.
Growing managers is the highest-leverage thing I do. Shipping a release helps customers this month. A manager who is good at the job helps every engineer who ever reports to them.
Strategy is writing, and writing is thinking. Which brings me back around to why this post exists.
Why I’m Writing Again
Twelve years of not writing was not a decision, but it did cost something. I have spent a decade working through problems that I now only half remember the reasoning for, because I never wrote them down anywhere that outlasted the meeting.
I do not have a posting schedule to promise. I have started that promise before and broken it. What I do have is a backlog of things I have learned about scaling teams, running platforms with a lot of internal customers, and retrofitting AI into systems that were not designed for it. Some of that is worth writing down.
The blog is on Hugo now, incidentally. It started life in 2007 as Coder Journal on WordPress, spent years on Ghost, and more than three hundred posts made the trip to Hugo intact. The URLs from 2007 still work, which I am unreasonably pleased about.
Twelve years is a long time between posts. Let us make the next gap shorter.