I talk often about Conway’s Law, both here and in real life. I also talk often about working in the public sector. But for some reason I’ve never mentally put the two topics together and drawn the inevitable conclusion discussed in this article: Conway’s Law at Government Scale.
I’ve read Recoding America, and generally agree with the notion that a product operating model makes sense and leads to better outcomes. But here’s the thing: change comes slowly, if it comes at all, and solutions are needed now, across many domains.
I’m grateful for the work of those whose role is to reorganize and rethink government, and wish them success. But in the meantime, I see my role as working within the structures that exist now and doing the best that can be done. Projects can succeed, even with constraints.
I’ve had space on the brain recently, having in the last month visited both the Jet Propulsion Laboratory and the Johnson Space Center. There’s a paradox of sorts when thinking about interstellar travel called the wait calculation (or, in simpler form, the wait/walk dilemma). The gist is that we should not launch a slow spacecraft now because one sent later with a faster propulsion system would simply overtake it. Repeat that argument ad infinitum, and you’ll never launch, hence the paradox.
These days I’m feeling caught in a similar sort of trap when it comes to learning about AI, with announcements almost daily (just this week, GPT-5 launched, right on the heels of the release of a bunch of open weight models that can run all sorts of places, including Amazon Bedrock). Just when I think I’ve identified the technology I want to really embrace, new ones arrive that create new capabilities, deprecate old ones, and demand rethinking workflows. It’s disorienting.
You know what isn’t disorienting? Photos of cool command centers. Here’s three of them from my recent travels:
The room where it happened
The room where it’s been happening a while
The room where it’s still happening
I’m a sucker for a goodcommandcenter, that’s for sure. But enough distractions; I know I just need to dig into deeper AI learning. That’s the trick: just start.
Dropdowns in web forms are generally good; they make it simpler for users to input options correctly and ensure back-end data integrity. They can be limiting at times, though, so I have to respect an event registration website I used yesterday that tried to be all-encompassing in the selections for “Title”:
Henceforth I expect to be addressed as “Lord Neer”
I have to imagine this list came from some out-of-the-box form generation tool. Or created by GenAI, perhaps? I’m curious what it could have been. And was it not modifiable? Suffice it to say several of the choices are fairly pretentious given the event I was buying tickets for, so I feel like maybe the developer should have looked into culling the list.
Technologists love to collect and share horror stories (see, for example, The Daily WTF). It’s one of the reasons this blog exists, as a brag document of a different sort.
No matter how stressful the situation may be, no matter how long the debugging session took, no matter how brain-melting the eventual solution was to implement, on those days when you experience a moment that you know will go into the annals of “how the heck did this happen” infamy, it brings a smile to your face.
For me, yesterday was one of those days.
It’s a little too early to tell the full story in a public place (the key stakeholders should get to hear it first), but I hope to eventually. It’s an all-time head-scratcher.
Despite having had a number of opportunities to do so throughout my career, I’ve never progressed beyond being an average public speaker.
Thankfully I don’t have any particular phobias about it, and I can do a decent job relaying facts while being mildly interesting, but I’m far from a great orator, especially when I have to speak on the fly.
Still, every once in a while I’m happy with my ad libbing. This past week I spoke at a conference, and came up with this turn of phrase that I quite liked:
Universal problems are often best-solved through many local partnerships.
Perhaps that’s why I enjoy building for state government so much?
Sitting on a late flight to New York City last night, I spent a few minutes time rereading my previous writing on radical responsiveness (yes, I do this sometimes). In the former post I said the following (and yes, it’s absolutely self-indulgent to quote myself, but here we go):
Being known as a responsive person 95% of the time usually means others will assume the best of you for the 5% of time you fail.
That ratio got me thinking: at what response rate will others start losing faith that you’re a responsive person, and thus begin not giving you the benefit of the doubt? It’s gotta be higher than 50%, because I can’t imagine thinking a person who’s likelihood to respond is no better than a coin flip could be viewed as a reliable responder. Maybe 70% or so? I bet a plot of actual response rate against fraction of people who will perceive said rate as responsive would look something like this:
The lesson: earning trust in responsiveness is hard, and keeping it is even harder!
Just when you thought there couldn’t be more (oh trust me, there’s more), here’s comes another round of my series on computer setups (earlier posts are here and here).
Years: 2005-2010 Machine: The box in the closet from my last post but with a snazzy new LCD monitor (my first flat panel), wireless peripherals, and my wife’s great-grandmother’s writing desk. What I was doing: writing daily on Xanga; applying to family camp; traveling to Tennessee to watch Revenge of the Sith with a high school friend; hosting LAN parties for Age of Empires III; traveling a lot for work.
Years: 2006-2007 Machines: A plethora of cast-off parts coalesced into a couple functional boxes in the garage. What I was doing: running a NAS for storing all my media; trying to get Gentoo Linux to compile; realizing that running a bare web server on the Internet is asking for trouble.
Years: 2008-2010 Machines: A beefed up HTPC rig in a rackmount case with a whole bunch of amps to power my Magnepan speaker system, plus the plethora of beige boxes from before, but better arranged. What I was doing: shivering in the garage when using this desk in the winter, listening to Comfortably Numb at peak volume, hosting movie-watching parties.
Year: 2011 Machines: Same rack mounted setup, but relocated from our old garage in Ohio to a closet in our new garage in San Diego, plus a random beige box (those just won’t go away). What I was doing: Playing with some audio recording gear; streaming Game of Thrones; stocking up on printer paper, apparently.
Years: 2012-2014 Machines: Moved inside, and rebuilt the guts from the rack mount case into a traditional one (albeit black and silenced), also got my first Mac laptop (from which I’ve never looked back). What I was doing: writing election software; creating hand-made guitar effect pedals (that were terrible); developing back problems thanks to a crummy chair.
Yesterday I re-passed the AWS Certified Solutions Architect – Professional exam. It’s my third time taking it (2019, 2022, and now 2025). I went into it cold turkey, absolutely zero preparation, and while my score wasn’t as good as before, a pass is a pass.
I guess at this point, I’ve done enough work in AWS that the knowledge is pretty much permanently ingrained. It reminded me of this article on Buying Moves:
There’s an amazing quirk of the human body that makes fitness adaptations work a bit like powerups in a video game. You can spend some resource to add capabilities to your “character,” and if you do it right they’re always available to you (for the most part). I’m talking about muscle memory and our ability to “lock in” physical skills.
Perhaps there’s a mental version of the same phenomena?
Continuing from earlier, this post covers early adulthood, pre-kids.
Year: 2000 Machine: Custom built beige box with NewEgg parts What I was doing: Enjoying early married life; playing Deus Ex; renting movies from Blockbuster and watching them on my computer’s DVD-ROM drive; organizing documents in that green file box (which I still have).
Year: 2001 Machine: Same beige box but with new guts What I was doing: starting to explore high-end audio and home theater, applying to Ph.D. programs, anticipating The Fellowship of the Ring.
Years: 2002 Machine: Still the same beige box but a new “flat” monitor What I was doing: trying to understand Topology, skipping graduate school classes to play Diablo II; anticipating The Two Towers (I still remember waiting for this trailer to download… it was was mind-blowing, despite the low quality of the day).
Year: 2003 Machine: You guessed it, a beige box, though I believe I’d upgraded it yet again What I was doing: starting a blog on Xanga, anticipating The Return of the King (it was on this setup that I created a postcard inviting 40 of my closest friends to a private midnight screening at the Arena Grand).
Year: 2004 Machine: A second beige box filled with scraped together cast-offs from my main computer, complete with double optical drives What I was doing: embracing my minimalist tendencies by trying to hide boxy components inside closets and running cables through the walls; ripping my entire CD collection to FLAC; preparing for our first kid.
