“Always code as if the guy who ends up maintaining your code will be a violent psychopath who knows where you live.”
— Martin Golding

‍

‍

A few weeks ago, I stepped back nearly two thousand years on a private tour of Roman London. We marveled at the remains of the Roman wall—more of a flex than an actual defensive structure!—and the high-tech formulation of the cement, which contained hot-mixed quicklime that would self-heal cracks when it rained. But after the Romans left in 410CE, what took over 300 years to build fell apart in a few decades.

‍

Maintenance is continuous, while collapse is abrupt and unpredictable. Left to themselves, systems degrade; entropy is always at work, pushing toward disorder (and once in a while, the demise of a civilization).

‍

In IT, this is not entirely a technology problem, it’s also a human problem. This is understandable: careers are often built on “game-changing” disruption, not on dogged diligence. Ambitious young technologists chase the moonshot team with its bright, shiny objects, while the teams that keep the lights on—and the planes flying, and the markets moving—are the ones doing the work that fundamentally matters.

‍

An ideal organization would recognize the importance of a continuous modernization mindset and set incentives accordingly. Alas, most organizations aren’t ideal, and so they’re staring down a hard reset to correct years of neglect, often cunningly disguised as a “transformation project.”

‍

And if the can has been kicked down the road long enough, this realization reveals a painful truth: how do you know what to build when you don’t know what’s already there? The people who built the system are long gone. The documentation is nonexistent. Rome has left the island.

‍

That makes modernization an act of archaeology, digging through layers of legacy code, dependencies, and surprising integrations. But it’s also anthropology. Systems are the living records of the people who built and maintained them, shaped by incentives and institutional priorities. That’s the incidental complexity that sinks so many modernization efforts.

‍

So, what to do? Three considerations:

1. Take a behavior-first approach to modernization efforts. Start by observing what the system actually does in practice, not just what the code or documentation claims. Map user behaviors and workflows to uncover hidden dependencies.

‍

2. Adjust incentives to recognize those who continually reduce complexity and improve resilience. Bake “the cost of change” into TCO calculations when issuing RFPs for modernization projects: assign a dollar value so it sticks.

‍

3. Activate continuous modernization as an ongoing state of being rather than a point-in-time, risky project. Embed modernization into the operating rhythm of the company, just like compliance and patching are today, by orchestrating between old and new environments.

‍

Like a thriving Roman city, maintenance isn’t drudgery; it’s the engine that perpetuates progress, strategy in its purest form. It’s what separates the organizations that sleepwalk into big, hairy technology catastrophes from the ones that earn the right to innovate. It’s precisely this quiet, constant work that makes the extraordinary possible. Respect.

‍

News and Views

Speaking of eschewing bright, shiny objects, two Princeton professors published a counter to AI hype, “AI as Normal Technology.” Their argument is that the challenge isn’t the speed of invention or innovation, it’s diffusion—the speed of human, organizational, and institutional change.

‍

Operational structure impacts your attack surface and threat landscape. The Justice Department is continuing to find this out the hard way, despite warnings. Reminds me of the immortal words of Anthony Weldon, a 17th century British politician: “Fool me once, shame on you; fool me twice, shame on me.”

‍

As if we need any more reminding that digital transformation is as much a cultural issue as a technological one, CIO Magazine writes yet another article about it. What caught my attention, though, is calling out that CIOs need to manage peoples’ fears.

‍

‍

From the Orchard

At the Gartner Application Innovation & Business Solutions Summit in EMEA last week, it was encouraging to see the keynote call out “rethink the way we modernize core systems” as one of three essential prerequisites to prepare for the AI era. Listen to a recording of Edward Hieatt’s presentation about “in production” as a better metric of modernization than lines of code transpiled.

‍

Fun stuff: if an ASCII-based game and AI got together, what could it look like? Jeff Schomay ran a few experiments on rendering a game in real time with AI.

‍

We’re launching a series of networking events around the perils and rewards of transformation with our partner, Thoughtworks. Called “Unfinished Business,” our inaugural event is a breakfast at the Rosewood in London on Tuesday, October 28th. Spots are very limited; reserve yours here.

‍

Curious to learn more? Say hello@mechanical-orchard.com.

‍

—

*Issue first published on September 30th, 2025. View all newsletters here

‍