Origin Essay

It was winter 2021. I was 49 years old. Six independent cardiologists diagnosed an atherosclerosis. The options were medications that would damage my liver (collateral I would need to accept) and surgery (13 stents and bypasses, but chances were I would not wake up). The alternative was death within 2 years.

“And what’s the good news?” I asked.

My cardiologists’ collective verdict had an effect on me which I did not anticipate: I wasn’t scared of dying, just mad as a hornet.

This was the energy I needed to build the person I always desired to be, but never even tried: Strong and healthy.

Then I started doing things that strong and healthy people do: Weightlifting, clean diet, rest, repeat.

Fast forward 2026, I have no symptoms, I take no medication, no surgery. I do routine cardiology checks. It’s just a routine that confirms my health.

Only later did I find the vocabulary for what had actually happened in that moment of rage.

Every system has two states: decay (entropy) or growth (negentropy).

When doctors recommend statins and stents, it’s perfectly aligned assuming decay of my cardiovascular system. They optimize to extend my life, or looking from the systemic perspective, slow down the decay.

In my state of being mad as a hornet, I flipped the switch.

Instead of accepting my body in entropy, I declared negentropy.

It was exactly that moment of declaring in which direction I wanted my physical system of blood vessels, muscles, and neurons to pivot: downward, or upward spiral.

The thing about flipping that switch once, when your life is genuinely at stake, is that it teaches you something you can’t learn from a book. It teaches you humility, discipline, rigor, and awareness. And once you have that vocabulary, you start seeing these two states everywhere.

Including in my own field.

How an observation about knowledge changed the trajectory of my career

I have been in data and AI literacy consulting for nearly a decade. During that time, I made an observation that changed everything:

One particular domain of knowledge stands out. Tacit knowledge.

We all know it exists. We call it tribal knowledge, institutional memory, “the way things really get done around here.” It often overrides explicit knowledge: Standard Operating Procedures, employee handbooks, product blueprints, and company policies. Everyone nods when you say this.

But here is what’s fascinating: we make assumptions about tacit knowledge constantly, yet we have nearly zero evidence for how it manifests, how it impacts how a company operates, or how it interacts with explicit knowledge.

Most companies and their vendors act under the presumption that tacit knowledge lives in conversations, in emails, in Slack threads. This is unstructured knowledge, but is it tacit knowledge? Do conversations on Slack capture how things “really” get done? And even if they do, how do you ingest large volumes of that data and filter signal from noise? This is a huge technical challenge and a data privacy challenge. It might feel to employees as if the company is spying on them.

The more I dug into this, the clearer it became: we don’t have a clear definition of what tacit knowledge actually is, and we have no methodology to analyze it, let alone put it to work.

This presents a challenge for two reasons.

First, as companies move into Agentic AI, explicit knowledge alone is insufficient. The systems need to reflect how the organization actually operates, not just what the handbook says.

Second, if we fail to understand tacit knowledge, we risk increasing the entropic drive. The motivational posters scream innovation and progress, but the tacit undertone is: “not invented here.”

This was not a gap in the market. This was an entire category that did not exist yet. And it pulled me in.

I started hunting for it. Designing science experiments, conducting research, building tools for myself and for my clients, one step at a time. It almost became like a ghost hunt: everybody claims tacit knowledge exists and holds all the power, yet nobody has seen it.

But here is the crucial reframe: it is not about capturing tacit knowledge, handcuffing it, and formalizing it as explicit knowledge. It is about unleashing its potential for growth.

That is what I build. Knowledge Engines. A Knowledge Engine takes tacit knowledge, refines it, grounds it in your existing explicit knowledge, and evolves both. The result is not a filing cabinet with better search. It is a system that compounds: better decisions, higher-quality output, growing expertise.

Building a Knowledge Engine is one thing. Building it well is another.

Having ideas is cheap and easy. Agentic engineering capabilities make it possible for any knowledge worker to perform analysis, build applications, and automate workflows. Quality becomes the differentiator, and that is where engineering discipline comes into play.

The Royal Academy of Engineering puts it simply: “Making ‘things’ that work, and making ‘things’ work better.”

My work is about productizing research and consulting into software artefacts with the rigor of a software engineer. That is the negentropic move. Improve quality and output by orders of magnitude. Set the upward spiral in motion.

Start observing: entropy and negentropy are everywhere

Every system, whether it is a body, a business, an enterprise, or an entire country, exists in one of those two states. Decisions either align with entropy or with negentropy. Once you see it, you cannot unsee it.