The Underappreciated Importance of Architecture

The Obvious and the Subtle

When we look at buildings, the importance of architecture is obvious. A building's beauty is directly connected to its architecture. We don't just say "that's a beautiful building," we say "wow, look at that architecture." Whether it's the old post office building in Chicago, the Duomo in Florence, or the Sydney opera house, we recognize great architecture.

But then there's subtle architectures all around us that we may not even realize. Look at a tree - there's an underlying architecture that gives it structure, fewer branches on the top, more branches towards the bottom, forming that triangular shape. A snail's shell shows that innate spiral phenomenon that nature has embedded. We don't look at these and point out the "architecture" directly, but we're drawn to their natural, unpredictable beauty inspired by the inherent design.

The Struggle in Business

In nature, architecture emerges through some mechanism or ecology we don't control. With buildings, there's an overt intention to create architecture. But in business, we often work against sound architectural principles when designing processes, information flows, software systems and more. It's right in the middle of organizations, systems, and processes where we struggle with architecture the most.

There's a huge temptation to skip upfront architectural work. We just want to get to the shiny object output - the code, the content, the deliverable. But going too fast without laying the architectural groundwork leads to chaos, just like trying to build a house without blueprints.

You can see the importance of upfront architecture even in the latest AI models. In a podcast I heard [the Neuron], the host cited architecture as the key differentiator between the quality of recent large language models. The architecture may be this extremely technical, intangible thing that only the "nerds" fully understand. But it propels everything forward.

Fighting Entropy

Maintaining good architecture is truly hard work - you're fighting against the entropy and tendency towards chaos. Individuals and organizations alike will gravitate toward little shortcuts that, unchecked, degrade the original architecture over time. Like weeds overrunning my garden if I don't constantly maintain it.

Organizations need to be purposeful about defining robust architectures upfront for key processes, systems, information flows and more. And then being vigilant about maintaining those architectures, "weeding" them regularly before they descend into chaos. You need to appoint "nerds" who understand both the technical architectures and the human tendencies working against them.

The Learning Architecture  

At Oxygen, we create many different types of architecture for our learning solutions. A full-scale learning program requires defined architectures for information design, data tracking, maintenance processes and more. Just like with software implementation - the software can't maintain itself or govern the data going into it.

Some may think AI will somehow automatically maintain these architectures. But we can't just handwave that - we need to be specific about the capabilities, build them purposefully, and integrate them into architectural maintenance systematically.

Because just like a home renovation, these learning architectures take real work to construct piece-by-piece before you get to that final, shiny output. Stakeholders often have expectations of quickly getting to that end deliverable without realizing all the unglamorous framing, demolition and installation that has to happen first.  

There's no shortcut past developing a proper learning architecture upfront, just like there's no shortcut past good instructional design. You have to do the upfront work or deal with an unmaintainable mess later.

Hooray for nerds

Architecture is one of those nerdy topics that's easy to overlook or take for granted. But architecture is crucial for creating solid, sustainable solutions - whether it's cathedral designs, a tree's branching patterns, an AI model, or a corporate learning program.  

We have to prioritize defining good architectures upfront. And be vigilant about maintaining them against the constant forces of chaos, complexity and human nature's path-of-least-resistance. It's laborious and often unseen, but that's the "nerd work" needed to avoid the shiny thing collapsing into a pile of rubble.

‍