🌀 Post-Causality Cognition

Field Notes on Perception

May 22, 2025

This is part 6 in the Field Notes on Perception series, where I’m exploring how we see structure—inside data systems, inside organizations, and inside ourselves.
This one’s about causality—or rather, what happens when it stops working the way we want it to. If you’re new here, Part 1 is All Code Is Pressure, and Part 5 is Language Is the First Bug.

Here’s a strange thing to admit for someone who works in data: I no longer believe in causality the way I used to. Not because I don't think causes exist. But because our ability to trace them—cleanly, confidently, linearly—has collapsed.

We live in a world of entangled systems.

Platforms shape behavior, which shapes platforms.
Algorithms respond to reactions that were themselves trained on algorithms.
Economic, political, social, and informational feedback loops are recursive, real-time, and opaque.

And we’re expected to make decisions inside this.

In this kind of environment, information doesn’t just flow—it rebounds. It gets packaged, optimized, weaponized. News stories are tailored for outrage. Metrics are gamed for headlines. Dashboards get cleaned for the boardroom. And beneath it all, algorithms amplify whatever gets the most reaction—not necessarily what’s most true.

We don’t just have feedback loops—we have adversarial ones. Ones designed to hijack attention, distort behavior, and reinforce whatever narrative benefits the most powerful actor in the loop.

It’s not just hard to trace cause and effect—it’s often impossible to tell whether what you’re seeing is signal, countermeasure, or bait.

📊 The Fog of Data Work

And yet, even in the middle of all this, we still try to reason our way through. We retreat to the one thing that feels solid—data.

We reach for data. But data assumes clarity of definition. We build models. But models assume stable inputs. We explain outcomes. But outcomes are emergent—not engineered.

The old questions—what caused X? what should we do about Y?—begin to feel incoherent. They don’t fit the reality we’re working in anymore.

Inside most orgs, we pretend our data environments are neat decision factories: clean inputs, rational analysis, predictable outputs. But what they actually are is a mess of partial signals, social compromises, stale assumptions, and human improvisation.

We’re trying to do causal analysis on systems made of shifting metrics, unspoken incentives, and half-broken definitions. We’re playing chess in fog, and pretending we’re still on a clean board.

And the more we wrap these systems in dashboards, the more convincing the illusion becomes.

But the fog is still there. And it’s getting thicker.

🧠 From Tooling to Cognition

On some level I believe this is a tooling problem. Better lineage. Better observability. Better interfaces. But the more I sit with it, the more I think it’s also a cognition problem.

We’re trying to apply linear thinking to non-linear systems. We’re trying to isolate signal in environments where everything is signal and noise at once.

And it’s exhausting.

This is why I spend so much time thinking about metadata. Not because it's a silver bullet—but because it gives you shape when the story breaks down.

In Metadata First, I argued that metadata isn't just exhaust—it’s architecture. And in Seeing Structure, I wrote about how you can trace the pressure inside a system even when the logic doesn’t hold.

🔍 Seeing Structure

Zac

Apr 14

Early in my career, I discovered I had an unusual skill.

Read full story

When causality is fuzzy, you don’t need perfect attribution—you need context.
When data doesn’t tell a clean story, you don’t discard it—you ask:

Where did this come from?
Who touched it?
What else does it affect?

That’s why I default to structure over speculation. It’s not always elegant. But it’s traceable. And when the system is too complex to explain, traceable beats explainable every time.

🧭 A Different Kind of Knowing

So what do you do?

You stop pretending that clarity is upstream. You stop waiting for perfect attribution. You stop demanding certainty where only navigation is possible.

And you start building something else:

A new kind of judgment.
A tolerance for ambiguity.
A feel for timing, tension, resonance.

A way to move through the world with less grasping—and more attunement.