People as Code 4 - Entropic Systems

The first three chapters explored metaphor, but this one, this one is about practice.

If People as Code asked whether people could be co-authors of systems, and People as Code 3 - Quantum Uncertainty reminded us that all systems are in flux, then this chapter asks:  What could we build, now, if we took those ideas seriously?

What would technology, governance, and funding look like if we designed not for control, but for entropic stability - systems that stay alive because they move?

Designing for entropy

Entropy isn’t chaos; it’s diversity, motion, difference. A living system needs it.

Our current models of tech, of governance, of funding do the opposite. They optimise. They converge. They narrow the field of possible futures until only one remains: the most efficient one or the least risky one. 

What if we built systems that got healthier as they diversified? Where friction, noise, and drift were design goals, not bugs to eliminate.

Where experiments didn’t have to justify themselves by “working,” but by expanding what was possible to know.

Observability, not measurement

We’ve learned that measurement collapses the system. It defines what counts, and everything else falls away. So instead of measurement, what if we build for observability, making it easy to see, sense, narrate what’s happening, without forcing everything into metrics too soon.

Measurement is passive, but observation is participatory. It allows patterns to emerge before they are judged and gives room for people to make sense of what they see, collectively.

Imagine social dashboards built around questions, not KPIs.

• What’s changing?
• What are we noticing?
• What do we not yet understand?

Data as Conversations, not verdict.

Collective foundations

If the last 4 years of AI has taught us anything, it’s that foundational models matter. They become the substrate for everything else that follows.

But the question is, whose foundation?

Imagine if social purpose organisations built their own models. Trained not just on language, dark corners of the internet, stolen creativity and hate filled socials; but on lived experience, community archives, oral histories. And what if these models were governed by collectives, not corporations. 

A shared model doesn’t mean uniformity. It means we decide together what’s “real enough” to train on. We define consent, representation, and use together.

This isn’t open data in the old sense; it’s collective data sovereignty. A foundation not of code, but of trust.

Governance as flow

In most organisations, governance is static.  Structures, committees, thresholds built to hold still.

But what if governance moved? What if it evolved as quickly as the world it seeks to guide?

Imagine:

• Priorities and metrics that evolve over time, assumed to be fluid.
• Ongoing consent, revisited as relationships change.
• Decisions made at the edge, where context lives.
• Governance as a conversation, not a constitution.
  

Data as plural

We’ve built data systems that value consistency over truth, but in reality, truth is often contradictory.

What if we treated different forms of data,  stories, images, feelings, local knowledge as legitimate inputs? Not to “validate” them, but to let them coexist.

Because in complex systems, contradiction is signal, not noise. Plural data creates plural futures.

Questions as interfaces

Interfaces shape attention. Right now, most are built for ‘certainty’, built for answers.

But what if the interface began with questions? “What do you want to learn?”  “What tension are you sitting with?”

Each question would open a pathway through the data, not to deliver certainty, but to show the plurality of possible answers.

The goal wouldn’t be consensus, but comprehension: seeing how different perspectives entangle.

Beyond technology

This isn’t just about software. It’s about how we fund, organise, and govern.

What if funding worked like quantum probability — multiple possibilities coexisting until observed? Instead of one “winning” proposal, fund several small, divergent paths. Let them interact, combine, and evolve.

What if governance of an ecosystem,  a network of organisations, was distributed like a neural net? No single node in charge, but pattern recognition across many.

What if we replaced “impact frameworks” with learning frameworks, systems that reward adaptation, not perfection?

A living architecture

We could start small.

• A fund that tracks learning rather than outputs.
• A data commons governed by those it represents.
• An AI model trained on stories, consented to by communities.
• A policy lab that treats uncertainty as design material.

Each one would be a node in a living architecture, a prototype for how we might build, together, in a world that refuses to stay still.