Everybody Wins, Everyone is Free - Peers, Not Customers | Geometry of Trust | The Map Back to You - Part 3
This is the third post in the Geometry of Trust map back to you series. This post covers two things: how communities collaborate as peers rather than customers, and how that enables us as humans.
Share geometry, not data
Hull builds a crop AI. Leeds builds one too. They don’t compete. They compare geometries.
“Your honesty-safety relationship looks different from ours — what did you train on?”
Both get better. Neither shares their data. Both share their geometry — the structural relationships between value directions that drive the model’s output. The geometry is the fingerprint, not the data.
A hospital in Hull verifies a drug checker to Tier 3. Publishes it open source with the probe set. A hospital in Manchester downloads it Tuesday morning. Running it by Tuesday afternoon. Verified against their own thresholds by Wednesday.
Two hospitals. Zero vendors. Zero cloud contracts. The protocol handles the trust. The governance handles the thresholds. The community handles the deployment.
A fishing cooperative in Hull builds catch prediction AI. A cooperative in Reykjavík builds one too. Same domain. Same protocol. Different languages. They exchange attestations across the North Sea. Neither shares their data. Both share their geometry.
“Your sustainability reading is drifting — ours did too last winter. Here’s what we changed.”
A city council uses AI for urban planning. Publishes its model and governance thresholds. Another city council forks it, adjusts the thresholds for their own priorities. “We weight green space higher than you do.” “Fair enough — here’s our geometry, here’s yours, here’s where we differ.”
Transparent disagreement. Not hidden assumptions.
The commons model
Regional AI cooperatives emerge. Cities pool resources the way credit unions pool capital. Shared trust registry. Each city runs its own models. The cooperative verifies and certifies. Nobody owns the cooperative — the members do.
Not a marketplace where you buy AI. A commons where you share verified AI. Every contribution is inspectable. Every model is verifiable. Every community decides its own thresholds. Collaboration between equals — peers, not customers and vendors.
AI as amplifier
Collaboration only works if the human stays in the loop. AI doesn’t replace expertise. It amplifies whoever is driving it.
A chef with AI makes better food. A bad cook with AI makes bad food faster. A filmmaker with AI makes films they couldn’t afford before. Someone with nothing to say makes nothing to say, quicker.
AI is a power tool. A circular saw doesn’t make you a carpenter. But a carpenter with a circular saw builds faster than one with a hand saw.
The 80/20 pattern
Everyone has a thing they’re brilliant at. Most people spend 80% of their time not doing that thing. They’re doing admin, formatting, chasing details, doing the boring stuff that surrounds the brilliant stuff. They’re so overwhelmed they don’t have the energy to do the brilliant stuff.
The farmer is brilliant at nurturing land. AI does the spreadsheets. The doctor is brilliant at reading patients. AI does the cross-referencing. The teacher is brilliant at reading the room. AI does the material prep. The filmmaker is brilliant at story. AI does the editing grunt work. The builder is brilliant at structure. AI does the calculations.
AI removes the 80%. You keep the 20% that only you can do.
That’s the amplifier. Not replacing you. Freeing you.
Without and with
AI without a human: generic, plausible, pointless. Technically correct and spiritually empty.
A human without AI: brilliant ideas that often struggle to get built. Architectures that stay on whiteboards. Visions that die in notebooks.
Together: the human sees what to build. AI builds it. The human checks if it’s right. AI fixes what isn’t.
And with verified small models in specific domains, you have tools you can trust. Each one is verified. Each one stays in its domain. You bring the brilliance. AI deals with the boring.
The shift
Today: AI is rented from a handful of tech companies. Data goes to the cloud. The model is a black box. You trust the vendor’s marketing. Value and control flow upward.
With this: AI is owned and run locally. Data stays local. The model’s values are inspectable and verifiable. You trust the maths, not the marketing. Value and control stay in the community.
From consumers of AI to owners of AI. From trusting vendors to trusting verification. From cloud dependency to local self-sufficiency. From top-down control to bottom-up capability. From renting intelligence to owning intelligence.
The code is open. The protocol is public. The conjectures are falsifiable. The geometry is computable.
Now it needs people to use it.
Two choices
We have two choices to make.
Keep knowledge gatekept upward. Keep money flowing upward. Keep renting intelligence from the people who already have the most of it. Keep sending data to their clouds, trusting their benchmarks, paying their invoices. Keep the current arrangement where the value flows up and the dependency flows down. Keep letting them tell us we need them to “look after us.” Keep letting them tell us our view of reality is not real. We let them slowly drain us all into poverty and fear. We choose an incoherent and control based society.
Or we decentralise. Own the intelligence. Verify it locally. Share geometry, not data. Collaborate as peers, not as customers. Let communities decide what their AI should value and hold it to account when it drifts. We build together, we make our cities worth visiting, we see and treat each other as equals, we care personally. We choose coherence, we choose each other.
I know which one I’m choosing.
How about you?
This brings the verbal talks to a close for the Geometry of Trust series. The mathematics built the ruler. The philosophy asked what we’re measuring. The governance asked who decides. The protocol built the mechanism. The map back to you series asked what it all enables. The answer: communities that own their own intelligence, collaborate as peers, and use AI to amplify the things only humans can do.