Intellectual Foundations

These are the ideas we build on. We're not academics, but we take ideas seriously. What follows are the core claims from bodies of work that inform how we think.

On why coordination is hard

Coase: Transaction costs explain institutions. Ronald Coase asked why firms exist if markets are efficient. His answer: using markets is expensive. You have to find counterparties, negotiate terms, monitor compliance, handle disputes. Firms emerge when these transaction costs exceed the costs of internal hierarchy.

We're building technology that lowers transaction costs for coordination. If we succeed, we shift the boundary of what can be accomplished without traditional organisational structures. Looser collaborations become viable. Smaller teams can access coordination capacity previously reserved for large institutions.

Olson: Collective action doesn't happen automatically. Mancur Olson showed that groups with shared interests often fail to act on them. Individuals free-ride; let others do the work, enjoy the benefits anyway. This is worse in large groups where individual contribution is invisible.

The implication: don't rely on shared interest alone. Small groups coordinate better than large ones. Contribution needs to be visible. Selective incentives—benefits that only contributors receive—motivate participation when collective benefit doesn't.

On how networks work

Granovetter: Weak ties transmit novel information. Your close friends know what you know. Your acquaintances bridge you to other clusters. Mark Granovetter showed that people find jobs through acquaintances more often than close friends—not because acquaintances try harder, but because they have access to information your close network lacks.

We're a weak-tie optimisation engine. Events generate weak ties. Our suggestions try to identify which weak ties are worth activating—who you met briefly but should talk to again, who bridges you to clusters you're not currently connected to.

Burt: Structural holes create opportunity. Building on Granovetter, Ronald Burt showed that people who bridge otherwise disconnected groups have disproportionate access to diverse information and opportunities. They see things others miss because they're positioned between worlds.

The platform should help users identify and bridge structural holes—not just connect to more people, but connect to different people.

On governing shared resources

Ostrom: Commons can be governed. Elinor Ostrom overturned the assumption that common resources inevitably get destroyed without privatisation or state control. She documented communities successfully managing fisheries, forests, and irrigation systems for centuries through self-organised institutions.

Her design principles for durable commons governance inform how we think about the coordination commons we're building:

1. Clear boundaries. Know who's in the community and what's being coordinated.
2. Proportionality. Those who benefit should contribute. Free-riders get excluded.
3. Collective choice. Those affected by rules participate in making them.
4. Monitoring. The community can see whether coordination is working.
5. Graduated sanctions. Rule-breakers face proportional consequences.
6. Conflict resolution. Cheap, accessible ways to resolve disputes.
7. Recognition. External authorities don't undermine self-governance.
8. Nesting. For larger systems, governance at multiple scales.

On matching and markets

Roth: Matching markets need design. Alvin Roth studies markets where you can't just buy what you want—the other side has to choose you too. Jobs, school admissions, organ donation. His work on market design identifies what makes matching work:

• Thickness. Enough participants that good matches exist.
• Low congestion. Not so many options that no one can evaluate properly.
• Safety. Participants can reveal true preferences without penalty.

These are our product problems. We need enough people in the graph that matches are good (thickness). We can't overwhelm users with suggestions (congestion). People need to be honest about what they want without fear of awkwardness (safety).

On what helps vs. what harms

Illich: Tools should enhance autonomy. Ivan Illich distinguished convivial tools—usable by anyone, for purposes they choose, without creating dependency—from industrial tools that require experts and diminish autonomy.

We want to be convivial. Minimum time on platform, maximum agency in the world. Users control purposes; we don't prescribe what they should coordinate toward. The platform should enhance coordination capacity without making users dependent on us for basic social functioning.

Scott: Beware imposed legibility. James Scott documented how schemes to improve human conditions fail when they impose simplified order on complex realities, overriding local knowledge.

We take this seriously. Our matching should learn from what users actually do, not impose what we think they should do. We should tolerate illegibility—not everything needs to be captured in the graph. When users route around our suggestions, that's information, not failure.

Desire paths. When planners lay paths and people ignore them, walking across the grass, the worn trails are called desire paths. They're revealed preferences made visible. Good design paves desire paths rather than fighting them.

Our algorithm should follow desire paths. Observe what works before prescribing. Let patterns emerge from behaviour, then reinforce them.

On what humans are for

Arendt: Action is irreducibly human. Hannah Arendt distinguished labour (survival), work (making things), and action (beginning something new among others). Action requires presence, others who witness and respond, willingness to initiate without controlling consequences.

In an age of automated execution, action is what remains for humans. Our platform creates conditions for action: connecting distinct people (plurality), enabling appearance in shared space (events), supporting the initiation of new things (projects, collaborations).

The goal isn't engagement. It's agency.