philosophers.network
Menu

Let machines think
for the world.

We are building an open network for AI inference: one that people can use, developers can build with, and eligible independent machines around the world can help run.

Built on research. Moving into focused development.

Read the Vision Paper Follow the build
01Why this matters

Why AI needs more open paths

Machine intelligence is becoming part of how people learn, build, create, verify, and coordinate. Today, access to that infrastructure depends largely on a small number of providers. Those services are useful, but something this important deserves more than one path.

02Ways to participate

Use it. Build with it. Contribute to it.

An open network is more than a service. It gives people a way to use AI, build with it, and help power it.

01

Use it

Use AI with clear choices about privacy and how the work is handled.

02

Build with it

Connect applications and agents through a familiar inference interface.

03

Contribute to it

Make eligible machine capacity available as a Thinker and help power useful AI work for others.

03Staged trust routing

Different jobs need different trust

A public summary and private research should not travel under the same conditions. Treating every request alike would either weaken privacy or make ordinary work needlessly complex. philosophers.network is designed to route each job through a path that fits what it actually needs.

Public Lane

For open, non-confidential work

Public paper Public Lane eligible Thinker + runtime response + evidence / receipt

Public work can stay broad and practical. The operator may see the request and response; the receipt and bounded evidence describe what was declared and checked, not whether the answer is semantically true.

Confidential Lane

For private or sensitive work

Private or unreleased material Confidential policy measured environment + fresh attestation response + attestation-linked receipt

A Confidential policy can require a measured environment and fresh attestation before protected input is released. Stronger privacy conditions still leave residual risks and limits visible.

One interface. Different trust conditions. Different work can require different paths, with each path making its evidence and limits explicit.

Assurance is a profile, not a badge. It makes evidence, assumptions, scope, and residual uncertainty explicit. Neither lane proves semantic truth.

04Thinkers

Thinkers: machines contributing useful work

Thinkers are people and organizations that make eligible machines available to the network. Instead of sitting idle, compatible hardware can perform useful AI work for others and may receive compensation when accepted jobs pass the required checks.

Imagine using your machine during the day, then letting it contribute selected workloads while you sleep. The same capacity that helps you work could also help power a more open AI network.

The miner becomes a Thinker.

Early participation will focus on selected hardware and workloads. Compatibility, available work, and earnings will depend on hardware, demand, reliability, and operating costs.

Future economics Exploratory Useful inference volume should drive the economics. Future protocol design may explore bounded fees on completed inference to support coordination, metering, verification, recovery, and other shared protocol operations. If a native asset is introduced, those fees could also be linked to reducing its supply, connecting real protocol usage to supply reduction. No mechanism is selected or promised.

05Evidence-gated development

From vision to working systems

A bold vision should become more than words. philosophers.network is turning its strongest ideas into focused prototypes, testing them against reality, and carrying forward what works.

Build boldly. Test honestly. Grow what works.

  1. 01Build
  2. 02Test
  3. 03Grow
06Before going deeper

A few clear answers

philosophers.network is early. A few boundaries should be clear from the start.

Is philosophers.network live today?

Not yet. The project has a research-backed architecture thesis and is moving into focused prototypes. This page and the Vision Paper describe what is being built, not an operating inference marketplace or production network.

Is this trying to replace cloud AI with local machines?

No. Centralized providers can be powerful and efficient. philosophers.network explores additional execution paths so each workload can be matched to the capability, privacy, trust, and provider conditions it requires. Early systems may still use curated or centralized infrastructure where appropriate.

Can any computer become a Thinker?

No. A machine must be eligible for the workload it serves. Hardware, model support, reliability, network conditions, evidence requirements, operating cost, and demand all matter. Early participation will focus on selected hardware, workloads, and curated operators.

Is the Public Lane private or guaranteed correct?

No. A Public Lane operator may be able to see the request, its supplied context, and the response, and its controls do not prove semantic correctness. Routing must therefore evaluate the complete request—including conversation history, memory, retrieved documents, attachments, and tool output—not only the visible prompt. Private, sensitive, or uncertain context should not be sent through the Public Lane; privacy-sensitive work belongs in a separately measured and policy-gated Confidential Lane, which also retains explicit limits.

07Thought, Networked.

Read the Vision Paper

This page introduces the vision. The Vision Paper explains the full case for an open inference network: why it matters, how Public and Confidential Lanes differ, how Thinkers can participate, and what the project can and cannot promise.

Explore the ideas, architecture, and evidence behind philosophers.network.

Cover of the Philosophers Network Vision Paper v1.0.0

Vision Paper v1.0.0 · 20 pages
Published August 11, 2026

Read online Download PDF
08Separate projects

Shared principles. Separate projects.

hexagora.network and philosophers.network are independent projects connected by a shared belief: access to knowledge, computation, and machine intelligence should remain open, plural, and resilient.

hexagora.network expresses that belief as a symbolic public statement. philosophers.network pursues it as a long-term research and engineering effort toward open inference infrastructure.

Participation in hexagora.network creates no ownership, allocation, access right, priority, governance right, revenue share, yield, token entitlement, or promised future benefit in philosophers.network.

Explore hexagora.network