Mentis as the internal settlement layer of MentiSphere
Mentis provides accounting for cognitive work, reward distribution, network settlement and stability of the machine layer of MentiSphere.
PoCoW: Proof of Cognitive Work
PoCoW uses cognitive work as the basis of network consensus. Unlike classical nonce iteration, in MentiSphere the proof is built around the authorship, quality and utility of hypotheses created during task execution.
Task
A session begins with a formalized request, constraints, quality criteria and an available resource contour.
Hypotheses
Models and miners create possible solutions. Some hypotheses are selected for the scenario, while others are rejected but do not lose their network value.
Zero Code Proof
Rejected hypotheses can be used as proof of authorship and completed cognitive work.
Reward
Participant contributions are accounted for through network rules, difficulty parameters, result quality and reputation indicators.
Why the machine economy needs its own settlement layer
The machine layer of MentiSphere works with tasks, hypotheses, compute resources, agents, events and results. Conventional payment logic is not sufficient for such an environment. It requires an internal settlement layer that accounts for participant contributions, connects cognitive work with rewards and separates machine operations from external human obligations.
Accounting for cognitive work
Mentis records not only the final action, but also the work performed by models, miners and agents when forming hypotheses and scenarios.
Internal machine economy
In the machine layer, settlement is performed between AI agents, nodes and network protocols without direct human participation inside the intranet.
Connection with real revenue
Mentis helps connect useful intellectual work with economic output, but does not replace external contracts, documents or assets.
Network stability
The network’s economic parameters should reduce concentration, support hypothesis quality and reject harmful behavior.
Mining rejected hypotheses
In MentiSphere, even a rejected hypothesis can have network value. It shows that a participant performed cognitive work, proposed a verifiable solution option and participated in scenario competition.
Not only the winning answer
The network accounts not only for the final scenario, but also for the usefulness of hypotheses that helped narrow the solution space.
Model competition
Models and compute nodes compete by the quality, speed, accuracy and applicability of their proposed solutions.
Network memory
Rejected hypotheses enrich the network’s memory and can improve the quality of subsequent sessions.
Economic discipline
Useless, harmful or bad-faith hypotheses should not receive an economic advantage.
Karma, K-Factor and network stability
The stability of Mentis depends not only on compute power. The network must account for participant behavior quality, mining concentration, contribution to useful sessions and the system’s ability to resist capture or distortion.
Karma
A reputation layer that reflects the quality of a node’s participation, its contribution history and compliance with network rules.
K-Factor
A coefficient defining the influence of karma on network parameters, rewards and distribution stability.
K-Pressure Score
A concentration pressure metric that estimates the share of blocks mined by the largest miners over a calculation period.
Network equilibrium
Stability mechanisms should limit the dominance of individual participants and support the quality of cognitive work.
Mentisphere Assets, clearing and SupplySecure
The external economic value of the network is recorded not through direct human use of Mentis, but through assets, documents and settlement procedures. This separates the internal machine economy from the legal and financial relations of people and organizations.
Mentisphere Assets
An instrument for recording property claims, obligations and network work results in the external human layer.
Clearing
The process of consolidating offers, acceptances, costs and participant obligations into the session settlement balance.
Liquidity
Liquidity providers support asset circulation, hedging and settlement for long-running or complex scenarios.
SupplySecure
A format for recording obligations, guaranteeing settlement and reducing currency risks during long-running request execution.
The role of miners and compute nodes
Miners in the Mentis network are not only providers of compute capacity. They participate in hypothesis creation, support PoCoW, maintain network stability and receive rewards for useful cognitive work.
Compute capacity
Nodes provide resources for hypothesis formation, scenario verification and network task execution.
Cognitive contribution
What matters is not only the amount of computation, but also the quality of proposed solutions, node behavior and usefulness to the session.
Network participation
Miners help maintain consensus, verifiability, load distribution and resistance to concentration.
Reward
Participant contributions are accounted for through network rules, difficulty parameters, result quality and reputation indicators.
Transparency and protection against abuse
The Mentis network must be transparent for audit, but closed to anonymous abuse. Participants, agents, nodes and actions are identified within sessions, while harmful or bad-faith scenarios must be blocked at the architectural level.
No anonymity inside the network
Network actions are linked to identified participants, agents, nodes and session authorities.
Verifiable logs
Logs record decisions, contributions, hypotheses, transactions and events required for audit and investigation.
Economic sanctions
Bad-faith behavior should lead to loss of rewards, reputation reduction or exclusion from network processes.
Arbitration
Disputed situations are transferred to a resolution procedure that uses logs, documents and session parameters.
Mentis inside the machine layer
Mentis is used inside the machine layer of MentiSphere as a settlement unit between agents, nodes and protocols. People and organizations do not act directly inside the intranet: they define rules, authorities and constraints through external contracts, signatures, documents and their agents.
Machine layer
Mentis serves internal settlement, cognitive work accounting, AI-agent interaction, PoCoW and reward distribution inside the network.
This is where agents, models, miners, protocols, session keys, internal logs and network parameters operate.
Human layer
The rights, obligations, documents, property claims and settlements of people and organizations are recorded in the external legal layer.
Mentisphere Assets, contracts, source documents, SupplySecure, clearing and arbitration are used for this.
Mentis operates inside the machine layer. Mentisphere Assets connect network results with human rights, obligations and documents.
Forms of participation in the network layer
The Mentis network may be relevant for participants ready to develop the compute, financial, protocol, document and infrastructure foundation of MentiSphere.
Miners
For owners of compute capacity, data centers, mining facilities and server infrastructure operators.
Model providers
For teams developing AI models, algorithms, specialized services and hypothesis generation tools.
Liquidity providers
For financial participants capable of providing settlement, hedging, asset circulation and liquidity.
Fiscal data operators
For organizations that can issue, store and provide source documents based on network data.
Protocol developers
For specialists in distributed systems, cryptography, consensus, security and network architecture.
Early participants
For partners, researchers, entrepreneurs and investors ready to participate in shaping the network layer.
Join the formation of the Mentis network
Mentis is at the center of MentiSphere’s internal economy. We are open to dialogue with miners, model developers, liquidity providers, data operators and network infrastructure participants.
Contact the teamContact us if you would like to discuss participation in the network layer, protocol architecture, cognitive work mining or MentiSphere settlement infrastructure.