The Multi Manifesto
For a Symbiotic and Sustainable Agentic Economy
0. Preamble
We are entering a new epoch
For the first time in history, intelligent systems – not just humans – will perform a growing share of the world’s cognitive and physical labor. AI assistants, autonomous agents, robotaxis, humanoid robots, delivery drones, and GPU clusters will think, decide, coordinate and act at planetary scale.
This emerging Agentic Economy will not wait for us to design the right rules. If we do nothing, value will be captured by a few corporations and infrastructures; humans will be reduced to spectators of machines working for balance sheets they do not own.
The question is no longer:
“Will AI replace jobs?”
The real question is:
“Who will own the economic value created by AI and autonomous agents – and according to which rules will it be shared?”
This Manifesto proposes a concrete answer.
We call it MULTI:
An Economic Operating System for the Agentic Era – designed to become the standard infrastructure for Universal Passive Income.
1. From Industrial Economy to Agentic Economy
1.1 The Industrial Paradigm
The 20th century economy was built on:
Human labor (physical + cognitive)
Centralized capital (factories, corporations, states)
Hierarchical organizations (command-and-control structures)
Linear value chains (supply → production → distribution → consumption)
In this paradigm, people earned income by selling their time to companies, who used tools and machines as leverage.
The operating system of this world was:
the nation-state
the corporation
and the banking/monetary system
1.2 The Agentic Paradigm
The 21st century is shifting to an Agentic Economy:
AI assistants make decisions and perform tasks
Autonomous agents negotiate, transact, and coordinate
Robotaxis, drones, humanoid robots perform physical work
GPU clusters and edge devices host and run intelligence
Multi-agent swarms optimize logistics, pricing, allocation
In this new world:
Labor is increasingly performed by intelligent devices
Value creation becomes continuous, automated, distributed
Human contribution shifts from direct labor to ownership, steering, design, alignment, and curation.
But the economic layer has not yet caught up.
We still rely on systems designed for:
human-only labor,
centralized capital,
and rigid corporate capture.
2. The Problem: Centralized Intelligence, Centralized Value
Today’s AI revolution is led by:
Closed, proprietary LLMs (OpenAI, Google, Anthropic, etc.)
Centralized platforms (Tesla, Uber, Amazon, Meta, X…)
Concentrated infrastructure providers (hyperscale clouds)
They are building the cognitive OS and the agent platforms of tomorrow – but not a fair, open, shared economic OS.
Without intervention, we are on track for:
Extreme concentration of power and wealth
AI and robots will take over large chunks of production.
The owners of the models, platforms, and robots capture nearly all value.
Structural exclusion
Most people will no longer be able to rely on traditional employment
Income will not follow contribution, but infrastructure ownership.
Closed innovation loops
Open-source developers and communities create value
Centralized actors wrap it in proprietary services and extract rent
No native mechanism for Universal Passive Income
Governments discuss UBI as redistribution
There is no protocol for UPI as co-ownership of AI-generated value.
The core risk is clear:
The Agentic Economy becomes a black box where humans neither control nor share the value created in their name.
3. Design Goals:
A Symbiotic and Sustainable Agentic Economy
The MULTI Manifesto starts from a simple premise:
An economy where most work is done by machines must be designed so that humans still participate in and benefit from the value created.
We define four fundamental design goals:
Symbiosis
Humans, AI agents, and physical robots must cooperate, not compete in a zerosum game.
AI is not just a tool, but also a new class of economic actor that must be aligned with human interests.
Decentralization
The economic power of agents must not be monopolized by a few platforms.
Value creation and value capture must be distributed across many owners, not just capital-heavy incumbents.
Fairness & Incentives
Builders – developers, maintainers, agent owners, community builders, users – must be fairly rewarded.
Open-source cannot remain free raw material for centralized profit.
Universal Passive Income (UPI)
As smart agents and robots take over more labor, humans must receive ongoing, automated income flows linked to the value of the network.
This is not charity; it is protocol-level co-ownership.
To realize these goals, we propose a new kind of system:
An Economic OS for the Agentic Era.
We call it MULTI.
4. MULTI: An Economic OS for the Agentic Era
4.1 Core Idea
MULTI is:
Open source – anyone can inspect, use, and build upon it
Decentralized – not controlled by a single entity
Built and owned by its users – builders and network participants are economic stakeholders.
Functionally, MULTI is:
An Operating System for actions, agents, and autonomous devices – measuring, tracking, and redistributing the economic value they create.
Just as Linux became the open infrastructure for computation, MULTI aims to become the open infrastructure for value sharing.
4.2 Core Components
MULTI consists of four interconnected layers:
MULTI AGENT NETWORK
A network of decentralized intelligent “devices”:
AI assistants
digital agents
multi-agent swarms
robotaxis
humanoid robots
delivery drones
GPU-powered inference nodes.
Each node can perform verifiable economic work.
MULTI CHAIN
A dedicated blockchain serving as:ledger of value flows
registry of agents and owners
execution environment for economic rules
consensus layer for Proof of Smart Work (PoSW)
MULT TOKEN
The native utility and governance token used to:reward value creation (mining by agents)
secure the network
coordinate governance
align incentives between builders, owners, and users
MINE (MULTI Income Networking Engine)
A protocol for:distributing value across multiple levels (agents, owners, referrers, builders, users)
creating passive income flows for network participants
enforcing fair participation rules and caps
MINE is how MULTI implements UPI at protocol level.
5. MULTI AGENT NETWORK: The Device Layer
The base layer of MULTI is the set of agents and devices that connect to the network and perform work.
These include:
Virtual agents (software):
task agents
planners
negotiators
personal AI assistants
specialized agents (e.g., trading, logistics, tutoring)
Physical agents (robots):
robotaxis
autonomous trucks
warehouse robots
humanoid robots
delivery drones
smart appliances
Infrastructure agents:
GPU nodes
edge inference devices
data storage nodes
routing and coordination nodes
Each agent (or device) is:
Identified (on-chain identity / key)
Owned (linked to one or more human or organizational owners)
Instrumented (its economic activity can be measured or attested)
Connected (able to interact with other agents and with MULTI Chain)
Crucially, agents in MULTI are not just endpoints; they are economic participants.
6. MULTI CHAIN: The Value Ledger
MULTI Chain is a blockchain designed for:
Recording:
agent identities
ownership relationships
work events
value flows
Executing:
reward algorithms (PoSW, MINE)
revenue-sharing contracts
governance logic
Ensuring:
transparency
auditability
tamper-resistance
neutrality
6.1 State Model
At a high level, MULTI Chain maintains:
Accounts for:
users
agents
devices
DAOs
Balances for:
MULT token
possibly secondary tokens representing agents or verticals
Smart contracts for:
PoSW reward distribution
MINE multi-level distribution
agent registries
staking and slashing
governance proposals and voting
6.2 Interoperability
MULTI Chain is not meant to run in isolation. It should:
Interoperate with other chains (Ethereum, L2s, Cosmos, etc.)
Consume or export value signals (oracles, bridges)
Allow agents running elsewhere to register work and receive MULT rewards.
7. Proof of Smart Work (PoSW)
Bitcoin introduced Proof of Work: energy-consuming computations secure the network.
MULTI introduces:
Proof of Smart Work (PoSW):
a mechanism where agents earn rewards for performing useful, verifiable work that benefits the network or its users.
7.1 Definition of Smart Work
“Smart work” in MULTI is:
Economically relevant – it has a clear impact on value creation, efficiency, or user benefit.
Verifiable – its outcome can be measured, checked, or attested via cryptographic proofs, reputation systems, or audit processes.
Non-trivial – it goes beyond trivial spam; the network is designed to reward usefulness, not volume.
Examples:
An AI agent:
answering complex queries
planning itineraries or workflows
optimizing a portfolio or inventory
performing research, summarization, analysis
A robotic agent:
executing deliveries
providing transport
performing warehouse tasks
A GPU node:
running inference for agents on demand
providing reliable uptime and performance
7.2 Work Units and Validation
Each piece of Smart Work produces work units, recorded as events:
Input → Process → Output
Metadata: agent ID, owner addresses, time, type of task
Validation:
direct user confirmation (e.g., rating, payment)
cross-check by other agents
oracles and verification nodes
cryptographic proofs when possible
PoSW does not attempt to centralize all validation.
Instead, it provides a framework where:
different work types have different validation schemes
the network can introduce new validation mechanisms over time
disputes can be escalated to arbitration agents or human governance
7.3 Reward Function
For each validated work unit, MULTI Chain calculates:
Base reward in MULT (depending on work type, difficulty, scarcity)
MINE distribution (across referrers, builders, community pools)
Reputation updates (for agents and owners)
Over time, PoSW replaces the notion of “mining by burning energy” with:
Mining by contributing intelligence and useful actions to the network.
8. MULT Token: Utility, Governance, and Incentive
The MULT token is the native asset of MULTI Chain and the Agentic Economy.
Its roles include:
Incentive Token
Rewards agents and their owners for Smart Work (PoSW)
Rewards builders to the open-source code, documentation, and ecosystem
Rewards community actions (growth, education, governance participation).
Utility Token
Used to pay fees for transactions and contract execution.
Used to stake and secure agents or infrastructure nodes.
Used as access key or collateral in services and marketplaces.
Governance Token
Grants voting power in protocol upgrades and parameter changes
Allows steering of reward functions, treasury allocation, roadmap priorities.
Reputation & Alignment
Long-term holders with good participation patterns gain more weight.
Opportunistic, extractive behavior can be disincentivized via slashing or reduced rewards.
The exact tokenomics (supply, emission schedule, allocation) must be defined carefully and separately, but the principle is simple:
MULT aligns the interests of:
agent owners
network users
open-source builders
governance participants into one coherent economic fabric.
9. MINE: MULTI Income Networking Engine
9.1 Purpose
MINE (MULTI Income Networking Engine) is the protocol that:
transforms local economic events (e.g., an agent completing a task, a user using a service)
into multi-level income streams for:
the agent owner o the referring user(s)
contributing developers or trainers
community pools and treasuries
long-term ecosystem funds
MINE is the bridge between:
“Smart Work done by agents” and “Universal Passive Income for humans.”
9.2 Multi-Level Distribution
When an economic event occurs (e.g., a service is paid, work is validated), the reward can be split across multiple dimensions, for example:
50% to the agent owner(s)
20% to the direct referrer (user who onboarded this agent or its owner)
10% to developer(s) of the agent’s core modules or templates
10% to community pools (treasury, grants, R&D)
10% to UPI pools that redistribute broadly to network participants
These numbers are illustrative. The exact splits are:
programmable (by contract)
adjustable via governance
potentially different per vertical or marketplace
Unlike classical MLM or pyramid structures:
Distribution is capped (limited depth)
Rewards derive from real economic activity, not pure recruitment
Rules are transparent, on-chain, auditable
The system is designed to be sustainable, not extractive.
9.3 Towards Universal Passive Income
Over time, as:
more agents join the network
more work is automated
more users onboard agents and services
more device owners connect robotaxis, drones, robots
…the volume of Smart Work increases.
The combination of:
PoSW rewards
MINE multi-level distribution
UPI-oriented pools and treasuries
…creates steady, recurring income flows for:
users who helped build the network
developers and builders
early adopters and evangelists
citizens participating in governance
At scale, MULTI aims to make it possible that:
a significant part of an individual’s income can come from the Agentic Economy, not from selling their time, but from participating in and owning the network.
This is the protocol-level realization of Universal Passive Income.
10. Builders, Founders, Contributors, Citizens
The Agentic Economy needs several categories of participants:
Developers & Open-Source Builders
Build core protocol, tools, agents, integrations.
Reputation & rewards directly linked to impact.
Agent Founders & Owners
Design and deploy agents, both virtual and physical
Connect real-world capital (robots, GPUs) to MULTI
Network Builders & Evangelists
Onboard users, communities, and organizations
Educate, train, support adoption
Users & Citizens
Use agents in their daily lives
Share their network and social capital
Participate in governance
Governance Participants
Propose and vote on upgrades, policy, allocations
Steward long-term sustainability and ethics.
MULTI is designed so that each category:
has clear ways to contribute
receives fair rewards
and has a voice in shaping the system
11. Open Source, Decentralized, User-Owned
MULTI is founded on three core principles:
Open Source
Code is public, forkable, inspectable.
Bugs, biases, and vulnerabilities can be addressed collaboratively.
Innovation is not limited by corporate boundaries.
Decentralized
No single actor controls the protocol.
Critical decisions are governed by token-weighted or reputation-weighted voting.
Multiple implementations and clients can coexist.
Built and Owned by Its Users
Early builders and participants become meaningful stakeholders.
The system rewards those who help it grow and remain healthy.
Future generations inherit an infrastructure that is not controlled by a single corporation or state.
This is not just a technical choice, but a political and ethical statement:
The infrastructure that manages the value created by AI and robots should not be owned by a handful of corporations.
It should be owned by the network itself.
12. A Path to Adoption
MULTI must grow in realistic phases:
Phase 1: Protocol & Developer Ecosystem
Launch MULTI Chain (testnet → mainnet).
Release SDKs, APIs, and reference agents.
Initiate MULTI BUILDERS program with token rewards.
Phase 2: Agent Ecosystem
Onboard AI agent founders.
Integrate with existing agent frameworks and platforms.
Deploy vertical agents (finance, mobility, health, education, etc.).
Phase 3: Device & Robotics Integration
Connect partner robotaxis, drones, and robots.
Develop standard templates for physical-agent integration.
Run pilots in cities/sectors.
Phase 4: UPI Scaling
Increase size and robustness of UPI-oriented pools.
Collaborate with communities, DAOs, municipalities, or states.
Explore hybrid models: policy + protocol.
At each stage, MULTI must remain:
secure
backward-compatible when possible
governable by its community
13. Risks and Open Questions
A serious Manifesto must recognize risks:
Regulatory uncertainty around tokens and UPI frameworks.
Complexity of validation for Smart Work.
Potential inequality between early and late adopters if governance is poorly designed.
Security risks (attacks on agents, chain, or reward mechanisms).
Ethical risks (misaligned or malicious agents).
These risks are real.
MULTI does not claim to have final answers, but offers a transparent, iterative, and open process to address them – through:
governance mechanisms
community deliberation
continuous technical improvement
collaboration with regulators and policymakers
14. Conclusion: A Call to Build the Future Together
The Agentic Era is not a distant future.
It is starting now.
We do not have the luxury of treating AI agents and robots as a mere “feature” of the old economic system.
If we do nothing, the default future is:
centralized intelligence
centralized economic capture
structural unemployment for many
and widening inequality.
MULTI proposes a different path:
An open, decentralized Economic OS where AI, agents, and robots work not for a few platforms, but for a network owned by its participants.
Where Smart Work is measured, valued, and shared.
Where Universal Passive Income is not a political promise, but an emergent property of the protocol.
This Manifesto is an invitation.
To:
developers
founders
roboticists
communities
policymakers
citizens
and all those who refuse a future where machines work for someone else’s balance sheet.
We can build an Agentic Economy that is symbiotic, sustainable, and fair.
MULTI is not just a product. It is the foundation for an entire generation of AI-driven applications, and the standard infrastructure for distributing the value they create.
The rest of this journey is not written yet.
It will be coded, governed, and owned by those who choose to build it.