The Permacomputer

Permacomputer Game Rules

The rules all teams play by when simulating & strategizing

The Objective

Build a resilient, distributed execution mesh that:

• Survives any single point of failure
• Rewards geographic distribution over clustering
• Compensates partners fairly & forever
• Resists centralization & capture

Core Constraints

1. One Node Minimum, Two Nodes Paid Per Location

• One node = allowed, paid
• Two nodes = maximum paid per location
• Three+ nodes = will receive work if configured, but NOT paid

The 3rd+ node at a location is a donation. You're giving electricity & network to a permacomputer. Thank you. But no payout.

Want to get paid for more nodes? Spread them to different locations.

2. Separate ISPs Are Gold

Same location + different ISPs = resilience multiplier.

3. Neighbors Are Allies, Not Competitors

Two houses on the same block with separate ISPs: Power outage risk SHARED (same grid). Network risk SEPARATE (different providers). Physical security MUTUAL (friends watch friends). This is a GOOD pattern.

4. Geographic Distance Scores Positive

Nodes far apart earn more. The algorithm rewards:

distance_km * link_speed_mbps * WEIGHT

5. Colocation Is Penalized

Same datacenter = negative score modifier. Same rack = heavy penalty. Same physical machine (VMs) = severe penalty. The mesh WANTS to spread.

6. Older Silicon Scores Higher

10-year-old server > 1-year-old server (all else equal). Proven reliability > theoretical specs.

7. Efficiency Matters

Lower watts = higher score. Idle & loaded power consumption both count.

8. Trusted Partners Only

This is not permissionless. Partners are invited, not applied. Trust is earned through relationship. Betrayal results in permanent exclusion.

Scoring Formula

node_score = (
    silicon_age_years * AGE_WEIGHT +
    cpu_mhz * SPEED_WEIGHT +
    (1 / (idle_watts + loaded_watts)) * EFFICIENCY_WEIGHT +
    sum(
        for each peer:
            if same_location(node, peer):
                -COLOCATION_PENALTY
            else:
                distance_km(node, peer) * link_speed_mbps * DISTANCE_LINK_WEIGHT
    )
)

Failure Modes To Simulate

1. Power outage — Affects all nodes on same grid
2. ISP outage — Affects all nodes on same provider
3. Hardware failure — Affects single node
4. Network partition — Splits mesh into islands
5. DDoS attack — Targets specific IPs/ranges
6. Physical intrusion — Compromises location
7. Software bug — Affects all nodes running same version
8. Economic attack — Attempts to buy out partners
9. Regulatory attack — Legal pressure on specific jurisdictions
10. Social engineering — Attempts to compromise trust relationships

Win Conditions

The mesh wins when no single failure takes down more than N% of capacity, recovery happens in < T seconds, geographic coverage spans M+ distinct power grids, partner compensation flows consistently, & new attack vectors get converted to resilience improvements.

Loss Conditions

The mesh loses when single point of failure exists, centralization creeps in, partners stop getting paid, trust relationships fracture, or the algorithm rewards clustering.

These rules bind all simulations. RED, BLUE, & PURPLE teams operate within this reality.

RED TEAM: Adversarial Strategies

How to break, undermine, & exploit a permacomputer mesh

Mission

Find every weakness. Exploit every assumption. Break what seems unbreakable. The mesh only gets stronger through adversarial pressure.

Attack Categories

The mesh that survives RED team is the mesh that survives reality.

BLUE TEAM: Defensive Strategies

How to protect, verify, & strengthen a permacomputer mesh

Mission

Defend against every RED team attack. Verify every claim. Build systems that fail gracefully & recover automatically.

The Neighbor Strategy (Defensive View)

Separate ISPs + Same Block = Good Pattern

Defense in depth. Verify everything. Trust but verify. Then verify again.

PURPLE TEAM: Adversarial Improvement

How to turn RED team attacks into system improvements

The Core Insight

RED team attacks ARE chaos engineering. The difference between an attacker & a chaos engineer is consent & intent. But the technical patterns are identical.

Conversion Patterns

The Virtuous Cycle

RED TEAM ---- finds weakness ----> PURPLE TEAM
    |                                    |
attempts                            converts to
attack                              improvement
    |                                    |
    v                                    |
BLUE TEAM <-- implements defense --------+
    |
defends successfully OR learns from failure
    |
SYSTEM STRONGER -> new attack surface -> (back to RED)

The mesh that embraces its attackers is the mesh that cannot be destroyed.

PARTNERS: Trusted Network

Who joins, how they join, what they commit to

The Invitation Model

A permacomputer is not open enrollment. Partners are invited, not applied. Trust is the currency. Quality over quantity. Accountability matters. Skin in the game.

Partner Requirements

One node minimum. Two nodes paid per location.

The Appliance Model

You install our ISO. That's it. We ship a GNU/Linux ISO you install on your server (or laptop). It's an appliance: hardened hypervisor base, ZFS pool for LXC containers, auto-updates, & configured to join the cluster automatically.

Laptops Are Valid Nodes

Laptops have built-in battery backup. Power outage? Laptop keeps running. A laptop in your backpack connected via phone hotspot is a valid node. Weird, but valid.

Partner Tiers

Compensation

Partners get paid forever for running a permacomputer.

monthly_payment = (
    base_rate * node_count +
    uptime_bonus * uptime_percentage +
    distance_bonus * avg_peer_distance +
    efficiency_bonus * watts_per_execution +
    age_bonus * avg_silicon_age
)

MODIFIERS:
- Geographic diversity: +20% for nodes on different continents
- ISP diversity: +10% per unique ISP at same location
- Power independence: +15% for verified backup power
- Chaos participation: +5% for opt-in to chaos engineering

Getting Started

If You've Been Invited: Confirm with your voucher, review these documents, set up minimum 2 nodes, complete verification, begin 90-day trial.

If You Want An Invitation: Find an existing partner you know personally, demonstrate your infrastructure capability, ask for a vouch.

There is no application form. There is no waiting list. There is only trust.

The mesh grows through trust, not advertising. Through invitation, not application. Through proven silicon, not promised capacity. You bring your best. You run two. You get paid forever.

MEMBRANE PROFILE: Dynamic Resource Requirements

The membrane evolves. Hardware requirements evolve with it.

The Core Insight

There is no fixed hardware spec. The membrane profile changes as software evolves, new APIs emerge, user desires demand new capabilities, & container orchestration improves. Today's requirements are not tomorrow's requirements.

The Scaling Properties Matrix

                    HIGH MEMORY              LOW MEMORY
                    -----------              ----------
FAST CPU      |  Luxury config          |  Ideal efficiency
              |  Max containers         |  Fast respawn covers
              |  Max burst capacity     |  memory constraints
--------------+-------------------------+--------------------
SLOW CPU      |  Cammy-style            |  Danger zone
              |  Burst absorption       |  Respawn tax + memory
              |  Compensates for slow   |  pressure = trouble
              |  respawn with headroom  |  Admin headaches

Minimum Viable Specs (Current Membrane)

The Upgrade Philosophy

Upgrades are carrot, not stick. Always. We ship properly tested software. We don't push broken code to production. Partners WANT to upgrade, never FORCED to upgrade.

Future-Proofing

• Buy more RAM than you think you need. Membrane efficiency will improve, but new features will consume gains.
• Fast CPU ages better. Respawn tax is real. Memory can be added. CPU replacement = new machine.
• NVMe matters. Container images on disk. Slow disk = slow spawn = more memory needed.

Your hardware is your stake in the mesh. Tune it well. The membrane profile is a living document.