ALU GUBI for GOs & NGOs

Project of record

The 11 Commandments

This industrial CLI implements the public ALU GUBI charter for governments and NGOs. Every install path, driver tour, and CoL/adversity flow below is conditioned by the 11 Commandments of shyamalschandra/alu_gubi.

1. 01

   Cost of Living

   Find minimum and maximum CoL for every ZIP code in the USA.

2. 02

   Adversity Score

   Use Bindi to determine actual CoL from adversity — not averages alone.

3. 03

   Restricted coupons

   Issue ALU GUBI coupons that cannot fund excise goods (alcohol, drugs, smoking).

4. 04

   Bindi unlocks

   Unlock quality-of-life tiers from lossless attested input → output.

5. 05

   Hotel & vacations

   Earn travel rewards from work achieved without disturbance.

6. 06

   Personalized gifts

   Friend gifts that are NFT-backed and AI-generated.

7. 07

   Universal library & law

   A shared corpus of library knowledge and law for every citizen.

8. 08

   Point systems

   Gym, dance, biking, education, books, library credits, textbooks, lifelong learning, VC funding.

9. 09

   Book-It competition

   Gamified reading that gives back to Goodreads and bookstore comments.

10. 10

    Fair elections

    Gamified elections with advertising coupons so the market is fair and equal.

11. 11

    Civic duty

    Citizen policing, government, and military paths while still meeting CoL.

The 11 Commandments: github.com/shyamalschandra/alu_gubi · Operators install the encrypted CLI via Homebrew tap shyamalschandra/alu-gubi-cli.

Documentation

PDF preview — reference manual

Operator reference for Homebrew install, CLI/driver commands, REST routes, PlankServices, herd computing (googolplex / infinity / finite → RNG JSON tokens), stream-agnostic --pipe sinks, and mathi-alu-gubi queues. Companion to the industrial technical report.

PDF URL: https://shyamalschandra.github.io/alu-gubi-cli/docs/reference_manual.pdf · LaTeX: docs/reference_manual.tex

Live capture · Rio terminal · animated GIF

Herd swarm viz playing in Rio

The fog-herd token shower as it runs in a Rio / Ghostty 32-bit truecolor terminal — random-walk Knuth LCG shards, Plank fog mesh, never materializing the virtual population. The loop below is a pure-Rust animated GIF of that swarm, framed as a Rio window.

rio · herd swarm viz · alu-gubi-demo-media

Live GIF

$ cargo run -p alu-gubi-demo-media --release -- --random-walk

32-bit ▀ · 16:9 · random-walk swarm · googolplex / ∞ / finite

Replay live in Rio: cargo run -p alu-gubi-demo-media --release · Driver path: alu-gubi-driver herd-stream --random-walk --dry-run · Regenerate: ./scripts/render-demo-media.sh

How-to · herd demo simulator

Run the herd stream demo — step by step

Walk through the same simulator shown in demo media: dry-run the real-time TUI, tee NDJSON to a file, then go live with plank-ingest and herd-subscribe. Virtual scales googolplex / infinity / finite never materialize individuals — only compact RNG JSON tokens.

1. 01

   Install the driver

   any

   Same Homebrew tap as the server CLI. Skip if alu-gubi-driver is already on PATH.

   Copy step 01

   brew tap shyamalschandra/alu-gubi-cli
   brew trust --formula shyamalschandra/alu-gubi-cli/alu-gubi-driver
   brew install alu-gubi-driver
   alu-gubi-driver herd-stream --help

2. 02

   Dry-run the real-time TUI

   T1

   No API needed. Tick-driven RNG tokens at googolplex scale (never materializes people). Keys: q quit · space pause · [/] chunk · -/+ tick.

   Copy step 02

   alu-gubi-driver herd-stream --scale googolplex --dry-run

3. 02b

   Plank viz token shower

   T1

   alu-gubi-cli viz — live particle shower for the Plank viz sink (same aesthetic as the demo movie). Keys: q quit · space pause · d toggle dry-run.

   Copy step 02b

   alu-gubi-cli viz --dry-run

4. 03

   Try other virtual scales

   T1

   infinity = unbounded virtual herd · finite = legacy 100B cover. Still O(batch) compact tokens only.

   Copy step 03

   alu-gubi-driver herd-stream --scale infinity --dry-run
   alu-gubi-driver herd --scale finite --dry-run --raw -g 4 --batch-size 8 -r 1

5. 04

   Tee NDJSON to a file (no TUI)

   T1

   Stream-agnostic PipelineItem — one NDJSON line per tick batch. Open the file to inspect tokens.

   Copy step 04

   alu-gubi-driver herd-stream --dry-run --no-tui \
     --pipe file:///tmp/herd-demo.ndjson \
     --duration 5 -g 8 --chunk 2 --tick 40
   head -1 /tmp/herd-demo.ndjson | python3 -m json.tool

6. 05

   Start the API (live path)

   T1

   Leave this running. Health check should return {"status":"ok"}.

   Copy step 05

   alu-gubi-cli serve
   # other terminal:
   curl -s http://127.0.0.1:8080/health

7. 06

   Emit into Plank (sealed pages)

   T2

   plank-ingest POSTs batches; server seals them into Keystone-inspired enclave pages (AES-256-GCM + homomorphic seal).

   Copy step 06

   alu-gubi-driver herd-stream --scale googolplex --pipe plank-ingest --no-tui --duration 8 -g 8 --chunk 4 --tick 30

8. 07

   Subscribe over NDJSON or SSE

   T3

   Same --pipe sinks as emit. Or curl the HTTP stream while step 06 is running.

   Copy step 07

   alu-gubi-driver herd-subscribe --format ndjson --pipe stdout --max-messages 5 --raw
   # or SSE:
   curl -N 'http://127.0.0.1:8080/v1/plank/herd/stream?format=sse'

9. 08

   Confirm herd status + watch the movie

   any

   Status JSON lists scales, pipe options, and citations. Demo media on the site shows the same simulator UX.

   Copy step 08

   curl -s http://127.0.0.1:8080/v1/plank/herd/status | python3 -m json.tool | head -40
   open 'https://shyamalschandra.github.io/alu-gubi-cli/#demo-media'

# === Herd demo simulator — step by step ===
# 1) Install (once)
brew tap shyamalschandra/alu-gubi-cli
brew trust --formula shyamalschandra/alu-gubi-cli/alu-gubi-driver
brew install alu-gubi-driver

# 2) Dry-run TUI (q to quit)
alu-gubi-driver herd-stream --scale googolplex --dry-run

# 2b) Plank viz token shower
alu-gubi-cli viz --dry-run

# 3) File pipe (no TUI)
alu-gubi-driver herd-stream --dry-run --no-tui --pipe file:///tmp/herd-demo.ndjson --duration 5

# 4) Live path — Terminal 1
alu-gubi-cli serve

# 5) Live path — Terminal 2 (emit → Plank)
alu-gubi-driver herd-stream --scale googolplex --pipe plank-ingest --no-tui --duration 8

# 6) Live path — Terminal 3 (subscribe)
alu-gubi-driver herd-subscribe --format ndjson --pipe stdout --max-messages 5 --raw

One-way binary distribution

Install via Homebrew

Version 0.1.0 — ML-KEM-768 + ML-DSA-65 hybrid encrypted releases under CCSL-1.0a. Install only through the trusted tap; no direct binary or formula downloads on this page.

Post-quantum: ML-KEM-768 key wrap · ML-DSA-65 signed manifest · AES-256-GCM payload

brew trust --formula shyamalschandra/alu-gubi-cli/alu-gubi-cli && brew install alu-gubi-cli

brew trust --formula shyamalschandra/alu-gubi-cli/alu-gubi-driver && brew install alu-gubi-driver

Tap: shyamalschandra/alu-gubi-cli · ML-DSA-65 signed release envelope · install via Homebrew only

API driver UX

Live tour · herd stream · pipe sinks

alu-gubi-driver live (aliases watch, tui, demo --watch) is a fullscreen emoji + unicode + color-block showcase. Pair it with herd-stream / herd-subscribe for the stream-agnostic RNG pipeline — same --pipe sinks on both ends. Conditioned by shyamalschandra/alu_gubi.

alu-gubi-driver live --zip 94102 --display-name "Demo Citizen"

alu-gubi-driver live \
  --zip 10001 \
  --display-name "Alex" \
  --income-usd 900 \
  --unemployment-months 2 \
  --education-years 14 \
  --domain gym --points 50 \
  --interval 1500 --skip-agents

alu-gubi-driver demo --watch --skip-agents

alu-gubi-cli viz --dry-run

alu-gubi-driver demo

┌─ 🦀 ALU GUBI Driver LIVE  ▶ live ─────────────────────────┐
│ 📍 94102 · San Francisco, CA   🪪 Demo Citizen            │
│ 🔗 http://127.0.0.1:8080   ⏱ 2000ms gap   🔁 loop 1       │
├──────────────────┬────────────────────────────────────────┤
│  Steps           │  💰 CoL zip                            │
│  ▶ ✅ health     │  {                                     │
│    ✅ col-zip    │    "zip": "94102",                     │
│    ⏳ points…    │    "state": "CA",                      │
│    · coupons     │    "col_index": "1.62", …              │
│    · bindi       │  }                                     │
├──────────────────┴────────────────────────────────────────┤
│ 🎯 ████████░░ 80%   📈 ▁▂▃▅▇█   🧮 alu_gubi params       │
│ ✅4 ⚠️0 ❌0         latency     unemp=3  edu=14yr        │
└───────────────────────────────────────────────────────────┘
 q quit · space pause · n next · [/] cycle zip · r reset

# Terminal 1: API
alu-gubi-cli serve

# Terminal 2: emit → Plank (Keystone-inspired sealed pages)
alu-gubi-driver herd-stream --scale googolplex --pipe plank-ingest

# Terminal 3: subscribe (stream-agnostic sinks)
alu-gubi-driver herd-subscribe --format ndjson --pipe stdout
# or: --pipe file:///tmp/herd.ndjson
# or: --pipe tcp://127.0.0.1:9797
# or: --pipe unix:///tmp/alu-gubi-herd.sock

alu-gubi-driver herd-stream --dry-run --no-tui --pipe stdout --duration 5
alu-gubi-driver herd-stream --dry-run --no-tui --pipe file:///tmp/herd.ndjson --duration 5
alu-gubi-driver herd-stream --dry-run --no-tui --pipe tcp://127.0.0.1:9797 --duration 5
alu-gubi-driver herd-stream --dry-run --no-tui --pipe unix:///tmp/alu-gubi-herd.sock --duration 5
alu-gubi-driver herd-stream --pipe plank-ingest --scale infinity

PlankServices · fog network · invented by Shyamal Chandra

A fog network with a herd

PlankServices is a fog network: many local edge nodes (generators, COL, coupons, points, bindi) mesh through Plank hops, aggregate at herd-bridge, and fan out to logger / viz / sealed pages — without a single choke-point cloud. The herd is the virtual population those fog edges coordinate: googolplex / infinity / finite scales emit compact RNG JSON tokens, never materializing individuals. Every publisher→subscriber hop is an edge with ROS-style power/latency logs. One NDJSON PipelineItem tees to the same --pipe sinks on emit and subscribe. Batches ride Keystone-inspired sealed enclave pages (encrypted + homomorphic integrity).

Citations (both): conversations with Kailash Chandra (Introduction to Simulation, Florida Tech / Space Coast) · reading Victor R. Lesser online (UMass Amherst MAS Lab) · Keystone (keystone-enclave.org) · CITATIONS.md

[INFO] [fog/col] [/fog/mesh] col->coupons@/fog/mesh: edge fog hop seq=12 power≈41mW

alu-gubi-cli plank-status --pulse
alu-gubi-cli serve
alu-gubi-cli viz --dry-run
alu-gubi-driver plank-status --pulse
alu-gubi-driver herd --scale googolplex --raw
alu-gubi-driver herd-stream --scale googolplex --dry-run
alu-gubi-driver herd-subscribe --format ndjson --pipe stdout

# Terminal 1: API
alu-gubi-cli serve

# Terminal 2: emit → Plank (Keystone-inspired sealed pages)
alu-gubi-driver herd-stream --scale googolplex --pipe plank-ingest

# Terminal 3: subscribe (stream-agnostic sinks)
alu-gubi-driver herd-subscribe --format ndjson --pipe stdout
# or: --pipe file:///tmp/herd.ndjson
# or: --pipe tcp://127.0.0.1:9797
# or: --pipe unix:///tmp/alu-gubi-herd.sock

alu-gubi-driver herd-stream --dry-run --no-tui --pipe stdout --duration 5
alu-gubi-driver herd-stream --dry-run --no-tui --pipe file:///tmp/herd.ndjson --duration 5
alu-gubi-driver herd-stream --dry-run --no-tui --pipe tcp://127.0.0.1:9797 --duration 5
alu-gubi-driver herd-stream --dry-run --no-tui --pipe unix:///tmp/alu-gubi-herd.sock --duration 5
alu-gubi-driver herd-stream --pipe plank-ingest --scale infinity

alu-gubi-driver herd --scale googolplex --dry-run --raw
alu-gubi-driver herd-stream --scale infinity --dry-run
alu-gubi-driver herd --scale finite --dry-run --raw

Architecture

The forge stack

💾

In-Memory SQLite

Zero external DB — PRAGMA journal_mode=MEMORY

🔐

GubiVault

Vault KV v2-compatible AES-256-GCM secrets

Visit →

GubiMoby

Pure Rust Moby engine — no Docker Desktop

Visit →

Dual Ollama AI

llama3.3 generative + gemma3:1b agentic

⛓️

Rust Ledger

Hash-linked Bindi chain in-process

⏱️

Async RT Scheduler

EDF-style deadlines, admission control, low-power idle, /v1/rt/status telemetry

🪵

PlankServices

Fog network of Plank edges by Shyamal Chandra — local mesh hops + rosout power logs + herd aggregate (NDJSON/SSE) with Keystone-inspired sealed pages

🦬

Fog herd + stream pipeline

Edge fog shards → herd-bridge → googolplex|infinity|finite tokens → one NDJSON PipelineItem → stdout · tcp · unix · file · plank-ingest

🔒

Encrypted CLI

Homomorphic-sealed release binaries