RVEIRS

AI DISCLOSURE AND NOTICE

I use language models to model language. And pathfinders to find paths. Don't blame the tool for its misuse. I'll weaponize anything I can.

This file has been fully read and fairly heavily edited by me. The design doc was mostly generated by Claude as I pieced together 20 year old memories of a design that could not exist then. I have read through it but it's still growing. Some of the other docs were derived from that and I have maybe not ready them myself, IDR. Oh, and the interface-specs.md IDK what's in there exactly, I tried to describe what I need, but I think I mostly just got the boring stuff. Think I'll have to draw the good stuff myself.

I've made substantial improvements since having to drop it back then. 20 more years of study. The maths were unfinished. The tech evolved. I already had most of the math and visuals then.

This is an active work in progress, assembled fast mostly from twenty year old memories like post-it notes from the sky. I intend to rewrite these by hand eventually. Nothing here is verified code-wise

Regionally Variable Economic Impact Research System

An observational tool. An extension of the human nervous system, into the human economic systems.

Money moves. Where it goes after it leaves your hand is invisible to you, and it's invisible to everyone else too. Not because anyone's hiding it, but because nobody has ever built an instrument that could show it.

RVEIRS is that instrument. A privacy-preserving mesh network that measures how money actually flows (geographically) through a local economy: how much of a dollar spent locally stays local, and how far it travels before it leaks out.

Standard economic indicators measure wealth, meaning how much is held and where the totals sit. RVEIRS measures current: how much is moving, in which direction, through where.

Nothing.

It doesn't recommend, rank, shame, or optimize. It reports what's true and stops. The design assumption is that people make better decisions when they can see more clearly, and that deciding for them, however well-intentioned, is both presumptuous and less effective than making the picture legible and getting out of the way.

This is a deliberate constraint, not an unfinished feature.

• Cheap. A basic node is about $20 in parts. Cost is an adoption barrier, and adoption is the whole game.
• Low-friction. A business doesn't restructure, incorporate differently, hire a lawyer, or sign anything. It reports where money it was already spending went.
• Privacy by arithmetic, not by promise. Raw numbers never leave your device. The guarantee here is a property of the math, which cannot be revised by whoever ends up on your local mesh.
• No political prerequisite. RVEIRS doesn't require agreement about capitalism, policy, or what should be done. Only that someone wants see their economy more clearly.
• Un-enclosable. Free, open, forkable, and licensed so it stays that way.

Early. Nothing is deployed. The design is thorough, the maths are worked out, and the calculator demo runs. The mesh network itself is built but unvalidated.

The documents below mark their own maturity with [proven], [design], and [idea], so you can tell how much to trust any given section. Where something is untested, it says so.

The calculator (eddy-calculator-v3.html) computes a single business's local, close, far, and void breakdown from its own books. It's an older version with known flaws, published anyway. Incomplete and honest beats complete and estimated.

Specific gaps, if any of these are your thing:

• Mesh. MeshCore roles are unvalidated. Chunking across a ~150 byte packet budget needs building.
• Math. The privacy proof, and the temporal merging problem (deltas vs. cumulative snapshots; rates don't average).
• Security. Attack surface review, especially around clustering.
• PWA dev. The interface exists as a spec and a rough demo, not an app. Current code is LLM generated from my specs, I would prefer a human touch, just literally don't have time to code it myself.
• Anyone who makes things for an audience. See below.

Open an issue, or reach out. Rough questions welcome. The docs should be honest about what's unfinished and I'd like to know about it if they're wrong.

This is not here in this repo yet.

I intend to make some videos like explaining how this all works (it's dead easy once you get it, but it's multifaceted.) and post them to yt. Trying to be educational about the math and systems and this tool in particular. If you read or hear anything here that is incorrect, I'd really like to know about it.

It's supposed to be: a 13 week, 91 episode series covering all of this, outlined and researched and free to rebuild in your own voice and style. Most of those are scheduled to be shorts. If you make things and don't know what to make, any of that is yours to take. Credit the project, and if it ever makes you money and your own needs are met, kick something back to keep this project going too, if you like.

Treat it like it has the same AGPL license the code does. Copy, paste, modify, monetize even, but never closed. Anything that descends from this must also be open. Which means whatever you make with it has to carry the same offer.

Not offering a template. Just a subject maybe worth pointing a camera at, if you need one.

GNU Affero General Public License v3.0.

AGPL specifically, rather than GPL: the network clause means a modified version run as a service also has to release its changes.

• GitHub: https://github.com/OutlandishApparatus/rveirs
• YouTube: https://www.youtube.com/@OutlandishApparatus
• itch.io: (link later, try a search if you like. it's there somewhere. I'm just disorganized and haven't got back with the actual link)
• Patreon: (link later here also)

Outlandish Apparatus. Instruments for observing things unseen.