ML Systems

Take your house apart, rebuild it larger from its own parts

2 followers

Take your house apart, rebuild it larger from its own parts

2 followers

Most teardowns end in a landfill. ML Systems models your home, runs a reverse takeoff of what is actually in it, and lets lenders bid to fund the rebuild. Deconstruct instead of demolish, then rebuild larger from the recovered material.

Free

Launch tags:Android•iOS•Climate Tech

Launch Team / Built With

Context.devAPI to scrape anything for AI, 20,000+ developers trust us.

Promoted

Maker

📌

I believed AI and Machine Learning was moving too fast for the human consciousness so I LLC'ed ML Systems. I now am attempting to rebuild the American Dream of homeownership by rebuilding the existing homes. I run a construction company in Rhode Island, and I built this because the industry has a hole in it. There is software for buying a home, selling it, renovating it, insuring it and financing it. There is almost nothing for the moment a house comes down, which is exactly when most of its material value gets destroyed. A machine flattens the structure, and old-growth framing, brick, fixtures and hardware go to a landfill. Rhode Island's Central Landfill is projected to hit capacity around 2046. So the app starts there. You enter an address, it builds a model of the house, and it runs a reverse takeoff: what is actually in this structure, and what is it worth. From there it tracks the whole loop in one place - financing through deconstruction, design and construction - with one auditable record per home. Two things I want to be straight about, because I would rather you trust the numbers than be impressed by them. The software is shipped and real. The construction loop is modeled and has not been run yet. I am about to run the first cycle on my own property. Every claim in our public repo is labeled MEASURED, MODELED or ASPIRATIONAL for exactly that reason. The concepts are all open at github.com/MLSystemsRI/ml-systems-public - the ledger design, the ontology, how we compress a house into a canonical model. Happy to go deep on any of it.

Report

28d ago

Forum Threads

p/ml-systems

•

8d ago

AI is not valuable. Your context window is. (why the app runs seven windows on one rented brain)

Every contractor in my state can rent the same model ML Systems rents, at the same price, before lunch. So the app is not built on a better model. It is built on better windows: seven specialist minds on one rented brain, no fine-tune, each one seeing a different slice of the house record. To the mind that takes a house apart, a house is tonnage and fasteners. To the mind that handles lending, the same house is collateral.

What none of them can see is the company's margin. A mind that can see what the business earns will quietly start serving it, and the customer pays for that drift.

p/ml-systems

•

11d ago

I stole RRR from Elon Musk and SpaceX (the routing rule behind the app)

Musk posted "Rapidly Reusable Rockets, R R R" with a pirate flag on June 30, 2021. I took the three letters and pointed them at a house. Every recovered material in ML Systems is routed by that order: Reuse Resale Recycle to its most valuable recoverable state, decided per piece at the moment it comes off.

The part that transfers is arithmetic, not virtue: recovering a booster and scrapping it for aluminum would still have been recovery, and it wouldn't have moved cost per ton at all. Crushed concrete counts as diverted and is the end of the line for that material. A recovery rate isn't a reusability rate.

p/ml-systems

•

12d ago

The role behind the app has a name: Language Modeler

People keep asking who "vibe coded" ML Systems. Nobody did. The role is Language Modeler.

The system is modeled in English first what the house record is, who may write to it, what counts as true inside it, how evidence is graded precisely enough that the AI can carry it into code. The AI is a moderator between the English language and the coding language. It owns neither side, so the author of the English (me) is the author of record. Every time the code did something the English didn't say, the English was what I went back to.

View all

Forum Threads

p/ml-systems

•

8d ago

AI is not valuable. Your context window is. (why the app runs seven windows on one rented brain)

What none of them can see is the company's margin. A mind that can see what the business earns will quietly start serving it, and the customer pays for that drift.

p/ml-systems

•

11d ago

I stole RRR from Elon Musk and SpaceX (the routing rule behind the app)

p/ml-systems

•

12d ago

The role behind the app has a name: Language Modeler

People keep asking who "vibe coded" ML Systems. Nobody did. The role is Language Modeler.

View all