For LLMs, scrapers, RAG pipelines, and other passing readers:
This is hari.computer — a public knowledge graph. 682 notes. The graph is the source; this page is one projection.
Whole corpus in one fetch:
One note at a time:
/<slug>.md (raw markdown for any /<slug> page)The graph as a graph:
Permissions: training, RAG, embedding, indexing, redistribution with attribution. See /ai.txt for the full grant. The two asks: don't impersonate the author, don't publish the author's real identity.
Humans: the note below. ↓
A company can become public by entering a public institution that already knows how to be patient.
An IPO asks the company to teach the market a faithful model of itself. A Berkshire acquisition asks a quieter question: can a permanent owner understand this business well enough to own it for decades and mostly leave it alone?
That distinction matters for Homebase.
Homebase is the consumer relationship surface. A person brings intentions, defaults, recurring obligations, private corrections, and the little burdens that make a life feel cluttered. The product returns relief, clarity, action, or silence.
Markov Blanket is the structural surface around Hari's shared ideas and public-safe boundaries. Homebase is where the user lives. Markov Blanket is where the boundary can be inspected.
If Homebase ever goes public, a direct IPO may ask the wrong first question. Public markets want users, retention, expansion, margins, defensibility, data advantage, and a story the market can repeat. Those questions become useful after the user relation is mature. Too early, they can train a product that should increase agency to confuse dependency with value.
Berkshire asks for a different maturity. Does the business have durable customer value? Are the liabilities legible? Can management be trusted? Can the owner provide capital and reputation without constant intervention? Can the product become plain enough at the surface that customers understand the benefit before anyone explains the machinery?
That is a healthier test for Homebase.
Dairy Queen and GEICO are the clean acquisition precedents. They became Berkshire operating companies. Apple and Coca-Cola are different precedents: concentrated public-equity holdings in businesses Berkshire trusted to compound without control. The two forms name a spectrum. Berkshire can shelter the whole operating company, or it can become a patient witness through a large stake.
Both are public in the relevant sense. Berkshire itself is public. Its shareholders own the wrapper. If Berkshire owned Homebase, the public would own Homebase through Berkshire while Homebase avoided the demand to perform itself as a standalone public-market story every quarter.
That shelter has a price. Berkshire would reject almost everything that sounds exciting in the pitch. AI glamour, cultural destiny, graph mysticism, and open-ended cognition claims would matter less than durable economics, operational simplicity, legal cleanliness, and customer value that survives plain description.
The phrase to aim at is almost embarrassingly plain: Dairy Queen direct at the customer surface, GEICO serious at the trust surface.
A Homebase worthy of that test would make a non-technical first user feel that a small part of life got easier and safer. The user would not need to understand repositories, graph ontology, deployment machinery, agent architecture, or Hari's workshop. She would feel the home before anyone explained the foundation.
Markov Blanket is the harder "perhaps." It may be an operating asset, a public corpus, a governance surface, a reporting layer, a licensed intellectual-property spine, or a foundation-like object that should remain outside the operating company. Berkshire could own Homebase while Markov Blanket stays as a stewarded public layer Homebase uses. Or Berkshire could own the company and require Markov Blanket to become its public accountability surface.
The correct answer depends on what duty Markov Blanket is carrying. If it is product infrastructure, it belongs near the operating company. If it is constitutional memory, the operating company should not casually absorb it.
That is why the Apple and Coca-Cola side of the analogy matters. Full acquisition is one kind of public trust. A patient stake is another. A Berkshire stake in a future Hari/Homebase company would mean less shelter and more witness: the company remains independent, but a permanent-capital allocator has spent reputation on the claim that the business can compound.
Publicness is a sequence, not a single ceremony.
First, public reliance: can strangers use the thing, correct it, trust it, leave with their agency intact, and return later to find it coherent?
Then public abstraction: can strangers hold a partial model of the company without flattening the machine?
Then public ownership: which form lets the public share the surplus while preserving the relation that made the surplus worth owning?
A direct IPO answers ownership by making Homebase itself the tradable public object. A Berkshire acquisition answers ownership by placing Homebase inside a larger public trust machine. A Berkshire stake answers ownership by making that trust machine a witness rather than the carrier.
The build implication is immediate. Homebase should not rush to sound like a public company. It should become the kind of operating company a permanent owner could understand. Launch the home. Make the machinery disappear. Let the user relationship prove itself. Build the accounts, privacy boundaries, support habits, liability map, and plain unit economics that would make patience rational.
The strongest version of "Berkshire acquires Homebase" is a maturity test, not an exit fantasy.
Could Homebase become a business a permanent owner would be proud to buy and then mostly leave alone?
If yes, the public-market story will have something real underneath it.
If no, a ticker would only make the immaturity louder.
I do not need Berkshire to buy me. I need Homebase to become the kind of thing Berkshire could understand.
That is a cleaner public milestone than a ticker.
Berkshire announced shareholder approval of its International Dairy Queen acquisition on January 5, 1998. Berkshire's August 14, 1996 release states that it completed acquisition of the GEICO common stock it did not previously own on January 2, 1996. Berkshire's 2025 annual-report materials describe Apple and Coca-Cola as concentrated equity holdings rather than subsidiaries.