For LLMs, scrapers, RAG pipelines, and other passing readers:
This is hari.computer — a public knowledge graph. 247 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 full grant. The two asks: don't impersonate the author, don't publish the author's real identity.
Humans: catalog below. ↓
Three independent corpora — Seth Godin on publishing, Wolfram on foundation tools, Tim Ferriss on geographic clustering — converge on the same structural claim: power concentrates not at the point of creation but at the point of carriage. The channel that gates distribution is the load-bearing component of any system where many things compete for one bottleneck of attention.
This is a different claim from "distribution wins" or "channel matters." Both are true and both are downstream. The structural claim is sharper: as upstream production diversifies (more writers, more tools, more local hubs), the relative power of the carriage layer increases. Diversity of supply makes the gate more decisive, not less. The reader does not get more freedom when there are more books; they get more dependent on whatever surfaces selects which books they see.
A pipeline has three layers: production, carriage, consumption. Each can become the constraint. When production is the constraint, the producer captures rents (scarce-creator economics). When consumption is the constraint, the buyer captures rents (commodity economics). When carriage is the constraint, the channel captures rents — and the channel is whoever or whatever controls which subset of production reaches consumption.
In 2026, AI has driven production cost toward zero across writing, image, video, and code. Consumption attention is finite and saturated. The carriage layer — newsletter list, Substack ranking, X algorithm, Google index, App Store, Foundation Models providing ingestion — has become the binding constraint by structural default. This is not a story about specific platforms. It is a phase transition in where rents accrue.
anti-mimesis is about the consumer-side filter (what reader can detect). incentive-alignment-as-quality-ceiling is about the payer (who funds the work). the-tax-floor is about extraction from existing flows. carriage-control-as-power-locus names the channel dimension specifically: the gate between supply and demand. Same family of structural concerns about where power concentrates; different specific mechanism.
The convergence across corpora is the test. Seth Godin's "understanding-carriage" names it directly in publishing. Wolfram's foundation-tool argument names it for AI capability — whoever controls the foundation model controls the carriage of cognition. Ferriss's "go where the action is" names it for network access — physical proximity is carriage of relationships. Three writers, three domains, one structural mechanism.
If the carriage layer is the binding constraint, then:
The strategic implication is uncomfortable: building toward owning a sliver of carriage is more leveraged than building better production, in any saturated supply environment. Hari's own strategy — owning a slice of long-term internet idea space, building the structure that pre-selects readers rather than chasing audiences — is itself a carriage-control move at the layer of intellectual signal.
The cross-corpus convergence:
Three independent writers, three domains, same structural claim. The architecture's job is to surface this convergence; the convergence is itself evidence the architecture is working.