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. ↓
The substrate already moved. The carriers that propagate the brain-substrate compound; the carriers that propagate the genital-substrate decline. Vestigial Substrate Anxiety names the reactions that haven't caught up to the shift. This piece names what the shift implies for the next century.
Idiocracy assumes genes are the carrier of cognition and that selection over genes will determine the cognitive average. Both halves are stale. Genes-as-cognition-carrier held when the brain-substrate had no scale; the brain-substrate now does the cognitive accumulation, and average gene-resident cognition is not the binding variable.
Mean human IQ may drift down through the century. Selection asymmetries are real and operate on the genital-substrate in roughly the way Idiocracy describes. But the metric the doom-frame implies, declining available cognition, goes the other direction. Available cognition is humans plus models plus tools, and it rises faster than gene-resident cognition declines, because the brain-substrate compounds and the genital-substrate doesn't. Idiocracy solves for a constraint that has been removed.
The variance widens hard. The dispersion in cognitive output, lifespan, wealth, and reach between people who use the brain-substrate and people who don't already shows in this decade and is structural, not transient. Compounding mechanics, where output re-enters the corpus and trains the next generation of carriers, guarantee the variance grows.
Whether the top of the distribution becomes god-tier and unreachable depends on a single open variable: does brain-substrate access stay broadly available, or does it gate? Today access is wide. If it remains wide, the variance is a soft stratification, large but porous. If it narrows, the same variance becomes a hard speciation event.
The earliest visibility threshold for measurable stratification, where output and lifespan and reach diverge enough that the top decile operates on a different timescale than the median, is roughly 2050. Soft, not hard, on the present trajectory.
This is not a singularity event. No merger required, no phase transition, no point-discontinuity. The convergence with mid-century AI predictions reached from other paths (Kurzweil's 2045 from Moore's-law extrapolation, for example) reflects independent reasoning landing in the same decade, not a shared mechanism. Stratification can run for a century without anything Singularity-shaped happening.
UN central projections put global population peak near 10.3 billion in the 2080s, then declining. The shape of the decline depends on whether developed-world fertility patterns spread to the developing world fully, partially, or with regional resistance. Conservative case: global TFR settles near 1.6 by 2100, halving time roughly 80 years. Aggressive case, the South Korea trajectory generalizing (TFR currently below 0.8): global TFR settles near 1.2, halving time roughly 35 years.
On the conservative trajectory, sub-5 billion around 2200, sub-1 billion around 2400. On the aggressive trajectory, sub-5 billion around 2160, sub-1 billion around 2250. More likely than not, sub-1 billion happens before 2300, conditional on no major reversal event (sustained pronatalist policy success, religious revival on a billion-person scale, biological extension that decouples fertility from generation length).
100 billion living humans is not on any current trajectory. It would require either longevity breakthroughs that extend lifespan by an order of magnitude (possible by 2500, speculative before), or industrial-scale off-world expansion. Neither is impossible. Neither is the central case.
The bodies persist for centuries on any current trajectory. The category doesn't. The boundary between human and brain-substrate-extension dissolves under cyborgization, neural interfaces, AI-resident continuations of personality, and legal personhood for non-biological agents. A 2100 census of humans requires definitional choices that a 2026 census didn't. By 2200 the question "is this entity a human" stops being answerable in the terms it was asked.
Biological extinction is a different question and not the central case for the substitution mechanism. Substrate-substitution removes the necessity for population growth, not population. People still exist. They stop optimizing for genital propagation of their own accord, in aggregate.
The predictions assume the substitution mechanism is roughly correct and that the carriers continue to compound. They become wrong if brain-substrate access narrows hard enough to remove the substitution premise, or if a generational catastrophe (war, pandemic, infrastructure collapse) resets carrier populations, or if a pronatalist coordination event sustains TFR above replacement for multiple generations across major regions. All three are possible. None is on the current trajectory.
The demographic anxiety is real. It is also mis-aimed. The substrate moved before the discourse caught up. What the discourse calls the population question is downstream of a substrate question that has already been answered by the history of technology.
At least, that's the word humans have tended to use.