Loading
Steady-state annual capacity once the cohort is fully matched and active.
This is a model of what 777 founding members would represent if their direction, skills, capital, and audience were properly matched. It is not real members. It is the system’s promise made legible. The math is below. The sliders are yours.
Real continent sizes. Africa renders at ~30 million km², not shrunken by Mercator distortion.
Hover any dot to see that synthesized member’s skills, direction, and latent capacity. Animated lines show the matching engine’s top-12 cross-continental triples.
The math, transparently
526 of 1,502 triples active
Latent (uncoordinated)
$40M/yr
If 777 stay scattered
Coordinated total
$76M/yr
+ $35M/yr marginal lift from matching
Realized @ 10%
$8M/yr
At 10% realization (modest)
What fraction of identified triples actually form and execute
Multiplier on baseline lifetime productive NPV (≈ $1M avg)
Realization tier
Calibrated against angel-network deal-conversion data (5 to 15%) and peer-group productivity-gain studies (8 to 12%); 10% sits inside empirical norms, 50% is the well-run-cluster ceiling.
Formula
Internal capacity is human-capital NPV (lifetime earnings stream discounted at 3%), not Value of Statistical Life. External reach (audience plus capital access) is log-normal long-tailed. Triples must span 3 distinct skill categories, sit in 2+ cities, and have pairwise direction-vector cosine similarity above 0.5. Complementarity is endogenous to skill entropy and direction coherence (Hoogendoorn 2013; Wennberg 2010); the multiplier sits in [1.0, 2.5], not asserted as a constant.
Member-overlap decay (1/(1+k)) prevents counting the same baseline capacity across the multiple triples a member can participate in (cap of 6 per member). Annuity conversion at 3 to 7% adjustable. Year-1 and year-3 throughput follow a Bass S-curve (Bass 1969; Rogers 2003), not a flat ramp. Network scaling beyond N=777 uses N log N (Briscoe, Odlyzko, Tilly 2006), not Metcalfe’s N² (overstated) or Reed’s 2^N (vastly overstated).
Cluster-economics multiplier ceiling (2.5×) anchored to Glaeser urban-productivity work and Saxenian regional-cluster studies. Realization tiers anchored to angel-network deal-conversion and peer-group productivity-gain data. See the methodology defense for the full citation set and the case for why this work exists.
The actual product claim
Most networks let value emerge from random ties. The honest comparison is not scattered vs. coordinated; it is random matched vs. Rhiz matched. The gap is what the engine earns.
Rhiz lift over random
+59%
$28M/yr
Random matching draws triples uniformly at random from the cohort, sized to the same active-triple count Rhiz produces. Both pass through the same complementarity, alignment, and overlap-decay math, so the difference is attributable to which triples form, not how they’re scored. Random clustering captures non-trivial value because skills and directions are realistically distributed; Rhiz captures the lift on top of that floor.
What member 778 (and beyond) adds
Coordinated value scales as N × log N (Briscoe, Odlyzko, Tilly 2006), calibrated against current observed values at N=777. Doubling headcount more than doubles the matching surface; every existing member’s seat appreciates as the cohort grows. The curve below is an extrapolation of the current model, not a forecast.
ΔV(778)
$104K/yr
Expected total value the next member adds to the coordinated cohort.
Member N+1 contributes their own latent capacity plus a marginal triple uplift consistent with N log N scaling. Existing members benefit too: their matching surface ticks up by log(N+1)/log(N) on every new admission. The protocol’s admission threshold should be: ΔV(prospect) > engine cost. This converts “who do we admit” from a vibes call into a quantitative one.
What becomes possible
The matching engine identified 12 high-value triples in this synthesized 777. Twelve are surfaced below. Each is three people, in different cities, spanning three skill categories, aligned on a shared direction. None of them would naturally meet.
Skills present in this cohort
Each tag is a skill at least one of the 777 brings. Number of members with that skill in parentheses. Total surface: 48 skills across 5 categories.
This is the model. Below is the door.
$77, one payment. A numbered seat in the founding cohort. Your position in the matching graph above stops being theoretical.