THE SECOND INFRASTRUCTURE

Hypothesis

Definitions

S        = set of usable, available resources
D        = set of unmet human needs
M ⊆ S×D  = feasible connections (s, d):
           compatible in location, quantity, time, safety, transport

Π_i      = private profit of actor i
W        = collective welfare

Π_i      = Π_eff + Π_gap
Π_eff    = return on value created (less input per output, better matching)
Π_gap    = return on a difference held: information, access, bargaining
           power, jurisdiction, or risk moved onto others

R(N)     = N(N−1)/2     possible pairwise relations among N participants
V ⊆ M    = connections some participant can actually see

|R(N)|   grows ~ N²     (10× participants → ~100× relations)
|V|      grows ~ N·c    each actor sees only its own c contacts

M \ V    = feasible but invisible connections  → where Π_gap lives

R(N) is an upper bound, not a model of a real society. The point is the ratio: as participants grow, the share of feasible connections anyone can see shrinks. Whoever sees across that gap can charge for it.

Statements

S1   max Π_i  ⇏  max W      Π_gap can grow while W does not
S2   price(x) → 0   ⇏   utility(x) → 0
S3   expansion  ≠  coordination

• S1. Profit can reward efficiency, or it can be a rent on a difference: knowing first, being the only one with access, sitting between two sides that cannot see each other, or moving risk onto someone else. The number alone does not say which. The hypothesis is that a significant share of profit today is Π_gap. Not yet estimated; the instrument below is how to measure it.
• S2. An unsold apple has price 0 and the same nutrients.
• S3. The first infrastructure (roads, grids, factories, networks, the Internet) is good at producing and connecting. That is not the same capability as matching what exists to what is needed.

Observations

water        ~2.2 billion people lacked safely managed drinking water (2022)
             → more than 1 in 4 humans                          [2]

wealth       top 10%    ~76% of global wealth
             bottom 50% ~2%                                     [3]

• Phone vs water. The mobile phone reached most of the planet in about twenty years; safe water has not. The two are not equivalent: a phone is made once, sold everywhere, billed monthly, deployed tower by tower; water needs pipes and treatment in every town, high upfront cost, long-term maintenance and public governance, often for people who cannot pay much. The observation is narrower: where returns are low the market rarely steps in, and where governance is weak, often nobody does. The economy spreads what pays better than what is needed most.
• Doomsday, 1960. Von Foerster, Mora and Amiot extrapolated historical population growth to a singularity on Friday, 13 November, A.D. 2026 — now weeks away [1]. Real population growth departed from that curve decades ago. What remains is the question: what happens when a system's complexity outgrows the structures built to manage it?
• Orwell, 1949. In Nineteen Eighty-Four, surplus is destroyed on purpose; ours is destroyed because it is often individually rational — no planner, so no one to remove [4].

Proposed Principle

∀ s ∈ S, d ∈ D :  (s, d) ∈ M  ⇒  s ↛ waste   while d unmet

Normative, proposed, not proven. It does not ignore economic or physical limits — those are exactly what M encodes. It asks that finding the paths in M be a basic capability, not charity.

Predictions — How It Could Be Wrong

if hypothesis holds:
  better coordination information  → less waste, fewer empty km
                                     without new production

weakens if:
  most waste is explained by safety, legal, or physical limits
  (i.e. M is already small and mostly exhausted)

Test case, stated generally: one good g moving between nodes.

p_i        = price of g at node i
c(i,j)     = cost to move g from i to j
τ_i        = policy term at i (taxes, subsidies, regulated price)
f(i→j)     = flow of g from i to j in a time window

access gap   buyer at j is served from i, while some k ∈ M exists with
             p_k + c(k,j) + τ_k  <  p_i + c(i,j) + τ_i
             and k ∉ V for the buyer

counter-flow f(i→j) > 0  and  f(j→i) > 0   same g, same window

policy gap   part of p_i − p_k comes from τ_i − τ_k, not from c or M \ V

Instrument: decompose the spread on one route.

Δp       = p_delivered − p_alt        p_alt = best alternative in M
Δp       = c_freight + c_policy + Π_gap
Π_gap    ≈ Δp − c_freight − c_policy     (residual, upper bound)

The hypothesis predicts that access gaps and counter-flows shrink when V grows toward M, without new production. If they persist once alternatives are visible, the claim fails for that good. c_policy is subtracted first, so a policy gap is not counted as coordination failure. Where prices are administered, the residual is not interpretable; the test needs a period or market where they are not.

Feasibility is local; the resource is global. Coordination is making M visible across the boundary where it is invisible today, without centralizing ownership.

Patch vs Architecture

defect        → patch fixes it
architecture  → patches accumulate, core function still fails

A system that needs constant external correction — subsidies, donations, emergency aid — to achieve its core function indicates an architecture problem, not a bug. Many of those corrections are essential; the claim is only about what their permanence signals.

Limits

• Not a claim that markets create no value. Intermediaries consolidate, refrigerate, check quality and finance.
• Not a claim that one technology solves coordination.
• Not a claim that every captured difference is bad. Finding a gap is information, and closing one can create value. Separating Π_eff from Π_gap is hard and has to be measured per domain: logistics, finance, platforms.
• Not a claim that coordination needs a central owner. Putting everyone's data in one system recreates the problem with a new owner; participation should not require surrendering ownership or privacy.

References

• [1] Von Foerster, H., Mora, P. M., & Amiot, L. W. (1960). Doomsday: Friday, 13 November, A.D. 2026. Science, 132(3436), 1291–1295.
• [2] WHO & UNICEF (2023). Progress on household drinking water, sanitation and hygiene 2000–2022.
• [3] Chancel, L., Piketty, T., Saez, E., & Zucman, G. (2022). World Inequality Report 2022.
• [4] Orwell, G. (1949). Nineteen Eighty-Four, Part Two, Chapter IX.