In plain language

The idea

Quantum mechanics, taken literally, predicts that a measuring device ends up in a superposition of outcomes: the cat both alive and dead, the pointer at every reading at once. We never see that. The textbook patch is the collapse postulate, where at measurement the state jumps to one outcome, by hand, outside the unitary law. The many-worlds picture keeps unitarity but pays with an unobservable branching multiverse. Both are answers to the same question: why one world?

The hypothesis

QIQT-H’s entry hook is a physical premise drawn from black-hole thermodynamics and holography: a bounded region of space has only a finite amount of operational information capacity — a finite number $N_R$ of distinguishable records ( $N_R<\infty$ ). Finiteness is the postulate (the “Quantized Information” core). This is one of five postulates, not the whole program: it is (P4), and it sits alongside the $(\Phi,\lambda)$ ontology (P1) and the quantum-equilibrium typicality premise (P5) — the irreducible new physics being P4 + P5, on the P1 ontology (P2–P3 are the standard quantum-relativistic arena). This finiteness has two machine-checked layers (and they are provably different): in the finite-dimensional model it is a literal record count ( $\mathrm{card}\,R \le e^{Q_R}$ ); in the continuum — where the matter algebra is the infinite Type III $_1$ of relativistic QFT — it is the corresponding finite entropy bound $S_{\mathrm{vN}}+S_{\mathrm{rel}}\le Q_R$ , which is machine-checked (EntropyNotCardinality) to be strictly weaker than a count. The finiteness is always on the records/entropy, never on the matter Hilbert space. That the capacity then takes the holographic area form $Q_R = A/4\ell_P^2$ — scaling with boundary area, not volume — is not part of the postulate: it is derived in a conditional Sakharov / induced-gravity bridge (with the value of $G/\ell_P$ carried as a datum, not derived). (2026 — the relation $G = 1/(N\Lambda_s^2)$ promoted to derived; the numerical value stays carried.) That last carried datum can itself be reduced: positing a fundamental record-granularity scale $\Lambda_s$ (the finite-information “pixel size”, $a_0 = 1/\Lambda_s$ ) in place of $\ell_P$ delivers $G = 1/(N\Lambda_s^2)$ — the Sakharov/Dvali species bound, machine-checked axiom-free (InducedNewtonConstant) — collapsing P4-MICRO’s carried inputs to a single scale $\Lambda_s$ , from which the finite capacity and $G$ both follow. The numerical value of $G$ still needs the species accounting (a frontier), and $\Lambda_s$ stays the one carried scale (a length cannot come from a count). With this induced $G$ the granularity capacity also maps onto the holographic dictionary — the boundary Cardy microstate count of a BTZ horizon equals QIQT-H’s bulk capacity exponent $(A/4)N\Lambda_s^2$ (machine-checked, HolographicBridge) — a correspondence under the shared $G$ , not an import of a boundary CFT or AdS/CFT’s cross-check. Finiteness alone gives only $S_{\mathrm{vN}}(\rho_R)\le\log N_R$ ; the area floor $S_{\mathrm{vN}}(\rho_R)\le Q_R$ is then a theorem (given the capacity postulate), and the $1/4$ a separate machine-checked theorem — but a conditional one, resting on the Sakharov bridge, not on finiteness alone.

Here is the sharp point — and getting it right took the program a while. Finite capacity by itself does not forbid a superposition in the wave function $\Phi$ : a superposition of two records is one vector in the same finite-dimensional space, costing no more room than either record alone, and $\Phi$ evolves exactly unitarily, keeping every branch. So $Q_{\max}$ is not a constraint on the wave function — but (a correction we make honestly, 2026) it does not force a single outcome either. Decoherence makes the macroscopic records non-interfering and redundantly objective, yet that removes interference; it does not make one record actual. The single actual record is supplied by a non-dynamical selector $\lambda$ — an Everett-like selection in which $\Phi$ keeps every branch and $\lambda$ marks exactly one as the actual world. $Q_{\max}$ ’s honest role is the finite record stage: it bounds how many distinguishable records a region can hold ( $\le e^{Q_R}$ ), not whether two of them can be actual.

The move. Collapse is not added as a new law and the wave function is never touched: Φ stays exactly unitary, and a non-dynamical λ marks the one actual record. Correction (2026). An earlier framing claimed the capacity bound forbids two actual records because “classical record-contents add up” — that is a category error. The holographic bound counts independent degrees of freedom, not a sum of redundant classical records (R redundant copies of one fact carry H(X), not R·H(X)); and ordinary record entropy is capped at ~A^3/4, about 10⁹¹ for the observable universe against ~10¹²² for the bound — a permanent ~31-order gap (only black holes saturate A/4, and a black hole has no records). So capacity never counts records out of existence; the single outcome is λ’s, and “two actual records can’t coexist” reduces to a classical carrier holding one value (local single-valuedness) — itself supplied by λ, not by the bound. The genuinely hard open piece is stitching the per-region actualities into one global, Lorentz-covariant λ.

What this buys, and what it doesn’t

If the hypothesis holds, finite information supplies a bounded, decoherence-selected record stage — but definiteness itself is the work of decoherence + the non-dynamical selector λ, not of the capacity bound. (Capacity bounds how many distinguishable records a region can hold; it does not forbid a multi-record superposition or select the actual one — that “capacity forbids records” reading is retired as a category error. λ makes exactly one record actual.) No collapse term, no branching ontology.

It does not, by itself, hand you the probabilities. That a given run yields outcome $k$ with frequency $|c_k|^2$ , the Born rule, is reduced to a single typicality premise (P5) about microscopic initial conditions across runs — provably underivable from unitarity alone; what stays open is justifying that premise as forced (Born itself is reduced, not open). Until that is settled, QIQT-H is an account of definiteness — why there is one outcome — not yet a complete interpretation that also says with what frequency.

Where it stands

This is a research program with a sharp core, not a finished interpretation. An earlier version leaned on a capacity-exclusion conjecture — that a region cannot hold two macroscopically distinct actual records because their information would overflow its holographic budget. We now regard that as a category error (see the correction above), and the numbers make the point vivid: the budget is so vast — about $10^{66}$ bits for a cm² boundary, $\sim 10^{122}$ for the cosmological horizon — against ordinary record entropy of $\sim 10^{25}$ bits and a structural ceiling of $\sim A^{3/4}\approx 10^{91}$ (for the whole observable universe), parametrically below the $\sim 10^{122}$ bound, so the budget is never even remotely approached. Even the universe’s total realized entropy — black-hole-dominated, $\sim 10^{104}$ bits, which carries no records — is only $\sim 10^{-18}$ of its holographic capacity. So capacity is not what gives single outcomes; λ is. What survives — and is genuinely distinctive — is the no-collapse single-world ontology: Φ exactly unitary, one non-dynamical actuality selector λ, the holographic bound supplying only the finite record stage. Absent an additional dynamical law (a new postulate with a free parameter), the framework is empirically equivalent to standard quantum mechanics — an interpretation, not new physics, stated plainly. The mathematical substrate it borrows — modular theory and relative entropy, the bookkeeping of regional information cost — is machine-verified in Lean 4. Read on for the mathematics or the open problems.