Cosmology, gravity, and the reach of the idea

The right neighborhood. Erik Verlinde’s emergent gravity treats gravity itself as an entropy/information effect; in a universe with dark energy, the holographic bookkeeping leaves an extra pull that mimics dark matter, reproducing flat galaxy rotation curves with no dark-matter particle. And holographic dark energy uses the same finite-information premise to get a dark-energy density of roughly the observed size.

One real win — borrowed. Recent work (Chandrasekaran–Longo–Penington–Witten) shows that including gravity and an observer turns a region’s algebra of observables from “Type III” (no well-defined entropy — everything divergent) into “Type II” (finite, well-defined entropy). That genuinely removes one infinity — the divergence in regional entropy — and QIQT-H is built on it. But it makes entropy finite; it does not renormalize gravity or cure singularities. And it is their result, not QIQT-H’s.

“But you machine-checked the Einstein equations — isn’t that quantum gravity?” No, and it is worth being exact, because the showcase theorem (qiqt_gr_ppwave_showcase) makes the gap precise rather than closing it. That result is a Jacobson “equation of state” derivation: the classical Einstein field equations emerge as the thermodynamics of a horizon, for a free field on a fixed background, conditional on three carried inputs — the matter equation of motion, the area-law relation S = A/4ℓ_P² (its ∝A form and coefficient enter conditionally via the Sakharov bridge; P4 itself postulates only finiteness, from which the area floor S_vN ≤ Q_R is a derived theorem), and a localization map. It supports exactly one claim: the metric need not be a fundamental quantum field — gravity can be emergent, like a temperature, so quantizing g_μν (gravitons, a path integral over geometries, the perturbative infinities) may be the wrong problem. It does not support “no quantum gravity needed.” Two reasons. (i) The coefficient η = 1/4ℓ_P² = 1/4ℏG is Newton’s constant — so the area-law coefficient inserts the gravitational coupling through the entropy density; the derivation has the equation of state but not the statistical mechanics (the microstates whose counting would give S = A/4). (ii) The hard cases — singularities, the Planck regime, black-hole microstates and information, and the lab tests of gravitationally-induced entanglement — are deferred to exactly that undelivered micro-theory, not solved. So the problem is relocated, not removed: from “quantize the metric” to “find the quantum degrees of freedom whose entanglement yields S = A/4 and the localization flux” — and QIQT-H currently carries the area form and its coefficient (conditionally, via the Sakharov bridge) rather than deriving them from a microstate count. The honest slogan: QIQT-H can remove the need to quantize the metric as fundamental; it does not remove the need for a quantum statistical micro-theory of spacetime.

Everything observable is slack; only horizons saturate. A trapped-ion qubit, a cubic centimetre of gas, the Earth, the Sun, the cosmic microwave background — every ordinary system sits 30–60 orders of magnitude below its holographic budget. Even the whole observable universe runs far below the ceiling: its total entropy, dominated by supermassive black holes ($\sim 10^{104}$ bits), is about $10^{-18}$ of capacity ($\sim 10^{122}$ bits) — and its ordinary record content is smaller still ($\sim A^{3/4}\approx 10^{91}$, a $\sim$31-order gap; see the capacity argument). The only systems that reach the ceiling $S = A/4\ell_P^2$ are black holes and cosmological horizons — the maximal-entropy gravitational objects, which sit on the line by construction.

The relativistic rungs — Dirac and Klein–Gordon. The study extends cleanly to the relativistic case. A particle in a box has no position floor (you can localise it arbitrarily); the oscillator has a floor in phase space (the $2\pi\hbar$ cell); the Dirac equation — the richest in structure — has a floor in position itself, set by the mass: the Compton wavelength $\lambda_C = \hbar/mc$. Try to localise an electron below it and the energy cost exceeds $2mc^2$, so the vacuum makes electron–positron pairs and the single-particle position record dissolves. The resolution saturates at $\lambda_C$ ($\sim$35 bits in a centimetre), and the 4-spinor adds a clean 2-qubit internal record (spin ⊗ particle/antiparticle) — the bit → qubit ladder realised in a real relativistic equation. Strikingly, that handoff is not arbitrary: in limited bit space the non-relativistic Schrödinger equation already acquires a maximum signal speed $v_{\max} = \hbar/ma$ on its position grid (a lattice ‘light cone’), and that speed reaches the true light speed $c$ exactly at $a = \lambda_C$ — so the bit-limited Schrödinger picture runs out of room and gives way to Dirac and Klein–Gordon precisely at the Compton wavelength.

Klein–Gordon, the spin-0 case, is the minimal rung — and the one that teaches the most. It shares the same Compton floor but strips the record to its smallest: no spin, just a single charge/sign qubit (and zero for a neutral scalar). It also exposes an obstruction Dirac avoids — its conserved density is indefinite (it can go negative), so $|\phi|^2$ is not a record law at all and single-particle position records are ill-defined. The honest fix is the field: Klein–Gordon is a tower of oscillators (one per mode), whose records are particle-occupation numbers — looping the relativistic ladder straight back to the oscillator rung.

Yet for both the verdict is identical: the floor is particle creation (set by the mass — standard relativistic QFT), not λ and not holography (the budget stays $\sim$70 orders slack); λ inert ⇒ still operationally Everett. The richest rung and the minimal one teach real physics and add no λ — thesis-empty, like the rest.

Counting is not dynamics. Finiteness of information tells you how much can be stored; it does not, by itself, give you a force, a rate, or a short-distance cutoff. Turning the premise into any of those means adding a holographic dynamics — a new theory of gravity, and the genuine unsolved problem.

What a dynamical bit-limit would do — and why we don’t see it. Make the finiteness dynamical (a genuine cutoff $Q_{\text{eff}}$, a real minimum length) rather than an inert tag, and the impact is concrete: the canonical commutator $[x,p]=i\hbar$ becomes impossible — a finite Hilbert space forces $\operatorname{Tr}[x,p]=0\neq\operatorname{Tr}(i\hbar\mathbf{1})$, so the Heisenberg algebra must deform; position turns discrete and bounded (a finite lattice with a hard edge — a built-in UV cutoff); and a minimum length gives a generalized uncertainty principle that shifts every bound spectrum. But the size of each is the ratio $(\Delta x_{\min}/L)^2$, so it is order one only when the floor is comparable to the system — a tiny-budget toy, or a saturated horizon (where it simply is Bekenstein–Hawking). For an atom the only real floor is the Compton length, and its impact — $\alpha^2$, the fine structure — is plain relativistic QM; anything Planckian is $\gtrsim$45 orders below measurement. Crucially, QIQT-H’s λ is inert and produces none of this: to predict any of it you must add the dynamical $Q_{\text{eff}}$ — a free parameter, not a consequence of the bound. That added postulate is exactly what would turn the interpretation into testable physics, and exactly what it does not yet contain.

The one falsifiable version — λ as a finite-information generator. Instead of an inert tag, let λ be a finite-information deterministic generator of the actual history — a finite rule, in the spirit of ‘t Hooft’s deterministic quantum mechanics. Then the finiteness has teeth. A generator with a budget of $B$ bits can reproduce Born statistics only up to a data size $\sim 2^B$: beyond that its output must repeat or reveal structure, so quantum ‘randomness’ would carry finite-information signatures — periodicity, compressibility, faint correlations — at large enough scales. Unlike the inert reading, this is a concrete, falsifiable prediction: quantum randomness is pseudo-random. So far every test of quantum random-number generators finds no such structure — fully consistent with a very large budget, but a genuine ongoing test.

Two honest catches keep it grounded. The break is at $\sim 2^{Q_R}$; for the holographic budget $Q_R \sim 10^{70}$ that is $2^{10^{70}}$ events — never reached (the observable universe holds $\sim 10^{120}$), so it stays operationally = QM in practice, becoming observable only near a horizon, or if quantum randomness is far more information-limited than holography suggests. And it is a different, deterministic ontology: it must pay Bell’s price (nonlocality or superdeterminism — the generator’s seed correlating with measurement settings) and still owes an account of why the generator reproduces the Born measure. So it is not a free prediction — but it is the one place the finite-information idea becomes a genuine, testable-in-principle claim rather than an interpretation.

Pushed to its end — the wall is the Poincaré recurrence. A sharper look closes the testability question with a fact, not a free parameter. The generating code — the laws of physics plus a simple initial state — can be tiny, so the universe may well be a simple deterministic program. But a program’s repeat-time is set by the state it evolves, not by the size of its code: it cycles only when its full microstate recurs. That state carries the universe’s realized entropy, $\sim 10^{104}$ bits, and ordinary thermalizing dynamics wanders through all of it — so the generator would only “repeat” at the Poincaré recurrence time, $\sim 2^{10^{104}}$, longer than the age of the universe by some $10^{103}$ orders of magnitude. Nothing shrinks that: the universe’s high entropy is a measured fact. So “the world is a simple generator” can be true and holographically motivated, yet its one fingerprint — the randomness eventually repeating — is buried at the recurrence time, forever out of reach. What is left is then not an experiment but a choice of picture: a simple deterministic program whose randomness is the ergodic unfolding of a low-complexity seed, versus an inert λ on Everett — two descriptions of exactly the same observations.