Haute Lumière
Commerce · VII.10 · MMXXVI · daylight
TEN CONCEPT BRIEFS · Chapter VII.10 — The Cosmological Frame
One page each. A reader who reads only these ten pages has the chapter.
BRIEF 1 — The Open System
The idea. A closed system cannot build order. An open one can, and the order it builds is paid for by what it throws away.
The second law says the entropy of an isolated system never falls. That is often quoted as though it forbids the accumulation of structure, and it does — for an isolated system. Earth is not isolated. It sits in a beam of low-entropy radiation from a source at 5,772 K and radiates the same quantity of energy back out at 255.1 K into a sky that is effectively at three kelvin. Entropy rises overall, exactly as the law requires, and a local decrease inside the planet is permitted because a larger increase is being exported.
Worked example. Take the numbers. The planet absorbs 122,526 TW and emits 122,526 TW. Net energy gain: zero. Net entropy export, on the blackbody convention: 6.12 × 10^14 W/K, continuously. That flow is the account out of which forests, soils, reefs, cities and every balance sheet in this edition are funded.
Why it matters. It is the physical permission slip for the whole idea of regeneration. Regeneration is not optimism and not a moral posture; it is a thing an open system under a star is thermodynamically allowed to do, and the only question is at what rate. Every chapter in Volume IV is an argument about that rate. This is the reason there is a rate to argue about.
You already know this because you have watched a room warm up and a fridge stay cold inside it, and you knew without being told that the fridge was paying for the difference by dumping heat out the back.
BRIEF 2 — Effective Radiating Temperature
The idea. A planet's temperature is not a property of the planet. It is the temperature at which the planet has to radiate in order to get rid of exactly what it takes in.
T_eff = ( S (1 - albedo) / 4 sigma )^(1/4)The numbers, end to end.
total solar irradiance 1,361 W/m2
spread over the sphere 1,361 / 4 340.25 W/m2
Bond albedo 0.294
absorbed per square metre 240.2 W/m2
sigma 5.670374419e-8 W m^-2 K^-4
T_eff 255.1 K = -18.0 CWhat to notice. Three inputs: a flux, a reflectivity, and a constant. No chemistry, no atmosphere, no oceans. The actual surface is warmer than 255.1 K because the atmosphere delays the departure of infrared, but the outgoing temperature is fixed by the budget and nothing a civilisation does to the surface changes it — the planet must still radiate 240.2 W/m² on average or it warms until it does.
Why it matters. It gives you the sink temperature for every thermodynamic argument at planetary scale, and it is the denominator of the entropy export. It is also the first number a sceptical reader will check, which is why every input above appears in lib/verify/VII_10.py beside the result.
You already know this because you know a kettle left on a hob reaches a temperature where it is losing heat as fast as the hob supplies it, and that the temperature is set by the losing, not by the kettle.
BRIEF 3 — Entropy Export, and the Factor of Twenty-Two
The idea. What a planet consumes is not energy. Energy is conserved and nothing consumes it. What is consumed is the low entropy of the incoming radiation, and the measure of how much is consumed is the ratio of the two temperatures.
The arithmetic. Blackbody radiation carries entropy flux equal to four-thirds of its energy flux divided by its temperature.
entropy in (4/3) x 1.2253e17 / 5,772 = 2.8304e+13 W/K
entropy out (4/3) x 1.2253e17 / 255.1 = 6.4035e+14 W/K
net export 6.1205e+14 W/K
ratio = 5,772 / 255.1 22.6 xThe four-thirds cancels in the ratio. On the bare Q/T convention the export is 4.59 × 10^14 W/K and the ratio is still 22.6.
The physical reading. Mean photon energy scales with temperature, so for a fixed quantity of energy, one visible photon arriving leaves as about twenty-three infrared photons. Same joules, twenty-three times as many places to put them. That spreading is the entropy export, and it is what pays for every ordered thing on the surface.
Why it matters. It converts a philosophical claim — life runs on a gradient — into a number with units that can be audited. It also sets up the chapter's central inversion: the gradient has two ends, and only one of them has ever been priced.
BRIEF 4 — The Atmospheric Window, and the Cold Sky as an Asset
The idea. Between roughly eight and thirteen micrometres the atmosphere is nearly transparent to infrared. A surface facing up through that band is not exchanging heat with the air above it. It is exchanging heat with deep space.
Worked example. In the Iranian plateau and later in Bengal, ice was made commercially for centuries on nights when the air never fell below freezing: shallow trays, insulated beneath, shielded from wind, left open to a clear sky. Mehdi Bahadori documented the architecture in 1978. In 2014 Aaswath Raman, Shanhui Fan and colleagues published a photonic surface that reflects almost all sunlight while emitting strongly in that window, and held it several degrees below ambient in direct midday sun, with no power input at all.
The engineering figure. A good radiative surface nets on the order of 40 W/m² of cooling. Over a year, at a duty cycle of 40 percent, that is 140.2 kWh-thermal per square metre; through a chiller of coefficient of performance four, it displaces 35.0 kWh of electricity per square metre per year.
Why it matters. It is the one place in this chapter where the cosmological frame produces a cash flow. A clear view of the sky is an asset with a yield, and in most jurisdictions nobody owns it, records it or protects it — which is both the opportunity and the risk.
You already know this because you have scraped frost off a windscreen on a morning when the air temperature never went below zero, and wondered.
BRIEF 5 — What the Biosphere Actually Takes
The idea. Photosynthesis is a poor engine, and the entire living world runs on a rounding error of the incoming flux.
global net primary production 104.9 Pg C/yr
as dry biomass at 0.45 kg C per kg 2.3311e+14 kg/yr
at 17.5 MJ/kg 4.0794e+21 J/yr
the whole biosphere 129 TW
human primary energy, 620 EJ 19.6 TW
absorbed solar flux 122,526 TW biosphere as a share of absorbed flux 0.106 %
humanity as a share of absorbed flux 0.0160 %
humanity as a share of the biosphere 15.2 %The two readings, and you need both. Against the Sun, humanity is negligible — sixteen thousandths of one percent. Against the living world, it is not: fifteen percent of everything photosynthesis produces, everywhere, every year, in one species' hands.
Why it matters. It is the cleanest available answer to how big are we, and the answer depends entirely on what you divide by. Abundance and pressure are the same number against two denominators, which is why arguments about limits so often turn out to be arguments about which one was meant.
You already know this because you have seen a company described in the same week as tiny (against its market) and dominant (against its sector), with both descriptions correct.
BRIEF 6 — Landauer's Limit
The idea. There is a floor under the energy cost of computation, it is a consequence of the second law rather than an engineering limitation, and it has been measured.
minimum dissipation per bit ERASED = k_B · T · ln 2
at 300 K 2.87 zJ
at the planet's 255.1 K 2.44 zJThe subtlety that matters. The cost attaches to erasure, not to computing. Rolf Landauer showed this in 1961; Charles Bennett showed in 1982 that logically reversible computation has no such bound at all. It is discarding information that costs, because erasure shrinks the number of accessible states and the second law charges for that. Antoine Bérut's group measured the floor directly in 2012 and found it where the theory put it.
Worked scale. World data centres drew about 415 TWh in 2024 — 47.3 GW continuous, 0.24 percent of world primary energy. At the Landauer floor that budget would fund 1.6490e+31 bit erasures per second.
Why it matters. It is the only hard physical limit in this edition that sits inside an ordinary planning horizon, and it is the reason the next brief exists.
You already know this because you have felt a laptop get hot and understood, correctly, that the heat was the by-product of throwing information away.
BRIEF 7 — The Efficiency Horizon
The idea. Computation currently sits about six and a half orders of magnitude above its thermodynamic floor. That gap is finite, it is closing at a measured rate, and when it closes the information economy's free lunch ends.
Green500 leader, June 2024 72.7 GFLOP/W
energy per FLOP 1.3755e-11 J
bit operations per FLOP 1,000 ASSUMED
energy per elementary bit op 1.3755e-14 J
Landauer floor at 300 K 2.87e-21 J
----------------------------------------------------
headroom 4,791,104 x = 6.68 orders
as halvings 22.2
at Koomey's post-2000 rate, 2.6 yr/halving 58 yr
at the pre-2000 rate, 1.57 yr/halving 35 yrThe honest part. The thousand-bit-operations-per-FLOP factor is an order-of-magnitude engineering estimate, not a measurement. Change it to a hundred or ten thousand and the headroom moves by a factor of ten and the horizon becomes roughly forty years or seventy-five. The horizon is decades either way, which is the only claim being made.
What follows. Jonathan Koomey's series shows energy per computation halving every 1.57 years to 2000 and every 2.6 years since. When the halvings stop, growth in computation has to be bought with raw energy — and raw energy lands straight on the waste-heat constraint in Brief 9.
Why it matters. Every plan that assumes compute gets cheaper forever is assuming something with an expiry date that a person now alive may see.
BRIEF 8 — Kardashev as Accounting, Not as a Ranking
The idea. The Kardashev scale is useful as an energy account and useless as a league table.
Nikolai Kardashev proposed it in 1964 to classify the radio power available for interstellar signalling. Carl Sagan gave it a continuous form:
K = ( log10 P - 6 ) / 10 P in wattsWhere things actually sit.
| Power | K | |
|---|---|---|
| Humanity, 2023 | 19.6 TW | 0.729 |
| Sagan Type I | 10¹⁶ W | 1.0 |
| The whole Sun | 3.828 × 10²⁶ W | 2.058 |
| The Milky Way | ~7.7 × 10³⁶ W | 3.089 |
Humanity is 509 times below Sagan Type I. And Type I is 8.16 percent of the entire absorbed solar flux.
Read as accounting, it says something specific. Every rung is a statement about where waste heat goes. A civilisation at 8.16 percent of its planet's whole radiative budget is not a civilisation with a bigger power station; it is one that has had to solve heat rejection at planetary scale. The ladder cannot be climbed on a surface, because each rung is defined by a sink and a surface has only one sky.
Why it matters. It converts a piece of science-fiction furniture into a constraint diagram, and the constraint it diagrams is the subject of Brief 9.
BRIEF 9 — The Waste-Heat Constraint
The idea. Every watt a civilisation dissipates must eventually leave as infrared, and a planet can only emit more by getting warmer. This is true of fusion, fission and photovoltaics alike. It is a limit on energy use that has nothing to do with where the energy came from.
At equilibrium T ∝ P^(1/4), so about the present point dT/T = (1/4)(dP/P), and T_eff / 4 = 63.78 K.
warming from today's 19.6 TW 0.010 K
dissipation that adds one whole kelvin 1,921 TW
warming at Sagan Type I, 10,000 TW 5.2 K
T_eff if the absorbed flux were doubled 303.4 KHow long that takes, from 19.6 TW:
| Growth | +1 K (1,921 TW) | 10,000 TW | Whole flux (122,526 TW) |
|---|---|---|---|
| 1.0 %/yr | 461 yr | 626 yr | 878 yr |
| 2.0 %/yr | 231 yr | 315 yr | 441 yr |
| 2.3 %/yr | 202 yr | 274 yr | 384 yr |
| 3.0 %/yr | 155 yr | 211 yr | 296 yr |
The reframe. The sink binds two hundred years before the source does, and the source outlasts both by five billion. The scarce good in a planetary economy is not sunlight. It is cold — and that is the single most useful sentence in this chapter. Tom Murphy put this arithmetic in Nature Physics in 2022; there is nothing in it but Stefan-Boltzmann and compound interest.
Why it matters. It makes the word limit mean a specific thing with a date attached, which is the difference between an argument and a rhetorical gesture.
BRIEF 10 — The Intensity Identity, and the Factor That Empties the Frame
The idea. The long-run budget constraint, written so that an economist can use it, is one subtraction.
growth in energy use = growth in output - improvement in energy intensity world GDP, 2023 1.054e+14 USD
world primary energy, 2023 620 EJ
value per gigajoule 170 USD/GJ
value per kilowatt-hour 0.61 USD/kWh
recent intensity improvement 1.3 %/yr
assumed long-run output growth 2.0 %/yr
implied growth in energy use 0.7 %/yr
years to the one-kelvin threshold at 0.7 %/yr 657 yrAt the recent intensity trend, the horizon of Brief 9 stretches from 231 years to 657. Raise intensity improvement to two percent a year and energy use stops growing and the constraint never binds at all. The gap — 0.7 points a year — is the only term in the entire cosmological budget that a firm, a code or a procurement policy actually moves.
And the honest negative, as a number. At a three percent discount rate a certain payoff 231 years out has a present value of 0.000978; at one percent, 0.0993; and at three percent the present value of anything falls below one part in a million at 461 years. No business case written this century should contain a cosmological term. Know the frame; never fund on it.
Why it matters. Because the frame's job is to tell you which quantity to manage, not to justify the managing — and the quantity it points at is the denominator, which is exactly what the quarterly numbers already asked for.
All figures in these briefs are computed in lib/verify/VII_10.py and sourced in the chapter's Works Cited. The 9,105× solar-to-primary ratio belongs to Chapter I.01, and the exergy apparatus to Chapter II.06; neither is re-derived here.