A unit of account anchored to physics — and an honest accounting of how little of a price physics can actually anchor.
Pure Python standard library. No dependencies. It measures value; it cannot move value.
python3 -m talanton # the witness
python3 -m talanton budget # the solar exergy budget
python3 -m talanton price 0.17 kWh electricityτάλαντον — the balance-beam, then the weight it measured, then the money named after the weight. The word already contains the idea.
One talanton = one gigajoule of exergy — energy weighted by its capacity to do work.
Not one gigajoule of energy. Energy is conserved, so it can never measure scarcity. Exergy is destroyed by every real process, which is exactly what makes it a cost. The user's instinct to call it "free energy" was the technically correct one: it's Helmholtz/Gibbs available work, not raw joules.
| solar irradiance (TSI) | 1361 W/m² |
| intercepted by Earth | 1.735 × 10¹⁷ W |
| absorbed after albedo (0.294) | 1.225 × 10¹⁷ W |
| per year | 3.867 × 10²⁴ J |
| exergy at the Landsberg limit | 3.606 × 10²⁴ J |
| humanity's share of it | 0.0160% |
| headroom | 6,236× |
This part is solid. It's the SI move: the kilogram stopped being a lump of metal in a vault and became Planck's constant. Nobody votes on it, and any laboratory can realise it independently.
A joule does not have a price. At one moment, in one country:
| carrier | $/GJ |
|---|---|
| coal, thermal | $2.50 |
| natural gas | $2.84 |
| utility solar | $8.33 |
| gasoline, retail | $26.38 |
| electricity, retail | $47.22 |
A 19× spread on the same joule. And that isn't market noise — burn the coal in a 40%-efficient plant and the fuel embedded in a gigajoule of electricity costs $6.25 against a $47.22 retail price:
Fuel exergy is 13% of the price. 87% is capital, grid, and institutions.
Physics prices the 13%. The other 87% is precisely the part an energy standard is hoped to make objective, and it is the part it cannot touch.
So the module's one hard gate: price_in_talanta refuses to convert without
being told the carrier.
CONVENTIONAL $0.17 = 0.0036 talanta — exergy accounts for 13% of this price;
87% is capital, delivery and institutions
UNDETERMINED $0.17 = 0.0036 talanta — no carrier named — the same joule trades
from $2.50 to $47.22/GJ (19x), so no
anchored answer exists
| unit of account — a measuring rod | physics can anchor this |
| store of value — a claim on future goods | physics cannot |
| medium of exchange — a network | physics cannot |
| distribution — who receives what | not a measurement at all |
Every energy-currency proposal since Soddy and the Technocrats anchors the first
and then quietly claims the other three. This one anchors the first and refuses
the rest in code — the witness asserts there is no issue, mint,
transfer, pay, buy, sell, or wallet function.
All primary energy humanity produces, distributed perfectly evenly to every living person:
| priced as | per person per year |
|---|---|
| coal | $189 |
| natural gas | $215 |
| utility solar | $630 |
| gasoline | $1,995 |
| retail electricity | $3,570 |
| global GDP per capita | $13,415 |
Two things fall out. The choice of price swings the answer 19× on identical physics — so the number is a political choice wearing a physics costume. And even the most generous figure is a quarter of output, because energy is ~5.6% of GDP: it is an input to value, not where value lives.
A universal dividend is fundable. It is not fundable from energy, and no unit definition changes that.
drift() — measuring how far a currency moved against a fixed physical quantity.
The exergy is identical at both dates by construction, so the entire change is in
the money. That is a deflator that cannot be revised by decree, and it is the
single most useful thing this buys you.
- The exergy factors are hard; the prices are soft. Every
Carrierrow carries one number that will be correct in a century and one that was already stale when it was written. The module never treats them as equals. - A talanton is not a claim on anything. No one owes you a gigajoule. It is a measuring rod, in the sense that a metre is not a claim on a metre of cloth.
- The Landsberg limit is a ceiling, not a yield. Real single-junction PV tops out near 33.7% (Shockley–Queisser), and captured exergy is bounded by land, materials, and transmission long before it is bounded by physics.
- Anchoring the unit does not stabilise purchasing power. If output grows faster than exergy capture you get deflation — the gold standard's failure mode, reproduced. A stable rod is not a stable economy.