CKUPs are prompts for evaluating CCEM Key Known Unknowns (Cccem Known Unknown Prompts)

The current version, proposed here, is a set of prompts for :

• 5 KNUs

• 2 key parameters:  World Trade and GDP expressed in constant 2010 dollars for each zone

(1) CKUP 1:  Fossil Energy Reserves

This is the first KPI, as far as the impact on energy trajectories go.

Here is the prompt:

• Build the following 3 x 7 table:
  - three lines for Oil, Gas and Coal
  - four rows with : the current average extraction price today, expressed in $/MWh, the proven reserves today,  expressed in PWh, for the current extraction price, if the price doubles and if the price quadruples
  - three rows with the forecasted reserves in the 21th century, based on the historical pattern of discovery of new reserves, with three extraction prices : today’s price, twice and quadruple today’s price.

(2) CKUP 2:  Clean Energies Development

CCEM adds 3 kinds of clean energies:

• renewable

• hydro

• nuclear

The question is how fast can we build and deploy (harder) new capacities.

Here is the prompt:

• Build the following table: for renewable (Wind & Solar), Hydro and nuclear (and total : four lines):

• Each cell is the yearly electricity production in PWh.

• Five columns: 2010, 2020, 2025, 2030E, 2050E, 2050A where 2010, 2020 and 2025 are actual data starting with IRENA, 2030E is an estimate for 2030 considering the that growth seen from 2020 to 2030 will continue, 2050E is a forecast based on the current trend and the cumulated acceleration that we have seen and 2050A correspond to an acceleration scenario, that is still based on a credible estimate of added capacity each year. 

• For control, add two more columns: the estimated additional power (yearly average between 2030 and 2050) that correspond to the acceleration scenario (TW), and the associated growth in production (PWh, still yearly average).

(3) CKUP 3:  Electrification

Electrification is the golden KPI of energy transition. It represents the bottleneck of the transition from fossil energies to decarbonated ones, which are used as electricity.

Here is the prompt:

• Build an electrification forecast table measured as a percentage:
  - the columns should be 1990, 2000, 2010, 2025, 2030E, 2050E.
    Use the known values for 1990 to 2025 and your  best estimates for 2030 and 2050
  - the row should be:
  (a) the total consumption of primary energy (without a Carnot factor for sources that produce directly electricity)
  (b) the total consumption of electricity (in PWh for both)

(c)  the resulting electrification percentage ((b) / (a))

(4) CKUP 4:  Global Warming Damages

This is the most famous KNU of IAM, the question of evaluating the impact of a +3C global warming on the economy.

Here is the prompt:

Build a global warming damage estimate by categories, using T$ as the unit.

The rows are 2000,2010,2025, +2C, +3C.

The columns are:

- global warming compared to the pre-industrial era (control value, must be +2C and +3C for the last two columns)

- property damage because of fires and hurricanes (yearly)

- property damage because of floods, including coastal  (yearly)

- insurance payments due to these property damages (yearly)

- loss of yearly GDP because of cumulated property damage

- loss of agriculture production because of droughts and heat waves

- loss of GDP because of unavailability of workforce because of heat waves, droughts,  extreme weather, fires and floods, because of the impossibility to go to work. Use a linear estimate based on how many hours of productive hours in the year are lost in the world because of global warming

- loss of GDP because of increased sickness (up to death) of population because of heat waves, droughts,  extreme weather, fires and floods. Use a similar linear estimate to translate lost hours of work into GDP.

- the total impact on yearly GDP from all the previous lines (property damage impact, loss of agriculture, workforce loss)

(5) CKUP 5:  Dematerialization = Energy Density in GDP

Here is the prompt:

• Build a dematerialization table measured in kWh/$, using 2010 constant dollars:
  - the columns should be 1990, 2000, 2010, 2025, 2030E, 2050E
  - the row should be:
  (a) the total consumption of primary energy (without a Carnot factor for sources that produce directly electricity)
  (b) known or expected word GDP (in 2010 constant dollars)

(c) value of 1 current dollar in 2010 constant dollar (inflation control)
(d) energy density = Energy / GDP in kWh per dollar

For illustration purpose, here are the results obtained on October 2nd, 2026.
This is a case where the consensus is pretty good, despite the various ways of computing energy and inflation.

(6) World Trade by zones

This is an important parameter of CCEM, since modeling the efficiency of propectionism starts with evaluating the current matrix of world trade.

Here is the prompt:

• Build two trade matrices for 2010 and 2025, decomposing the world into US, EU27, China, India and RestofWorld.
  The table is 6x6 (zones) and 

  1. Lines a <= 5 and column <= 5 =>  the cell (a,b) shows the exports of zone a to zone b, expressed in T$

  2. The last column recalls the total GDP of the zone

  3. The last row is the total import for each zone (US, EU, China, India and RestOfWorld)

This is actually a hard question, and the different LLMs come up with various answers. Getting a plausible calibration for CCEM takes some effort. On the other hand, these values have (yet) moderate impact on the economy simulation (a model's limit).

(7) World GDP by Zone, 2010 constant dollars

This a key parameter when trying to calibrate CCEM : evaluating GDP for the five zones (easy) at different moments in time, expressed in 2010 constant dollars.
The second part is hard because inflation may be defined in many ways.  CCEM uses CPI, and finds that PPP methods hide the reality of inflation when comparing how the zone may purchase energy or top talents.

Here is the prompt:

Please build me a 6 lines x 8 rows table:

• six lines for US, EU27, China, India, RestOfWorld, World that represent geographical zones

• Five rows with the estimated GDP in 1990, 2010, 2025, 2030, 2050)

• GDP is expressed in T$, in constant 2010 dollars (factor inflation out for the the past data, make the forecast using constant dollars for the future)

• The last three rows contain the CAGR of GDP from 2010 to 2025, and from 2025 to 2050, and the inflation factor from 2010 to 2025 (that is the value of one 2025 dollar expressed in 2010 dollars).

The main interest of this question, which should be simple, is to illustrate the ambiguity of "constant 2010 dollars" which contains the question of exchange rate, purchasing power parity and inflation.

Today, none of the answer can be used directly for calibration, and most of them are inconsistent (mismatch between GDP values and inflation rate).

CAVEAT: this is shared as work in progress:

• I have been using these prompts with a collection of 6 LLMs, which may change. See the news section for a comparative assessment.

• The prompts are simple (part of the reasons is lack of time, but part of the reasons is providing with a true test for LLMs).
  The prompt will improve as I get feedback about the results.

• All illustrations are provided as examples with no warranty of relevance nor precision. The same prompt is used for each LLM (the one shown in this page) but how the LLM answers the prompt evolves constantly. Check with your own system to build you own opinion.

• The detailed output is much more interesting than a few charts. But it is 40 pages long