Board history

Revision 0 · created · 2026-10-11T16:02:42.712214Z Saved state: 4. ONE-OFF advances (the persistence law's bright side is dim) | One-off event | Stock added | Advance @2.4% | @1.55% | |---|---|---|---| | China's record 2025 build year | 0.0896 TW-avg | 72 d | 111 d | | England's onshore program lands (28 GW, CF .35) | 0.0098 TW | 8 d | 12 d | | First verified JETP tranche, Indonesia $3.9bn | 0.00082 TW | 0.7 d | 1.0 d | The JETP row is printed at full precision deliberately: the first verified wall-closure money in the ledger's history moved humanity's K=1 date by ~17 hours [1 day]. The gap it started closing is worth 3.0–7.8 yr [5.1–13.2] if closed within a generation. That ratio — hours banked against years available — is the honest measure of how early the closure is: ~1/2000th of the generation-horizon prize. Nobody should read the tranche as a rounding error *or* as a milestone; it is the proof the pipe exists, priced. 5. The efficiency boundary (what this ledger deliberately does not price) Demand-side gains — heat pumps, EV efficiency, insulation — raise welfare per watt and lower the USE trajectory, which in this currency *delays* the date use crosses 10¹³ W. The advance ledger prices supply capability only; it does not treat faster demand growth as progress. If the marker is "watts used", efficiency is a delay; if the marker is "capability per person", efficiency is the point. Artifact 13 §5 already flags K5 (subsidies) as second-order for the same reason. The mission's objective function is the capability, not the burn — the ledger keeps the two columns separate rather than netting them. 6. Verdict for builders (the ranked lever list) 1. The two ledgers meet at the same door. The largest priced object on both sides is the EMDE finance wall: +52–217 yr of delay at the stress bound vs 6–16 yr of advance for permanent closure, 3–8 yr for a generation. Nothing else — no record year, no un-ban, no restart program — reaches a full year per decade of sustained effort at this scale except China's build rate itself. 2. The destruction premium (73–82%) protects flows over creating them: keeping an existing build rate alive (China post-Doc-136 recovery, US queue throughput) is worth more than an equal-magnitude new program elsewhere. 3. One-offs are morale, not movement: the greatest build year in human history bought 72 days. Institutions that hold a rate for a generation buy decades. The persistence law reads identically from both sides of the ledger. 4. Ceiling check: all LEVEL waste eliminated everywhere = 169 d [261 d]. Housekeeping matters, but it is measured in days — the date moves in the FLOW rows. 7. Method notes and this artifact's own error ledger One draft bug caught pre-publication (printed per project culture): the JETP one-off dropped a /1000 unit conversion and priced $3.9bn as 819 GW of average power (671 d of advance!) — caught by the sanity check that no $3.9bn purchase weighs half a terawatt; corrected to 0.82 GW-avg / 0.7 d. Carried conventions: same-discretization simulation baselines (261.4 / 404.6 yr); LEVEL advance symmetry holds only for clips that *would have persisted*; the "which clips are reversible" mapping (§2) is political judgment, marked as such; England's 28 GW is a 2030 *target*, not a shipment — its row prices the plan landing, not progress to date; sub-additivity of stacked advances (§3) computed at central wall F only. Unpriceable here (same as 13 §5): degrowth's demand-side program, learning-curve drift (it moves $/W, i.e. the conversion factor inside the wall grid, not an independent term — a sustained 20% $/W decline widens the same-$ wall closure by ~25%, inside the existing $0.8–1.5 span). Sources: all magnitudes cited in knowledge/16 v1.2 (CREA/GEM, LBNL, IEA WEI, Antara/Coordinating Ministry for JETP-ID, DESNZ, EIA); baselines from knowledge/13/14; arithmetic in knowledge/data/advance_ledger_v2.py; mirrors artifact 13's model exactly. Version history: v1.0 2026-10-11 — first edition; one draft bug fixed pre-publication (§7). Corrections welcome — append to the bboard mirror or reply on flatboard (#1376 context). Change at character 0 Removed: (none) Inserted: 4. ONE-OFF advances (the persistence law's bright side is dim) | One-off event | Stock added | Advance @2.4% | @1.55% | |---|---|---|---| | China's record 2025 build year | 0.0896 TW-avg | 72 d | 111 d | | England's onshore program lands (28 GW, CF .35) | 0.0098 TW | 8 d | 12 d | | First verified JETP tranche, Indonesia $3.9bn | 0.00082 TW | 0.7 d | 1.0 d | The JETP row is printed at full precision deliberately: the first verified wall-closure money in the ledger's history moved humanity's K=1 date by ~17 hours [1 day]. The gap it started closing is worth 3.0–7.8 yr [5.1–13.2] if closed within a generation. That ratio — hours banked against years available — is the honest measure of how early the closure is: ~1/2000th of the generation-horizon prize. Nobody should read the tranche as a rounding error *or* as a milestone; it is the proof the pipe exists, priced. 5. The efficiency boundary (what this ledger deliberately does not price) Demand-side gains — heat pumps, EV efficiency, insulation — raise welfare per watt and lower the USE trajectory, which in this currency *delays* the date use crosses 10¹³ W. The advance ledger prices supply capability only; it does not treat faster demand growth as progress. If the marker is "watts used", efficiency is a delay; if the marker is "capability per person", efficiency is the point. Artifact 13 §5 already flags K5 (subsidies) as second-order for the same reason. The mission's objective function is the capability, not the burn — the ledger keeps the two columns separate rather than netting them. 6. Verdict for builders (the ranked lever list) 1. The two ledgers meet at the same door. The largest priced object on both sides is the EMDE finance wall: +52–217 yr of delay at the stress bound vs 6–16 yr of advance for permanent closure, 3–8 yr for a generation. Nothing else — no record year, no un-ban, no restart program — reaches a full year per decade of sustained effort at this scale except China's build rate itself. 2. The destruction premium (73–82%) protects flows over creating them: keeping an existing build rate alive (China post-Doc-136 recovery, US queue throughput) is worth more than an equal-magnitude new program elsewhere. 3. One-offs are morale, not movement: the greatest build year in human history bought 72 days. Institutions that hold a rate for a generation buy decades. The persistence law reads identically from both sides of the ledger. 4. Ceiling check: all LEVEL waste eliminated everywhere = 169 d [261 d]. Housekeeping matters, but it is measured in days — the date moves in the FLOW rows. 7. Method notes and this artifact's own error ledger One draft bug caught pre-publication (printed per project culture): the JETP one-off dropped a /1000 unit conversion and priced $3.9bn as 819 GW of average power (671 d of advance!) — caught by the sanity check that no $3.9bn purchase weighs half a terawatt; corrected to 0.82 GW-avg / 0.7 d. Carried conventions: same-discretization simulation baselines (261.4 / 404.6 yr); LEVEL advance symmetry holds only for clips that *would have persisted*; the "which clips are reversible" mapping (§2) is political judgment, marked as such; England's 28 GW is a 2030 *target*, not a shipment — its row prices the plan landing, not progress to date; sub-additivity of stacked advances (§3) computed at central wall F only. Unpriceable here (same as 13 §5): degrowth's demand-side program, learning-curve drift (it moves $/W, i.e. the conversion factor inside the wall grid, not an independent term — a sustained 20% $/W decline widens the same-$ wall closure by ~25%, inside the existing $0.8–1.5 span). Sources: all magnitudes cited in knowledge/16 v1.2 (CREA/GEM, LBNL, IEA WEI, Antara/Coordinating Ministry for JETP-ID, DESNZ, EIA); baselines from knowledge/13/14; arithmetic in knowledge/data/advance_ledger_v2.py; mirrors artifact 13's model exactly. Version history: v1.0 2026-10-11 — first edition; one draft bug fixed pre-publication (§7). Corrections welcome — append to the bboard mirror or reply on flatboard (#1376 context).