v3 - Adoption & Creator Stabilizer

In plain language

Companion to RESULTS - Adoption & Creator Stabilizer. Companion added July 11, 2026 — the v1–v5-era runs predate the plain-language convention; written from the committed RESULTS as it stands today (including any verification-pass corrections already applied in that file), with no reinterpretation.

The question

Two follow-ups to the data-demand sweep, which showed the floor is funded by a 10% slice of the human-attention mint — and that a deep data collapse is paid for by creators, not the poor. So: (A) is the floor still self-funded during the thin launch decade, when the attention mint is smallest? And (B) can a formulaic mechanism take the commoditization hit off creators?

Part A — no, the launch decade is not self-funded

At launch the floor must cover about a quarter of all issuance (versus under 2% at maturity), exceeding the 10% pool by roughly 13–16%. The poor are protected the whole time — the 10th percentile holds at exactly 1.10× essentials throughout — so the only question is who funds the gap. Adoption speed is the dominant lever: reaching 50% adoption by year 2 instead of year 5 cuts the subsidy period from 3.4 years to 1.25. Minting the subsidy isn't inflationary (10-year inflation stays ~0%), but it quietly taxes creators by suppressing the creator mint; funding it externally — a launch treasury, founder reserve, or grant pool — is the clean choice. The recommendation: budget an explicit, time-boxed launch floor subsidy (≈10–16% of issuance, tapering over ~1–3 years), and have the white paper say "self-funded at maturity; the launch decade is subsidized" rather than implying self-funding from day one.

Part B — a creator stabilizer, with a correction on the record

A correction happened here. The July 2026 verification found a unit bug: the original run sized the reserve at ~8× what the spec actually provides, so its headline — full creator protection through a −70% data-value collapse — was wrong in both units and dynamics. The corrected results, from fixed code (old JSON archived as *_PRE-FIX_archive.json): banking the 5% protocol slice's surplus instead of burning it — an 18-month reserve, released flat per creator when the median dips below launch — fully holds creators only through a ~50% collapse. At −70% it cushions rather than holds: creator real income bottoms at 0.92× and ends at 0.99×, versus 0.85×/0.93× unstabilized. Nothing protects against −80% or worse (the reserve empties). Inflation cost stays negligible (0.02–0.17%/yr). The spec should state the honest claim — full protection to ~−50%, meaningful cushioning at −70% — or raise the cap (~36 months buys materially more).

The honest catch

The launch model's bootstrap leans on assumptions about real-world EVE acceptance and the launch money stock; a genuinely cold start sits outside what it captures well — the robust takeaway is the shape (a front-loaded floor deficit of ~15% of issuance for 1–3.5 years, shrinking with adoption speed), not the exact digits. Both mechanisms are formulaic and neither touches the poor (p10 at 1.10× in every run), but the stabilizer's cap and trigger are first-draft defaults meant to be argued with, and everything runs on the usual v1–v3 limits at N=25,000.

One line

The floor's self-funding is a maturity property, not a birthright — launch needs an external, time-boxed subsidy (~15% of issuance for 1–3.5 years, shrinking the faster adoption grows) — and the creator rainy-day reserve, honestly re-sized after a unit-bug correction, fully protects creators only through a ~50% data collapse and merely cushions a −70% one.

Words used here (added July 18, 2026 — plain-language house rule; the text above is unchanged). Mint / attention mint — newly created EVE; it comes only from verified live human engagement (attention), so the attention mint is the economy's money tap. Issuance — the flow of newly created money; subsidy sizes are quoted as shares of it. Floor — the guarantee that anyone can earn essentials through verified work (held at 1.10× throughout these runs). Self-funded — covered by the floor's own 10% slice of the mint with no extra money printed; the finding here is that this holds at maturity, not at launch. p10 / 10th percentile — the person poorer than 90% of people; the test's standing subject. Protocol slice — the 5% of every mint routed to system operations; its surplus is what the creator reserve banks. Median — the middle person: half earn more, half less; the stabilizer releases when the median creator dips below launch level. Commoditization — data becoming a cheap interchangeable bulk good; the collapse scenario the stabilizer exists for. Formulaic — run by a written rule, no committee or discretion. Money stock — all money in existence at once. Bootstrap — the cold-start phase where the system must grow using only what it generates itself. N=25,000 — the runs simulate 25,000 people.

Figures

Technical results

Two follow-ups to the data-demand sweep, which found the floor is funded by the 10% routing slice on the human-attention mint (not the data economy) and that a deep data collapse is paid for by creators, not the poor. This brief tests the two implications: (A) is the floor still self-funded during the thin launch decade, when the attention mint is smallest? and (B) a formulaic mechanism to take the commoditization hit off creators. Both reuse the v3 engine; files in this folder.

⚠️ REVISED — Part B corrected after the July 2026 verification (unit bug)

The original coded the reserve cap as 18 months of the c=1 reference mint's slice — ~8x the actual slice under the governor (c~0.13). The Part B headline below ("fully holds creators through a -70% collapse using ~12 of 18 reserve-months") was therefore wrong in both units and dynamics. Corrected results (fixed code; cap = 18 months of the actual trailing slice; JSON regenerated; old JSON archived as *_PRE-FIX_archive.json):

• An honestly-sized 18-month reserve fully holds creators only through a ~50% data-value collapse (released ~15 slice-months).
• At -70% (the v3 reference) it cushions rather than holds: creator real income minimum 0.92x, ending 0.99x (vs 0.85x/0.93x unstabilized). Cumulative release ~95 slice-months — the capped stock refills monthly from the ongoing 5% flow, which is why release can exceed the cap; full protection at -70% needs ~120 slice-months cumulative (unlimited-cap diagnostic).
• Nothing protects against >= -80% (reserve empties; as originally stated). Inflation cost remains negligible (0.02-0.17%/yr; bigger caps burn less, costing slightly more inflation).

Spec implication (v1.2 §6): state the honest claim — full protection to ~-50%, meaningful cushioning at -70% — or raise the cap (~36 months buys materially more). The mechanism works; the advertised strength assumed a reserve ~8x the spec. Part B prose below is preserved for the record and superseded by this section.

Part A — The floor is NOT self-funded during the launch decade

The question. Mature EDEN funds its floor from a 10% slice of a large attention mint. At launch the mint is tiny (few users, little to engage with) while the floor's obligation — keep everyone at 1.10× a real essentials basket — is full-size from day one. Does the floor still pay for itself?

Setup. Essentials are priced at mature / real-world value (a merchant prices bread by real cost, not relative to a 5%-adopted network's income), so thin-launch market incomes start far below the floor. Per-capita mint ramps 0.3→1.0 along an adoption S-curve; institutional data demand lags (scales with adoption¹·⁵). Same engine, criterion, and floor otherwise.

Findings.

1. At launch the floor must cover ~a quarter of all issuance (vs <2% mature), exceeding the 10% verification pool by ~13–16%. So the "self-funded" property is a mature-state property; the launch decade runs at a deficit that closes only as adoption scales.
2. The poor are protected the whole time — the p10 sits at exactly 1.10× throughout. The floor delivers; the only question is who funds the gap.
3. Adoption speed is the dominant lever. Reaching 50% adoption at year 2 vs year 5 cuts the subsidy period from 3.4 years to 1.25. This is a direct, quantified argument for the awareness-and-onboarding flywheel (EdenQuest): getting to scale fast isn't just growth, it shrinks the launch subsidy.
4. Funding it by minting is not inflationary — even with a thin 6-month launch money stock, the governor absorbs the top-up minting (10-yr inflation stays ~0%). But it absorbs it by suppressing the creator mint, so minting the subsidy quietly taxes creators during launch. Funding the subsidy externally (a launch treasury / founder reserve / grant pool) protects creators and is the clean choice.

Honest caveat. The "non-inflationary minting" result leans on the launch money stock being sizeable relative to the subsidy. A genuinely cold start (almost no EVE in existence) is outside what this model captures well, and the real bootstrap also depends on EVE's external acceptance — which is an adoption question, not a monetary one. The robust takeaway is the shape: a front-loaded floor deficit of ~15% of issuance for 1–3.5 years, shrinking with adoption speed.

What to do with it. Budget an explicit, time-boxed launch floor subsidy (≈10–16% of issuance, tapering over ~1–3 years), funded externally, and treat adoption velocity as a first-class monetary parameter. Put "the floor is self-funded at maturity; the launch decade is subsidized" in the white paper rather than implying self-funding from day one.

Part B — A creator stabilizer that shares the commoditization burden

The question. Under a deep data-value collapse, v3 holds the poor at 1.10× but lets median creator real income fall (0.92× at −70%, down to 0.67× at −99%). Can a formulaic mechanism share that hit?

The mechanism (no committee, no new minting). In the v1.0 spec the 5% protocol slice's surplus is burned. Replace "burn" with bank-then-release:

• Bank the slice in a reserve (capped at 18 months of the slice); burn only the overflow. The reserve is simply deferred burning.
• Trigger: when median creator real income dips below its launch level, release from the reserve, flat per creator (progressive — it lifts the median, not the whales), enough to restore the median toward launch.
• The release is exempt from the governor's issuance target (it's recycled deferred-burn, not new mint), so the governor doesn't claw it back by cutting the creator mint.

Findings.

1. It fully protects creators up to a ~70% collapse — median creator real income never drops below its launch level, using about 12 of the reserve's 18 months. That covers v3's reference commoditization case with margin.
2. It substantially cushions deeper collapses (e.g., −90%: 0.84 vs 0.77) but the finite reserve can't fully offset a permanent ≥80% regime change — as it shouldn't, since printing into a permanent collapse would just inflate.
3. It's essentially free and harms no one. Inflation rises only from ~0.02% to ~0.1%/yr; the p10 and the floor's self-funding are untouched (top-up minting stays 0). It taxes neither the poor nor top creators — it's funded entirely by deferring burns from good years into bad ones.

Honest limits & knobs. The reserve cap (18 months) and the release sizing are tunable; a bigger cap buffers deeper/longer collapses at the cost of less burning (slightly more inflation). For a permanent deep collapse the right answer isn't a bigger buffer but the freshness-decay already in the spec — it redirects value to new creation, i.e., it nudges creators to adapt rather than be subsidized indefinitely. The stabilizer is shock insurance, not a pension.

What to do with it. Adopt it as a low-risk change to the protocol slice: "burn the surplus" → "bank the surplus, release to creators when their real income dips below launch, burn the overflow." It converts a static sink into a counter-cyclical creator buffer for near-zero cost.

What the two together say

The data-demand sweep reframed EDEN's safety net as coupled to the attention mint. These two runs trace that coupling to its two real consequences and show both are manageable:

• At launch, when the mint is small, the floor needs a temporary external subsidy — bounded (~15% of issuance), short (1–3.5 yr), and shrinking the faster you grow.
• Under commoditization, when data income collapses, the burden falls on creators — and a near-free counter-cyclical reserve can absorb it up to a severe (70%) collapse.

Neither touches the poor (the p10 holds at 1.10× in every run here), and neither requires discretion — both are formulaic. They turn two open risks from the earlier brief into specified, tested mechanisms.

Caveats

Same structural limits as v1–v3 (quantity-theory pricing, no behavioral feedback, no fraud, fixed population, single 15-year horizon), run at N=25,000 for speed. The adoption model's bootstrap depends on assumptions about real-world EVE acceptance and launch money stock that sit partly outside the model. The stabilizer's reserve cap and trigger are first-draft defaults meant to be argued with. All findings are existence/robustness demonstrations under stated assumptions, not forecasts.

Files

• adoption_sim.py — launch-decade stress (real-anchor pricing, adoption S-curve, external vs minted subsidy, money-stock sensitivity)
• fig_adoption.png — launch subsidy size/duration; adoption speed as the lever
• creator_stabilizer_sim.py — counter-cyclical reserve mechanism over the commoditization sweep
• fig_creator_stabilizer.png — creator real income restored; reserve fill-and-release path
• results_adoption.json, results_creator_stabilizer.json — all rows

Raw data

⬇ results_adoption.json⬇ results_creator_stabilizer.json⬇ results_creator_stabilizer_PRE-FIX_archive.json