Calibration v0.3: 28 measured targets (cloud, frontier-model, search, OS shares; B; US attention; sentiment US/EU/AR; AR adoption), aext and auto fitted, 39/39 in range; WP-0011/0012 context updated

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Assistant: claude-code Assistant-Model: opus Assistant-Process: 14913@bnt-lap001 Assistant-Session: add7ef69-b901-4756-9a67-34cf1125d96f
2026-10-13 17:24:04 +02:00 · 2026-10-13 17:24:04 +02:00 · 3c2a155f48
commit 3c2a155f48
parent 9bd9f6f4b7
10 changed files with 447 additions and 324 deletions
--- a/checkpoints/2026/2026-10-13-review.md
+++ b/checkpoints/2026/2026-10-13-review.md
@ -27,6 +27,24 @@
 StatCounter's "Europe" region returned no rows, so the EU is approximated by
 the mean of DE, FR, IT and ES.

+## Further observations (same day)
+
+| Indicator | Jurisdiction | Value | Reference period | Source |
+|---|---|---|---|---|
+| P1 cloud infrastructure shares (Amazon / Microsoft / Google) | world | 28 / 20 / 15 % | Q2 2026 | [Synergy Research, 2026-07-30](https://www.srgresearch.com/articles/q2-cloud-market-passes-143-billion-highest-growth-rate-in-eight-years) |
+| P4 enterprise LLM API spend (Anthropic / OpenAI / Google) | mainly US | 40 / 27 / 21 % | Dec 2025 | Menlo Ventures (via [summary](https://www.usecarly.com/blog/llm-market-share/)); not checked against the original |
+| P4 US businesses paying (Anthropic / OpenAI) | US | 43.8 / 39.8 % | Aug 2026 | Ramp AI Index (via [summary](https://konabayev.com/blog/llm-market-share/)) |
+| F4 leisure time, of which TV | US | 5.2 h, 2.6 h per day | 2025 | [BLS ATUS release 2026-06-25](https://www.bls.gov/news.release/archives/atus_06252026.htm) |
+| F1 compensation of employees / GDP; social benefits / GDP | US | 0.502; 0.114 | 2026 Q2 | BEA via FRED (COE, GDP, B087RC1Q027SBEA) |
+| F1 compensation D1 / GDP; social benefits D62 / GDP | EU27 | 0.479; 16.5 % | 2025 | Eurostat nama_10_gdp, gov_10a_main |
+| same | DE | 0.544; 16.8 % | 2025 | Eurostat |
+| L6 AI's effect on daily life in 20 years (positive / negative) | EU | 55 / 35 % | fieldwork 2024 | Special Eurobarometer 554/557 (via [Verian](https://www.veriangroup.com/news-and-insights/artificial-intelligence-and-the-future-of-work)) |
+| L6 optimistic / sceptical about AI | AR | 30.7 / 44.2 % | June 2026 | La Nación Data (via [secondary report](https://blog.donweb.com/ia-futuro-argentinos-encuesta-2026/)); not checked against the original |
+| L3 firms applying AI in daily operations; planning to | AR | 20.4 %; 44.4 % | 2026 | Google / IDB study (via [iProfesional](https://www.iprofesional.com/management/424296-que-cantidad-de-argentinos-ya-usa-inteligencia-artificial-para-trabajar)) |
+
+Secondary-source figures are marked. They should be checked against the
+originals at the next review.
+
 ## Operators applied

 - Discovery channel Pd = leading search share.
@ -42,7 +60,15 @@ P-2026-004 (US sentiment falls while adoption grows): the historical series
 (2021 → 2026, Σ −0.12) runs in the predicted direction. This is a baseline,
 not an observation against the prediction, which starts at checkpoint 0.

+- Infrastructure Pi = top-3 cloud share. Lineage Plin = top-3 share of
+  enterprise model spend.
+- Attention A = share of leisure not taken by TV and engagement media.
+- Claims breadth B = compensation of employees / GDP + social benefits / GDP.
+- Σ for the EU and AR uses each survey's positive − negative balance. The
+  questions differ from Pew's, so the ranges are wide.
+
 ## Still priors after this review

-Pi (cloud shares), Plin (frontier developers), Pm (F2), A (ATUS), B level,
-E, EU and AR sentiment, AR adoption (L3). They remain WP-0015-T01/T02.
+Market concentration Pm (F2), ecological headroom E, EU and DE attention, AR
+claims breadth and AR adoption *growth* (only one level point). They remain in
+WP-0015-T02.
--- a/models/Calibration.md
+++ b/models/Calibration.md
@ -1,7 +1,7 @@
 # Calibration of GADM-0.4 to measured data

-**Status:** v0.2 (2026-10-13), ASSEMBLY-WP-0013, updated with measured
-channels and US sentiment (§ 3.1). Calibrates starting states
+**Status:** v0.3 (2026-10-13). ASSEMBLY-WP-0013, then the 2026-10-13 review
+(§ 3.1): 28 measured targets, all inside their ranges after fitting. Calibrates starting states
 for US, EU, DE and AR. Argentina has no measured adoption growth, so it is
 calibrated only on priors. Most state targets are literature priors that the
 next review must verify against checkpoint data.
@ -77,16 +77,52 @@ The present is between B2 and B3. B1 is ruled out as a description of
 today, and B4 as a description of today's level. B4 can still describe the
 future if self-enhancement starts later.

-### 3.1 Update with measured values (review 2026-10-13)
+### 3.1 Calibration v0.3: measured targets and fitted parameters

-`checkpoints/2026/2026-10-13-review.md` replaced several priors with
-measurements: the discovery channel Pd (leading search share: US 0.85,
-EU 0.89, DE 0.88, AR 0.94), the membrane channel Pmem (mobile OS HHI: US 0.50,
-EU 0.57, DE 0.60, AR 0.78), and US sentiment Σ (Pew 2026: 0.285, down from
-0.405 in 2021). The adoption fit and the anchor verdict did not change.
-Starting domination is now US 0.63, EU 0.67 and **AR 0.71**: Argentina starts
-above the floor, driven by foreign-operated discovery and membrane channels.
-Still priors: Pi, Plin, Pm, A, B level, E, EU and AR sentiment, AR adoption.
+The review `checkpoints/2026/2026-10-13-review.md` measured 28 of the 39
+targets:
+
+- adoption level and growth (BTOS, Eurostat; AR level from the Google/IDB
+  study);
+- search share (Pd), mobile-OS HHI (Pmem), cloud top-3 share (Pi),
+  frontier-model top-3 share (Plin);
+- claims breadth B (compensation plus social benefits, % of GDP) and the US
+  B trend;
+- US attention (ATUS 2025);
+- sentiment Σ for US (Pew 2026), EU (Eurobarometer) and AR (La Nación Data).
+
+Fitted parameters per jurisdiction:
+
+| | k (× B2) | t* | wL0 | ereg | aext | auto |
+|---|---|---|---|---|---|---|
+| US | 2.00 | 1.2 y | 0.478 | 0.0096 | 0.387 | **0.26** (fitted to the B trend) |
+| EU | 1.54 | 0.9 y | 0.523 | 0.0098 | 0.931 | 0.60 (default) |
+| DE | 1.41 | 1.5 y | 0.616 | 0.0103 | 0.983 | 0.60 |
+| AR | 2.00 | 1.1 y | 0.387 | 0.0500 | 0.580 | 0.60 |
+
+**Fit: 39 of 39 targets inside their range, including 28 of 28 measured
+targets.**
+
+- **Fitted against data:** adoption speed k and alignment t* (level and
+  growth), `wL0` (B level), `aext` (attention, US measured, others prior) and
+  `auto` (US B trend).
+- **Set by assignment:** the channel levels are set to the measured values,
+  and the channel counter coefficients are fitted so that those levels are at
+  rest. That is a calibration of the counters, not a test of them.
+
+Findings:
+
+1. **Adoption.** The anchor verdict holds. B1 cannot produce the observed
+   growth anywhere, and B3 fits best (US z 0.0, AR 0.0, EU +0.5).
+2. **Automation.** The measured US labour-share decline is slower than the
+   default automation strength implies. Fitted `auto` is 0.26 instead of 0.6.
+3. **Domination starts at or above the floor:** US 0.69, EU 0.72, AR 0.75
+   (floor 0.70). The large contributions are measured: frontier-model spend
+   is 88 % with three firms, and search is 85–94 % with one.
+
+Still priors: market concentration Pm, ecological headroom E, EU and DE
+attention, AR claims breadth, and AR adoption *growth* (only one level point).
+They stay in ASSEMBLY-WP-0015-T02.

 ## 4. Readings from the calibrated start (T04)

@ -97,24 +133,25 @@ present, with the legitimacy loop off and on.
 |---|---|---|---|---|
 | R1 rights-first levers give the best VFAC (§ 4.3 of AdoptionBand) | holds | holds | holds | **survives** |
 | R2 the legitimacy loop damps adoption (§ 3.3) | fails under J1 | holds | fails under J1 | **revised** |
-| R3 lagging alone costs early output (§ 4.3) | −0.080 Ȳ | −0.078 | −0.068 | **survives, stronger** |
-| R4 the current regime breaks the domination floor by t 100 | yes (J3) | yes (J1) | yes (J2; already above at t 0) | **new** |
+| R3 lagging alone costs early output (§ 4.3) | −0.088 Ȳ | −0.084 | −0.121 | **survives, stronger** |
+| R4 the current regime breaks the domination floor by t 100 | yes (J3; P 0.69 → 0.82) | yes (J1; 0.72 → 0.82) | yes (J2; 0.75 → 0.86) | **new** |

 1. **R2 revised: the loop's sign depends on the regime.** Under rights-first
   levers, domination falls, perceived benefit rises, and sentiment
   *accelerates* adoption (US: K_Ω peak 0.21 → 0.26). Under extractive regimes
   it brakes adoption as before. Damping is not a property of the loop. It
   is a property of the loop in a regime where adoption raises domination.
-2. **R4: today's domination is close to the floor.** Calibrated to observed
-   concentration, aggregate P starts at 0.59–0.65 against a floor of 0.70.
+2. **R4: today's domination is at the floor.** Calibrated to measured
+   concentration (v0.3), aggregate P starts at 0.69–0.75 against a floor of
+   0.70.
   Under the current regime, continued adoption pushes it over the floor in
   all three jurisdictions. Uncalibrated GADM started at P = 0.37 and
   understated this.
 3. **The EU has no lever headroom left within J1.** The channel counters are
   calibrated so that the current regime is at rest. Lever effects are
   therefore *relative to the current regime*. The US gains by moving from J3
-   to J1 levers (P 0.64 → 0.54). The EU is already at J1 settings, and its P
-   rises to 0.77. Within this closed-economy model, the EU's domination comes
+   to J1 levers (P 0.69 → 0.55). The EU is already at J1 settings, and its P
+   rises to 0.82. Within this closed-economy model, the EU's domination comes
   from levels set abroad (lineages, platforms), which its levers do not
   reach. This is the strongest argument yet for WP-0011.
 4. **Readings that still need re-running:** the GSA per anchor (WP-0009-T05)
@ -138,8 +175,8 @@ When a checkpoint brings new data:

 ## 6. Limitations

- Most state targets are priors, not measurements. Verifying them is
-  WP-0013-T06.
+- Nine targets are still priors (§ 3.1). Several measured values come from
+  secondary reports (see the review record).
 - One adoption speed parameter (k) for four parameters (aO, ell, oseed,
  selfenh). k does not separate investment from learning.
 - The channel-counter calibration makes the current regime a rest point by
--- a/simulations/gadm02/calibrate_data.py
+++ b/simulations/gadm02/calibrate_data.py
@ -6,7 +6,9 @@ models/Calibration.md). Per jurisdiction:
 1. Set stocks with targets directly (power channels, E, Sigma) and fit the
   parameters that place them at rest: channel counters (gadm04.calibrate),
   ecological regeneration `ereg` (dE = 0), base labour share `wL0` (claims
-   breadth B on target). Archetype levers of the nearest archetype
+   breadth B on target), attention extraction `aext` (attention at rest on
+   target, bisection). Where a measured trend of B exists (US), the
+   automation strength `auto` is fitted to it. Archetype levers of the nearest archetype
   (models/AdoptionBand.md § 4.4) are applied except `perm`.
 2. Relax wealth, cohort attention and cohort agency to equilibrium.
 3. Adoption: with an anchor's adoption parameters scaled by k, find the model
@ -70,12 +72,14 @@ def relax(s: dict, p: dict, keys=gr.EQ_KEYS, steps: int = 4000, h: float = 0.25)
    return s


-def start_state(jur: str, anchor: str) -> tuple[dict, dict]:
+def start_state(jur: str, anchor: str, auto: float | None = None) -> tuple[dict, dict]:
    """(params, init) at rest with the jurisdiction's state targets, before adoption alignment."""
    t = targets(jur)
    vp, vi = gr.variant(gr.BAND[anchor])
    _title, lev, _s0 = ARCHETYPES[NEAREST[jur]]
    p = {**g.PARAMS, **vp, **{k: v for k, v in lev.items() if k != "perm"}}
+    if auto is not None:
+        p["auto"] = auto
    s = {**g.INIT, **vi}
    for k in CHANNEL_KEYS:
        if k in t:
@ -96,9 +100,23 @@ def start_state(jur: str, anchor: str) -> tuple[dict, dict]:
            p["ereg"] = drain * s["E"] / (1 - s["E"])
        p.update(g.calibrate(p, s))
        s = relax(s, p)
+        if "A" in t:
+            p["aext"], s = fit_attention(s, p, centre(t["A"]))
    return p, s


+def fit_attention(s: dict, p: dict, target: float) -> tuple[float, dict]:
+    """Bisection on the extraction strength aext so that attention at rest equals `target`."""
+    lo, hi = 0.05, 3.0
+    for _ in range(14):
+        mid = 0.5 * (lo + hi)
+        s2 = relax(s, {**p, "aext": mid}, steps=1500)
+        _d, a = g.derivatives(s2, {**p, "aext": mid})
+        lo, hi = (mid, hi) if a["A"] > target else (lo, mid)
+    aext = 0.5 * (lo + hi)
+    return aext, relax(s, {**p, "aext": aext})
+
+
 def align(rows: list[dict], ratio: float) -> int | None:
    """First row after the growth peak whose one-year K_Omega ratio falls to `ratio`."""
    k = int(round(1 / STEP))
@ -156,27 +174,39 @@ def main() -> None:
             f"range. Adoption scaled from anchor {BASE_ANCHOR} by k; t* aligned by observed growth.", ""]
    starts = {}
    lines += ["## Fitted adoption speed and alignment", "",
-              "| Jurisdiction | Archetype levers | k | t* (y) | K_Ω(t*) | fitted wL0 | fitted ereg |",
-              "|---|---|---|---|---|---|---|"]
+              "| Jurisdiction | Archetype levers | k | t* (y) | K_Ω(t*) | fitted wL0 | fitted ereg | fitted aext | auto |",
+              "|---|---|---|---|---|---|---|---|---|"]
    verdict = ["", "## Fixed anchors aligned by growth: K_Ω level miss", "",
               "| Jurisdiction | " + " | ".join(gr.BAND) + " |", "|---|" + "---|" * len(gr.BAND)]
    resid_lines = []
+    resid_all = {}
    for jur in ("US", "EU", "DE", "AR"):
        t = targets(jur)
        p, s = start_state(jur, BASE_ANCHOR)
        if "Ko_ratio_1y" not in t:
-            lines.append(f"| {jur} | {NEAREST[jur]} | — | — | — | {p['wL0']:.3f} | {p['ereg']:.4f} |")
+            lines.append(f"| {jur} | {NEAREST[jur]} | — | — | — | {p['wL0']:.3f} | {p['ereg']:.4f} | {p['aext']:.3f} | {p['auto']:.2f} |")
            starts[jur] = {"params": {k: p[k] for k in p if p[k] != g.PARAMS.get(k)}, "init": s, "t_star": None,
                           "note": "no measured adoption growth; adoption left at anchor B2 from its start"}
            continue
        kf, j, rows, q = fit_k(jur, p, s)
+        if "B_trend" in t:   # automation strength so that B's one-year change is on target
+            lo, hi = 0.05, 1.2
+            for _ in range(8):
+                mid = 0.5 * (lo + hi)
+                p, s = start_state(jur, BASE_ANCHOR, auto=mid)
+                kf, j, rows, q = fit_k(jur, p, s)
+                k1 = j + int(round(1 / STEP))
+                lo, hi = (lo, mid) if rows[k1]["B"] - rows[j]["B"] < centre(t["B_trend"]) else (mid, hi)
+            p, s = start_state(jur, BASE_ANCHOR, auto=0.5 * (lo + hi))
+            kf, j, rows, q = fit_k(jur, p, s)
        row = rows[j]
        lines.append(f"| {jur} | {NEAREST[jur]} | {kf:.2f} | {row['t']:.1f} | {row['Ko']:.3f} | {q['wL0']:.3f}"
-                     f" | {q['ereg']:.4f} |")
+                     f" | {q['ereg']:.4f} | {q['aext']:.3f} | {q['auto']:.2f} |")
        init = {k: row[k] for k in g.STOCKS}
        starts[jur] = {"params": {k: q[k] for k in q if q[k] != g.PARAMS.get(k)}, "init": init,
                       "t_star": row["t"], "k": kf}
        res = residuals(jur, row, q, rows, j)
+        resid_all[jur] = res
        resid_lines += [f"### {jur}", "", "| Target | Range | Model at t* | z | basis |", "|---|---|---|---|---|"]
        for name, (val, zz) in res.items():
            tt = t[name]
@ -191,6 +221,13 @@ def main() -> None:
            cells.append("cannot reach growth" if ja is None else
                         f"K_Ω {rows_a[ja]['Ko']:.3f}, z {z(rows_a[ja]['Ko'], t['Ko']):+.1f}")
        verdict.append(f"| {jur} | " + " | ".join(cells) + " |")
+    n_in = sum(1 for r in resid_all.values() for _v, zz in r.values() if abs(zz) <= 1)
+    n_all = sum(len(r) for r in resid_all.values())
+    n_meas = sum(1 for jur, r in resid_all.items() for name in r if targets(jur)[name]["basis"] == "measured")
+    n_meas_in = sum(1 for jur, r in resid_all.items() for name, (_v, zz) in r.items()
+                    if targets(jur)[name]["basis"] == "measured" and abs(zz) <= 1)
+    lines += ["", f"**Fit summary:** {n_in} of {n_all} targets inside their range; measured targets "
+              f"{n_meas_in} of {n_meas}."]
    lines += verdict + ["", "## Residuals at the calibrated start", ""] + resid_lines
    (OUT / "calibration.md").write_text("\n".join(lines))
    (OUT / "calibrated_starts.json").write_text(json.dumps(starts, indent=1, sort_keys=True))
--- a/simulations/gadm02/calibration_targets.json
+++ b/simulations/gadm02/calibration_targets.json
@ -22,10 +22,10 @@
    "op": "leading search engine share 85.2 % (Google); StatCounter Global Stats, Sept 2026, retrieved 2026-10-13 (checkpoints/2026/2026-10-13-review.md)"
   },
   "Pi": {
-    "lo": 0.5,
-    "hi": 0.72,
-    "basis": "prior",
-    "op": "infrastructure ~ top-3 cloud share (~63-66 %, Synergy Research)"
+    "lo": 0.56,
+    "hi": 0.7,
+    "basis": "measured",
+    "op": "top-3 cloud infrastructure share 28+20+15 = 63 % worldwide, Q2 2026 (Synergy Research, 2026-07-30) (checkpoints/2026/2026-10-13-review.md)"
   },
   "Pmem": {
    "lo": 0.4,
@ -34,10 +34,10 @@
    "op": "HHI of mobile OS shares (iOS 53.5, Android 46.5 %) = 0.50; operator uncertain, range ±0.10; StatCounter Global Stats, Sept 2026, retrieved 2026-10-13 (checkpoints/2026/2026-10-13-review.md)"
   },
   "Plin": {
-    "lo": 0.45,
-    "hi": 0.8,
-    "basis": "prior",
-    "op": "lineage ~ share of frontier-model usage or spend held by the top three developers"
+    "lo": 0.8,
+    "hi": 0.95,
+    "basis": "measured",
+    "op": "top-3 share of enterprise LLM API spend 40+27+21 = 88 % (Menlo Ventures, Dec 2025); US firms paying Anthropic 43.8 %, OpenAI 39.8 % (Ramp, Aug 2026) (checkpoints/2026/2026-10-13-review.md)"
   },
   "Pm": {
    "lo": 0.4,
@ -46,10 +46,10 @@
    "op": "market ~ top-firm sales concentration in digital sectors (F2)"
   },
   "B": {
-    "lo": 0.55,
-    "hi": 0.75,
-    "basis": "prior",
-    "op": "claims breadth ~ labour share plus broad transfers; F1 index 93.4 (2017=100) is measured and gives the trend"
+    "lo": 0.57,
+    "hi": 0.67,
+    "basis": "measured",
+    "op": "compensation of employees 16,350 / GDP 32,563 = 0.502 plus government social benefits 3,724 / GDP = 0.114: 0.616 (BEA via FRED COE, GDP, B087RC1Q027SBEA, 2026 Q2) (checkpoints/2026/2026-10-13-review.md)"
   },
   "B_trend": {
    "lo": -0.01,
@ -59,9 +59,9 @@
   },
   "A": {
    "lo": 0.4,
-    "hi": 0.6,
-    "basis": "prior",
-    "op": "self-directed leisure share, ATUS 2023 / DataReportal (WP-0008 grounding)"
+    "hi": 0.55,
+    "basis": "measured",
+    "op": "ATUS 2025 (BLS, 2026-06-25): leisure 5.2 h/day, TV 2.6 h (half); A = leisure share not taken by TV and engagement media, about 0.42-0.50 (checkpoints/2026/2026-10-13-review.md)"
   },
   "E": {
    "lo": 0.2,
@ -96,10 +96,10 @@
    "op": "leading search share, mean of DE 87.8, FR 88.1, IT 87.3, ES 93.8 % = 89.3 %; StatCounter Global Stats, Sept 2026, retrieved 2026-10-13 (checkpoints/2026/2026-10-13-review.md)"
   },
   "Pi": {
-    "lo": 0.55,
-    "hi": 0.8,
-    "basis": "prior",
-    "op": "top-3 (non-EU) cloud providers' share of the EU market"
+    "lo": 0.56,
+    "hi": 0.76,
+    "basis": "measured",
+    "op": "top-3 cloud infrastructure share 28+20+15 = 63 % worldwide, Q2 2026 (Synergy Research, 2026-07-30); EU market not separated, upper range widened for the low share of EU providers (checkpoints/2026/2026-10-13-review.md)"
   },
   "Pmem": {
    "lo": 0.47,
@ -108,16 +108,16 @@
    "op": "mobile OS HHI, mean of DE 0.60, FR 0.55, IT 0.54, ES 0.61 = 0.57; StatCounter Global Stats, Sept 2026, retrieved 2026-10-13 (checkpoints/2026/2026-10-13-review.md)"
   },
   "Plin": {
-    "lo": 0.45,
-    "hi": 0.85,
-    "basis": "prior",
-    "op": "mostly non-EU lineages (cross-border, WP-0011)"
+    "lo": 0.8,
+    "hi": 0.95,
+    "basis": "measured",
+    "op": "top-3 share of enterprise LLM API spend 40+27+21 = 88 % (Menlo Ventures, Dec 2025); US firms paying Anthropic 43.8 %, OpenAI 39.8 % (Ramp, Aug 2026); same foreign lineages (checkpoints/2026/2026-10-13-review.md)"
   },
   "B": {
-    "lo": 0.65,
-    "hi": 0.8,
-    "basis": "prior",
-    "op": "higher labour share and transfers than US (F1 EU proxy measured)"
+    "lo": 0.6,
+    "hi": 0.69,
+    "basis": "measured",
+    "op": "compensation D1 / GDP 0.479 (Eurostat nama_10_gdp 2025) plus social benefits D62 16.5 % of GDP (gov_10a_main 2025): 0.644 (checkpoints/2026/2026-10-13-review.md)"
   },
   "A": {
    "lo": 0.4,
@ -132,10 +132,10 @@
    "op": "as US; E2/E4 measured, not yet mapped"
   },
   "Sig": {
-    "lo": 0.38,
-    "hi": 0.55,
-    "basis": "prior",
-    "op": "Eurobarometer: majority see AI as useful but want regulation; uncalibrated"
+    "lo": 0.45,
+    "hi": 0.62,
+    "basis": "measured",
+    "op": "Special Eurobarometer 554/557 (fieldwork 2024): 55 % expect AI to affect daily life positively, 35 % negatively; Sigma = 0.5 + (pos - neg)/200 = 0.60; different question from Pew, so the range is wide and skewed down (checkpoints/2026/2026-10-13-review.md)"
   }
  },
  "DE": {
@ -162,14 +162,20 @@
    "hi": 0.7,
    "basis": "measured",
    "op": "mobile OS HHI (Android 72.3, iOS 27.7 %) = 0.60; StatCounter Global Stats, Sept 2026, retrieved 2026-10-13 (checkpoints/2026/2026-10-13-review.md)"
+   },
+   "B": {
+    "lo": 0.66,
+    "hi": 0.76,
+    "basis": "measured",
+    "op": "D1 / GDP 0.544 plus D62 16.8 %: 0.712 (Eurostat 2025) (checkpoints/2026/2026-10-13-review.md)"
   }
  },
  "AR": {
   "Ko": {
-    "lo": 0.02,
-    "hi": 0.15,
-    "basis": "prior",
-    "op": "no measured adoption (L3 manual); wide prior"
+    "lo": 0.05,
+    "hi": 0.17,
+    "basis": "measured",
+    "op": "20.4 % of firms apply AI in daily operations (Google / IDB study, Aug 2026); scaled like the US level fit (23.8 % -> 0.12) (checkpoints/2026/2026-10-13-review.md)"
   },
   "Pd": {
    "lo": 0.89,
@ -178,10 +184,10 @@
    "op": "leading search share 93.8 %; StatCounter Global Stats, Sept 2026, retrieved 2026-10-13 (checkpoints/2026/2026-10-13-review.md)"
   },
   "Plin": {
-    "lo": 0.5,
-    "hi": 0.9,
-    "basis": "prior",
-    "op": "foreign lineages only"
+    "lo": 0.8,
+    "hi": 0.95,
+    "basis": "measured",
+    "op": "top-3 share of enterprise LLM API spend 40+27+21 = 88 % (Menlo Ventures, Dec 2025); US firms paying Anthropic 43.8 %, OpenAI 39.8 % (Ramp, Aug 2026); same foreign lineages (checkpoints/2026/2026-10-13-review.md)"
   },
   "B": {
    "lo": 0.45,
@ -190,16 +196,22 @@
    "op": "wage share fell sharply in 2024 (INDEC income accounts)"
   },
   "Sig": {
-    "lo": 0.35,
-    "hi": 0.6,
-    "basis": "prior",
-    "op": "unmeasured (L6 manual); wide prior"
+    "lo": 0.36,
+    "hi": 0.5,
+    "basis": "measured",
+    "op": "La Nacion Data (June 2026, via secondary report): 30.7 % optimistic, 44.2 % sceptical; Sigma = 0.43; secondary source, range +-0.07 (checkpoints/2026/2026-10-13-review.md)"
   },
   "Pmem": {
    "lo": 0.68,
    "hi": 0.88,
    "basis": "measured",
    "op": "mobile OS HHI (Android 87.3, iOS 12.7 %) = 0.78; StatCounter Global Stats, Sept 2026, retrieved 2026-10-13 (checkpoints/2026/2026-10-13-review.md)"
+   },
+   "Ko_ratio_1y": {
+    "lo": 1.25,
+    "hi": 1.65,
+    "basis": "prior",
+    "op": "no AR series yet; assumed within the US/EU/DE observed range (1.31-1.48)"
   }
  }
 }
--- a/simulations/gadm02/results/calibrated_readings.md
+++ b/simulations/gadm02/results/calibrated_readings.md
@ -9,45 +9,45 @@ already contains today's permissiveness).

 | Regime | Loop | K_Ω t5 / t25 / peak | Ȳ 0–25 | VFAC t100 | G | A | P t0 → t100 | floors broken |
 |---|---|---|---|---|---|---|---|---|
-| J1 | off | 0.17 / 0.21 / 0.21 | 0.299 | 0.424 | 0.48 | 0.50 | 0.63 → 0.56 | none |
-| J1 | on | 0.10 / 0.25 / 0.25 | 0.297 | 0.420 | 0.47 | 0.49 | 0.63 → 0.56 | none |
-| J2 | off | 0.36 / 0.38 / 0.42 | 0.336 | 0.296 | 0.39 | 0.26 | 0.63 → 0.71 | P |
-| J2 | on | 0.18 / 0.22 / 0.26 | 0.292 | 0.341 | 0.42 | 0.32 | 0.63 → 0.66 | none |
-| J3 | off | 0.31 / 0.34 / 0.37 | 0.320 | 0.269 | 0.37 | 0.25 | 0.63 → 0.79 | P |
-| J3 | on | 0.16 / 0.18 / 0.21 | 0.277 | 0.319 | 0.41 | 0.32 | 0.63 → 0.72 | P |
-| J4 | off | 0.32 / 0.36 / 0.38 | 0.330 | 0.318 | 0.39 | 0.33 | 0.63 → 0.69 | none |
-| J4 | on | 0.16 / 0.30 / 0.30 | 0.301 | 0.338 | 0.41 | 0.35 | 0.63 → 0.66 | none |
+| J1 | off | 0.18 / 0.22 / 0.22 | 0.294 | 0.431 | 0.48 | 0.56 | 0.69 → 0.55 | none |
+| J1 | on | 0.11 / 0.27 / 0.28 | 0.292 | 0.425 | 0.47 | 0.54 | 0.69 → 0.56 | none |
+| J2 | off | 0.38 / 0.40 / 0.44 | 0.339 | 0.277 | 0.37 | 0.33 | 0.69 → 0.78 | P |
+| J2 | on | 0.19 / 0.22 / 0.27 | 0.288 | 0.311 | 0.40 | 0.40 | 0.69 → 0.76 | P |
+| J3 | off | 0.33 / 0.36 / 0.39 | 0.324 | 0.262 | 0.36 | 0.32 | 0.69 → 0.82 | P |
+| J3 | on | 0.16 / 0.19 / 0.22 | 0.275 | 0.304 | 0.39 | 0.39 | 0.69 → 0.77 | P |
+| J4 | off | 0.33 / 0.38 / 0.40 | 0.332 | 0.296 | 0.38 | 0.41 | 0.69 → 0.77 | P |
+| J4 | on | 0.16 / 0.31 / 0.31 | 0.297 | 0.307 | 0.39 | 0.43 | 0.69 → 0.76 | P |

 ## EU (current regime ≈ J1)

 | Regime | Loop | K_Ω t5 / t25 / peak | Ȳ 0–25 | VFAC t100 | G | A | P t0 → t100 | floors broken |
 |---|---|---|---|---|---|---|---|---|
-| J1 | off | 0.25 / 0.31 / 0.32 | 0.319 | 0.326 | 0.43 | 0.47 | 0.67 → 0.79 | P |
-| J1 | on | 0.22 / 0.19 / 0.25 | 0.295 | 0.347 | 0.45 | 0.53 | 0.67 → 0.76 | P |
-| J2 | off | 0.55 / 0.51 / 0.59 | 0.378 | 0.216 | 0.35 | 0.22 | 0.67 → 0.89 | P |
-| J2 | on | 0.40 / 0.13 / 0.40 | 0.278 | 0.275 | 0.41 | 0.37 | 0.67 → 0.86 | P |
-| J3 | off | 0.49 / 0.48 / 0.54 | 0.366 | 0.208 | 0.34 | 0.21 | 0.67 → 0.91 | P |
-| J3 | on | 0.35 / 0.11 / 0.35 | 0.266 | 0.264 | 0.40 | 0.36 | 0.67 → 0.88 | P |
-| J4 | off | 0.50 / 0.50 / 0.55 | 0.372 | 0.231 | 0.35 | 0.29 | 0.67 → 0.89 | P |
-| J4 | on | 0.44 / 0.34 / 0.47 | 0.338 | 0.255 | 0.38 | 0.35 | 0.67 → 0.87 | P |
+| J1 | off | 0.24 / 0.30 / 0.31 | 0.306 | 0.293 | 0.41 | 0.36 | 0.72 → 0.82 | P |
+| J1 | on | 0.26 / 0.18 / 0.28 | 0.285 | 0.311 | 0.41 | 0.42 | 0.72 → 0.80 | P |
+| J2 | off | 0.54 / 0.52 / 0.58 | 0.374 | 0.195 | 0.33 | 0.15 | 0.72 → 0.91 | P,A |
+| J2 | on | 0.46 / 0.14 / 0.47 | 0.274 | 0.248 | 0.38 | 0.26 | 0.72 → 0.88 | P |
+| J3 | off | 0.48 / 0.49 / 0.54 | 0.362 | 0.187 | 0.32 | 0.14 | 0.72 → 0.92 | P,A |
+| J3 | on | 0.41 / 0.12 / 0.41 | 0.262 | 0.239 | 0.38 | 0.25 | 0.72 → 0.90 | P |
+| J4 | off | 0.49 / 0.50 / 0.55 | 0.367 | 0.209 | 0.34 | 0.20 | 0.72 → 0.90 | P,A |
+| J4 | on | 0.51 / 0.35 / 0.53 | 0.340 | 0.230 | 0.36 | 0.24 | 0.72 → 0.89 | P |

 ## AR (current regime ≈ J2)

 | Regime | Loop | K_Ω t5 / t25 / peak | Ȳ 0–25 | VFAC t100 | G | A | P t0 → t100 | floors broken |
 |---|---|---|---|---|---|---|---|---|
-| J1 | off | 0.06 / 0.07 / 0.07 | 0.336 | 0.430 | 0.43 | 0.56 | 0.71 → 0.63 | none |
-| J1 | on | 0.06 / 0.10 / 0.11 | 0.343 | 0.424 | 0.43 | 0.54 | 0.71 → 0.63 | none |
-| J2 | off | 0.13 / 0.15 / 0.16 | 0.349 | 0.292 | 0.37 | 0.35 | 0.71 → 0.83 | P |
-| J2 | on | 0.12 / 0.10 / 0.13 | 0.332 | 0.306 | 0.38 | 0.38 | 0.71 → 0.81 | P |
-| J3 | off | 0.11 / 0.12 / 0.13 | 0.331 | 0.274 | 0.36 | 0.34 | 0.71 → 0.86 | P |
-| J3 | on | 0.10 / 0.07 / 0.11 | 0.315 | 0.286 | 0.37 | 0.37 | 0.71 → 0.85 | P |
-| J4 | off | 0.12 / 0.13 / 0.14 | 0.342 | 0.305 | 0.37 | 0.43 | 0.71 → 0.82 | P |
-| J4 | on | 0.11 / 0.12 / 0.13 | 0.338 | 0.308 | 0.37 | 0.44 | 0.71 → 0.82 | P |
+| J1 | off | 0.13 / 0.14 / 0.15 | 0.356 | 0.381 | 0.41 | 0.52 | 0.75 → 0.71 | P |
+| J1 | on | 0.12 / 0.18 / 0.18 | 0.362 | 0.377 | 0.41 | 0.51 | 0.75 → 0.72 | P |
+| J2 | off | 0.29 / 0.31 / 0.34 | 0.418 | 0.255 | 0.34 | 0.27 | 0.75 → 0.86 | P |
+| J2 | on | 0.22 / 0.16 / 0.23 | 0.363 | 0.280 | 0.36 | 0.33 | 0.75 → 0.85 | P |
+| J3 | off | 0.25 / 0.27 / 0.29 | 0.395 | 0.244 | 0.34 | 0.26 | 0.75 → 0.89 | P |
+| J3 | on | 0.19 / 0.13 / 0.20 | 0.343 | 0.266 | 0.36 | 0.32 | 0.75 → 0.87 | P |
+| J4 | off | 0.25 / 0.28 / 0.30 | 0.406 | 0.271 | 0.35 | 0.35 | 0.75 → 0.86 | P |
+| J4 | on | 0.20 / 0.23 / 0.25 | 0.382 | 0.279 | 0.35 | 0.37 | 0.75 → 0.85 | P |

 ## Readings check

 | Jurisdiction | R1 J1 best VFAC | R2 loop damps | R3 lagging costs early output (ΔȲ, ΔVFAC) | R4 current regime breaks P floor |
 |---|---|---|---|---|
-| US | holds | **fails** | holds (-0.080, +0.050) | yes |
-| EU | holds | holds | holds (-0.078, +0.029) | yes |
-| AR | holds | **fails** | holds (-0.068, +0.027) | yes |
+| US | holds | **fails** | holds (-0.088, +0.042) | yes |
+| EU | holds | holds | holds (-0.084, +0.026) | yes |
+| AR | holds | **fails** | holds (-0.121, +0.036) | yes |
--- a/simulations/gadm02/results/calibrated_starts.json
+++ b/simulations/gadm02/results/calibrated_starts.json
@ -1,58 +1,55 @@
 {
 "AR": {
  "init": {
-   "As": 0.398218911257917,
-   "Aw": 0.4100549238154302,
-   "Cch": 0.4,
-   "Cd": 0.4,
-   "Cs": 0.55,
-   "Cw": 0.55,
-   "Dch": 0.4,
-   "Dd": 0.4,
-   "Ds": 0.55,
-   "Dw": 0.55,
-   "E": 0.5,
-   "F": 0.4,
-   "G": 0.5,
-   "Gs": 0.3802001627244952,
-   "Gw": 0.4119616691879227,
-   "Ko": 0.05,
-   "Kp": 0.5,
-   "Lg": 0.475,
-   "N": 0.5,
-   "Nch": 0.2273,
-   "Nd": 0.0909,
-   "Ns": 0.1705,
-   "Nw": 0.5114,
-   "Pd": 0.935,
-   "Pi": 0.45,
-   "Plin": 0.7,
-   "Pm": 0.5,
-   "Pmem": 0.78,
-   "Ppol": 0.5,
-   "Q": 0.5,
-   "S": 0.5,
-   "Sig": 0.475,
-   "W": 0.25043547119345966,
-   "X": 0.4
+   "As": 0.39338136453238914,
+   "Aw": 0.40511361695776804,
+   "Cch": 0.4273843914823976,
+   "Cd": 0.3587407705906544,
+   "Cs": 0.5221043401339072,
+   "Cw": 0.5527861039172586,
+   "Dch": 0.4324721016696886,
+   "Dd": 0.40208384609168824,
+   "Ds": 0.5381858106668632,
+   "Dw": 0.5432637495309056,
+   "E": 0.5162507400717095,
+   "F": 0.4040863733578474,
+   "G": 0.3939397910097698,
+   "Gs": 0.3691676541584724,
+   "Gw": 0.40219866087736084,
+   "Ko": 0.11023441493770228,
+   "Kp": 0.48266382717035555,
+   "Lg": 0.43,
+   "N": 0.4663059741677877,
+   "Nch": 0.22729941048548938,
+   "Nd": 0.09090423054953169,
+   "Ns": 0.17049764194195746,
+   "Nw": 0.5114005895145106,
+   "Pd": 0.9352292554211166,
+   "Pi": 0.4510842967761467,
+   "Plin": 0.8763314787123644,
+   "Pm": 0.5007341069456409,
+   "Pmem": 0.7802943990654635,
+   "Ppol": 0.5023658343796548,
+   "Q": 0.4899755208859503,
+   "S": 0.4736980855522147,
+   "Sig": 0.43,
+   "W": 0.25085364433836677,
+   "X": 0.380525692886688
  },
-  "note": "no measured adoption growth; adoption left at anchor B2 from its start",
+  "k": 2.0,
  "params": {
   "L": 0.2,
-   "aO": 0.25,
   "aext": 0.58,
   "c_d": 0.0934,
   "c_i": 0.4354,
-   "c_lin": 0.3342,
-   "c_m": 0.083,
+   "c_lin": 0.1734,
+   "c_m": 0.0827,
   "c_mem": 0.2597,
-   "c_pol": 0.4633,
+   "c_pol": 0.4708,
   "care_load": 0.03,
-   "ell": 0.415,
   "expl": 0.2,
   "lin_open": 0.2,
   "m": 0.2,
-   "oseed": 0.0165,
   "r_d": 0.6,
   "r_i": 0.45,
   "r_lin": 0.45,
@ -60,124 +57,59 @@
   "wL0": 0.38659793814432997,
   "x_att": 0.7
  },
-  "t_star": null
+  "t_star": 1.1
 },
 "DE": {
  "init": {
-   "As": 0.5994680869112339,
-   "Aw": 0.5956745140145546,
-   "Cch": 0.4477308946229111,
-   "Cd": 0.3463065232298026,
-   "Cs": 0.5202136991854523,
-   "Cw": 0.5650065243033202,
-   "Dch": 0.4534177193202591,
-   "Dd": 0.4023715744586903,
-   "Ds": 0.5340524489690092,
-   "Dw": 0.5412381533859043,
-   "E": 0.3247890301689604,
-   "F": 0.40620515777497634,
-   "G": 0.4638249153794165,
-   "Gs": 0.44702012572352084,
-   "Gw": 0.4694274944281725,
-   "Ko": 0.11120201587003738,
-   "Kp": 0.4845529885718244,
-   "Lg": 0.465,
-   "N": 0.4682823225054386,
+   "As": 0.48980323387568897,
+   "Aw": 0.49482257426652876,
+   "Cch": 0.4441648034819571,
+   "Cd": 0.34617316016465904,
+   "Cs": 0.5184755053991729,
+   "Cw": 0.5619538231854824,
+   "Dch": 0.45222328828671904,
+   "Dd": 0.40237155735053637,
+   "Ds": 0.534051971434396,
+   "Dw": 0.5412207640198076,
+   "E": 0.3248298954736303,
+   "F": 0.40587535591155427,
+   "G": 0.44811082106136974,
+   "Gs": 0.43000444357509027,
+   "Gw": 0.45414733881795505,
+   "Ko": 0.11234398697833374,
+   "Kp": 0.4813608094712152,
+   "Lg": 0.535,
+   "N": 0.46215059748089454,
   "Nch": 0.227299207505914,
   "Nd": 0.0909055888190386,
   "Ns": 0.17049683002365584,
   "Nw": 0.5114007924940861,
-   "Pd": 0.8803752003298085,
-   "Pi": 0.6767101667605153,
-   "Plin": 0.6531140334151384,
-   "Pm": 0.5011880052136072,
-   "Pmem": 0.6002756261055299,
-   "Ppol": 0.5014127662652859,
-   "Q": 0.4938091540425548,
-   "S": 0.4921217856330394,
-   "Sig": 0.465,
-   "W": 0.14181427065953242,
-   "X": 0.3808892870931109
-  },
-  "k": 1.2968395546510096,
-  "params": {
-   "L": 0.6,
-   "aO": 0.3242098886627524,
-   "aext": 0.58,
-   "c_d": 0.1627,
-   "c_i": 0.3082,
-   "c_lin": 0.1545,
-   "c_m": 0.0675,
-   "c_mem": 0.1358,
-   "c_pol": 0.3932,
-   "care_load": 0.03,
-   "ell": 0.538188415180169,
-   "ereg": 0.010371111111111112,
-   "expl": 0.7,
-   "lin_open": 0.6,
-   "m": 0.7,
-   "oseed": 0.02139785265174166,
-   "r_d": 0.6,
-   "r_i": 0.45,
-   "r_lin": 0.45,
-   "r_mem": 0.45,
-   "std": 0.6,
-   "wL0": 0.6371706758304698,
-   "x_att": 0.3
-  },
-  "t_star": 1.5
- },
- "EU": {
-  "init": {
-   "As": 0.6022369802724169,
-   "Aw": 0.5983990586704753,
-   "Cch": 0.4296961098097199,
-   "Cd": 0.3658000277616374,
-   "Cs": 0.531114776281014,
-   "Cw": 0.5595638314034223,
-   "Dch": 0.4338969793417084,
-   "Dd": 0.40185192948881626,
-   "Ds": 0.5402876564242498,
-   "Dw": 0.5444708739472085,
-   "E": 0.32491072972305585,
-   "F": 0.40383829528416304,
-   "G": 0.46362209249277403,
-   "Gs": 0.44706936936879144,
-   "Gw": 0.46914067764474965,
-   "Ko": 0.09236214234529257,
-   "Kp": 0.49057792425006697,
-   "Lg": 0.465,
-   "N": 0.480428523544123,
-   "Nch": 0.22729951420364686,
-   "Nd": 0.09090351693347513,
-   "Ns": 0.17049805681458743,
-   "Nw": 0.5114004857963531,
-   "Pd": 0.8901481224236264,
-   "Pi": 0.6756195094254455,
-   "Plin": 0.6512409883764116,
-   "Pm": 0.5004206843711574,
-   "Pmem": 0.5701391873421384,
-   "Ppol": 0.5005199792487638,
-   "Q": 0.496121516063456,
-   "S": 0.4950179461291722,
-   "Sig": 0.465,
-   "W": 0.14153505661488275,
-   "X": 0.3881824342892735
+   "Pd": 0.8804850333571789,
+   "Pi": 0.661976211103832,
+   "Plin": 0.876740750775282,
+   "Pm": 0.50128095279319,
+   "Pmem": 0.6002811359417883,
+   "Ppol": 0.5023583760033836,
+   "Q": 0.4917127861508988,
+   "S": 0.4820035643089187,
+   "Sig": 0.535,
+   "W": 0.14836036568082947,
+   "X": 0.3778200367466775
  },
  "k": 1.4142135623730951,
  "params": {
   "L": 0.6,
   "aO": 0.3535533905932738,
-   "aext": 0.58,
-   "c_d": 0.1507,
-   "c_i": 0.3082,
-   "c_lin": 0.1545,
-   "c_m": 0.0675,
-   "c_mem": 0.1387,
-   "c_pol": 0.3933,
+   "aext": 0.9831283569335938,
+   "c_d": 0.1627,
+   "c_i": 0.3161,
+   "c_lin": 0.0741,
+   "c_m": 0.0683,
+   "c_mem": 0.1358,
+   "c_pol": 0.4045,
   "care_load": 0.03,
   "ell": 0.5868986283848344,
-   "ereg": 0.010371111111111112,
+   "ereg": 0.010267111111111112,
   "expl": 0.7,
   "lin_open": 0.6,
   "m": 0.7,
@ -187,71 +119,134 @@
   "r_lin": 0.45,
   "r_mem": 0.45,
   "std": 0.6,
-   "wL0": 0.6371706758304698,
+   "wL0": 0.6156930126002291,
+   "x_att": 0.3
+  },
+  "t_star": 1.5
+ },
+ "EU": {
+  "init": {
+   "As": 0.4937836203396411,
+   "Aw": 0.49896308458637634,
+   "Cch": 0.426294685819212,
+   "Cd": 0.3657278892931665,
+   "Cs": 0.5295695555856555,
+   "Cw": 0.5566611062469996,
+   "Dch": 0.4308098298656463,
+   "Dd": 0.4018519214238028,
+   "Ds": 0.5402872245541102,
+   "Dw": 0.5444431095847656,
+   "E": 0.32493207314438577,
+   "F": 0.4028309712113618,
+   "G": 0.4328125482458494,
+   "Gs": 0.4131728688642662,
+   "Gw": 0.4393603073192635,
+   "Ko": 0.09101991941647128,
+   "Kp": 0.4878961969619115,
+   "Lg": 0.535,
+   "N": 0.4765659398459323,
+   "Nch": 0.22729951420364686,
+   "Nd": 0.09090351693347513,
+   "Ns": 0.17049805681458743,
+   "Nw": 0.5114004857963531,
+   "Pd": 0.890186133559438,
+   "Pi": 0.6607139553167074,
+   "Plin": 0.8756873973062851,
+   "Pm": 0.5003930796435053,
+   "Pmem": 0.570135040320222,
+   "Ppol": 0.5009721717160969,
+   "Q": 0.4948158265527009,
+   "S": 0.4862857394652712,
+   "Sig": 0.535,
+   "W": 0.1754134603387214,
+   "X": 0.38616450279192865
+  },
+  "k": 1.5422108254079407,
+  "params": {
+   "L": 0.6,
+   "aO": 0.3855527063519852,
+   "aext": 0.9310928344726562,
+   "c_d": 0.1507,
+   "c_i": 0.3161,
+   "c_lin": 0.0741,
+   "c_m": 0.0735,
+   "c_mem": 0.1387,
+   "c_pol": 0.4273,
+   "care_load": 0.03,
+   "ell": 0.6400174925442954,
+   "ereg": 0.009816444444444445,
+   "expl": 0.7,
+   "lin_open": 0.6,
+   "m": 0.7,
+   "oseed": 0.025446478619231022,
+   "r_d": 0.6,
+   "r_i": 0.45,
+   "r_lin": 0.45,
+   "r_mem": 0.45,
+   "std": 0.6,
+   "wL0": 0.5226231386025201,
   "x_att": 0.3
  },
  "t_star": 0.9
 },
 "US": {
  "init": {
-   "As": 0.3807387060541207,
-   "Aw": 0.39231553791052187,
-   "Cch": 0.43207432091616804,
-   "Cd": 0.35544687293148663,
-   "Cs": 0.5204083909927092,
-   "Cw": 0.5545192473095811,
-   "Dch": 0.44021580444384073,
-   "Dd": 0.4021773452907872,
-   "Ds": 0.5371446868990992,
-   "Dw": 0.5427661062067165,
-   "E": 0.32477402360599716,
-   "F": 0.40386646797500675,
-   "G": 0.44746771325105483,
-   "Gs": 0.4273800228043242,
-   "Gw": 0.4541648103175681,
-   "Ko": 0.12085960981519464,
-   "Kp": 0.485568315370954,
+   "As": 0.4638561505217184,
+   "Aw": 0.47095929341363896,
+   "Cch": 0.4332185525265263,
+   "Cd": 0.3554804338324086,
+   "Cs": 0.5209267906156767,
+   "Cw": 0.5554727722506773,
+   "Dch": 0.43881655455166524,
+   "Dd": 0.402177333960535,
+   "Ds": 0.5371442836459209,
+   "Dw": 0.542748194853914,
+   "E": 0.32475029967213903,
+   "F": 0.4034487151448105,
+   "G": 0.4241337094438139,
+   "Gs": 0.4020491232007042,
+   "Gw": 0.4314965577581139,
+   "Ko": 0.1243134793968734,
+   "Kp": 0.484041964344501,
   "Lg": 0.285,
-   "N": 0.4750661899740872,
+   "N": 0.46840082573714836,
   "Nch": 0.22729935918753863,
   "Nd": 0.09090457861394226,
   "Ns": 0.17049743675015444,
   "Nw": 0.5114006408124614,
-   "Pd": 0.850338836992873,
-   "Pi": 0.61120172693781,
-   "Plin": 0.6282948532035832,
-   "Pm": 0.5007646038707543,
-   "Pmem": 0.5010688516394961,
-   "Ppol": 0.5012170779107693,
-   "Q": 0.4944329020485621,
-   "S": 0.48930053442912297,
+   "Pd": 0.8504430615397236,
+   "Pi": 0.6313462967709835,
+   "Plin": 0.8768277636112062,
+   "Pm": 0.5006822876704253,
+   "Pmem": 0.5011181593225171,
+   "Ppol": 0.5016418357629986,
+   "Q": 0.49220841683727995,
+   "S": 0.4828360374678574,
   "Sig": 0.285,
-   "W": 0.17364601012524727,
-   "X": 0.3834924328519065
+   "W": 0.18562318134613848,
+   "X": 0.3811184088807723
  },
-  "k": 1.8340080864093424,
+  "k": 2.0,
  "params": {
   "L": 0.1,
-   "aO": 0.4585020216023356,
-   "aext": 0.58,
+   "aext": 0.38697052001953125,
+   "auto": 0.25888671875,
   "c_d": 0.2661,
-   "c_i": 0.3904,
-   "c_lin": 0.5618,
-   "c_m": 0.0734,
+   "c_i": 0.3805,
+   "c_lin": 0.2603,
+   "c_m": 0.0757,
   "c_mem": 0.576,
-   "c_pol": 0.4114,
+   "c_pol": 0.4316,
   "care_load": 0.03,
   "disc": 0.6,
-   "ell": 0.761113355859877,
-   "ereg": 0.009851111111111113,
+   "ereg": 0.009643111111111111,
   "lin_open": 0.1,
   "m": 0.1,
-   "oseed": 0.030261133425754153,
   "r_d": 0.6,
   "r_i": 0.45,
   "r_lin": 0.45,
   "r_mem": 0.45,
-   "wL0": 0.529782359679267,
+   "wL0": 0.47841498652534076,
   "x_att": 0.8
  },
  "t_star": 1.2
--- a/simulations/gadm02/results/calibration.md
+++ b/simulations/gadm02/results/calibration.md
@ -6,20 +6,23 @@ range. Adoption scaled from anchor B2 by k; t* aligned by observed growth.

 ## Fitted adoption speed and alignment

-| Jurisdiction | Archetype levers | k | t* (y) | K_Ω(t*) | fitted wL0 | fitted ereg |
-|---|---|---|---|---|---|---|
-| US | J3 | 1.83 | 1.2 | 0.121 | 0.530 | 0.0099 |
-| EU | J1 | 1.41 | 0.9 | 0.092 | 0.637 | 0.0104 |
-| DE | J1 | 1.30 | 1.5 | 0.111 | 0.637 | 0.0104 |
-| AR | J2 | — | — | — | 0.387 | 0.0500 |
+| Jurisdiction | Archetype levers | k | t* (y) | K_Ω(t*) | fitted wL0 | fitted ereg | fitted aext | auto |
+|---|---|---|---|---|---|---|---|---|
+| US | J3 | 2.00 | 1.2 | 0.124 | 0.478 | 0.0096 | 0.387 | 0.26 |
+| EU | J1 | 1.54 | 0.9 | 0.091 | 0.523 | 0.0098 | 0.931 | 0.60 |
+| DE | J1 | 1.41 | 1.5 | 0.112 | 0.616 | 0.0103 | 0.983 | 0.60 |
+| AR | J2 | 2.00 | 1.1 | 0.110 | 0.387 | 0.0500 | 0.580 | 0.60 |
+
+**Fit summary:** 39 of 39 targets inside their range; measured targets 28 of 28.

 ## Fixed anchors aligned by growth: K_Ω level miss

 | Jurisdiction | B1 | B2 | B3 | B4 |
 |---|---|---|---|---|
-| US | cannot reach growth | K_Ω 0.059, z -0.9 | K_Ω 0.135, z +0.1 | K_Ω 0.230, z +1.4 |
-| EU | cannot reach growth | K_Ω 0.062, z -0.6 | K_Ω 0.142, z +0.9 | K_Ω 0.252, z +2.9 |
-| DE | cannot reach growth | K_Ω 0.080, z -0.5 | K_Ω 0.188, z +1.1 | K_Ω 0.384, z +4.1 |
+| US | cannot reach growth | K_Ω 0.055, z -0.9 | K_Ω 0.124, z -0.0 | K_Ω 0.209, z +1.1 |
+| EU | cannot reach growth | K_Ω 0.055, z -0.7 | K_Ω 0.125, z +0.5 | K_Ω 0.201, z +1.9 |
+| DE | cannot reach growth | K_Ω 0.074, z -0.6 | K_Ω 0.171, z +0.9 | K_Ω 0.347, z +3.6 |
+| AR | cannot reach growth | cannot reach growth | K_Ω 0.110, z +0.0 | K_Ω 0.163, z +0.9 |

 ## Residuals at the calibrated start

@ -27,16 +30,16 @@ range. Adoption scaled from anchor B2 by k; t* aligned by observed growth.

 | Target | Range | Model at t* | z | basis |
 |---|---|---|---|---|
-| Ko | 0.05–0.2 | 0.121 | -0.06 | prior |
-| Ko_ratio_1y | 1.3–1.6 | 1.434 | -0.11 | measured |
+| Ko | 0.05–0.2 | 0.124 | -0.01 | prior |
+| Ko_ratio_1y | 1.3–1.6 | 1.447 | -0.02 | measured |
 | Pd | 0.8–0.9 | 0.850 | +0.01 | measured |
-| Pi | 0.5–0.72 | 0.611 | +0.01 | prior |
+| Pi | 0.56–0.7 | 0.631 | +0.02 | measured |
 | Pmem | 0.4–0.6 | 0.501 | +0.01 | measured |
-| Plin | 0.45–0.8 | 0.628 | +0.02 | prior |
+| Plin | 0.8–0.95 | 0.877 | +0.02 | measured |
 | Pm | 0.4–0.6 | 0.501 | +0.01 | prior |
-| B | 0.55–0.75 | 0.634 | -0.16 | prior |
-| B_trend | -0.01–0.0 | -0.012 | -1.40 ⚠ | measured |
-| A | 0.4–0.6 | 0.389 | -1.11 ⚠ | prior |
+| B | 0.57–0.67 | 0.613 | -0.13 | measured |
+| B_trend | -0.01–0.0 | -0.005 | +0.01 | measured |
+| A | 0.4–0.55 | 0.469 | -0.08 | measured |
 | E | 0.2–0.45 | 0.325 | -0.00 | prior |
 | Sig | 0.22–0.35 | 0.285 | +0.00 | measured |

@ -44,28 +47,40 @@ range. Adoption scaled from anchor B2 by k; t* aligned by observed growth.

 | Target | Range | Model at t* | z | basis |
 |---|---|---|---|---|
-| Ko | 0.04–0.15 | 0.092 | -0.05 | prior |
-| Ko_ratio_1y | 1.35–1.6 | 1.458 | -0.13 | measured |
+| Ko | 0.04–0.15 | 0.091 | -0.07 | prior |
+| Ko_ratio_1y | 1.35–1.6 | 1.449 | -0.21 | measured |
 | Pd | 0.84–0.94 | 0.890 | +0.00 | measured |
-| Pi | 0.55–0.8 | 0.676 | +0.00 | prior |
+| Pi | 0.56–0.76 | 0.661 | +0.01 | measured |
 | Pmem | 0.47–0.67 | 0.570 | +0.00 | measured |
-| Plin | 0.45–0.85 | 0.651 | +0.01 | prior |
-| B | 0.65–0.8 | 0.713 | -0.16 | prior |
-| A | 0.4–0.6 | 0.599 | +0.99 | prior |
+| Plin | 0.8–0.95 | 0.876 | +0.01 | measured |
+| B | 0.6–0.69 | 0.636 | -0.21 | measured |
+| A | 0.4–0.6 | 0.498 | -0.02 | prior |
 | E | 0.2–0.45 | 0.325 | -0.00 | prior |
-| Sig | 0.38–0.55 | 0.465 | +0.00 | prior |
+| Sig | 0.45–0.62 | 0.535 | +0.00 | measured |

 ### DE

 | Target | Range | Model at t* | z | basis |
 |---|---|---|---|---|
-| Ko | 0.05–0.18 | 0.111 | -0.06 | prior |
-| Ko_ratio_1y | 1.2–1.45 | 1.312 | -0.10 | measured |
+| Ko | 0.05–0.18 | 0.112 | -0.04 | prior |
+| Ko_ratio_1y | 1.2–1.45 | 1.316 | -0.08 | measured |
 | Pd | 0.83–0.93 | 0.880 | +0.01 | measured |
-| Pi | 0.55–0.8 | 0.677 | +0.01 | prior |
+| Pi | 0.56–0.76 | 0.662 | +0.02 | measured |
 | Pmem | 0.5–0.7 | 0.600 | +0.00 | measured |
-| Plin | 0.45–0.85 | 0.653 | +0.02 | prior |
-| B | 0.65–0.8 | 0.708 | -0.22 | prior |
-| A | 0.4–0.6 | 0.597 | +0.97 | prior |
+| Plin | 0.8–0.95 | 0.877 | +0.02 | measured |
+| B | 0.66–0.76 | 0.693 | -0.33 | measured |
+| A | 0.4–0.6 | 0.494 | -0.06 | prior |
 | E | 0.2–0.45 | 0.325 | -0.00 | prior |
-| Sig | 0.38–0.55 | 0.465 | +0.00 | prior |
+| Sig | 0.45–0.62 | 0.535 | +0.00 | measured |
+
+### AR
+
+| Target | Range | Model at t* | z | basis |
+|---|---|---|---|---|
+| Ko | 0.05–0.17 | 0.110 | +0.00 | measured |
+| Pd | 0.89–0.98 | 0.935 | +0.01 | measured |
+| Plin | 0.8–0.95 | 0.876 | +0.02 | measured |
+| B | 0.45–0.65 | 0.540 | -0.10 | prior |
+| Sig | 0.36–0.5 | 0.430 | +0.00 | measured |
+| Pmem | 0.68–0.88 | 0.780 | +0.00 | measured |
+| Ko_ratio_1y | 1.25–1.65 | 1.443 | -0.03 | prior |
--- a/workplans/ASSEMBLY-WP-0011-cross-border-coupling.md
+++ b/workplans/ASSEMBLY-WP-0011-cross-border-coupling.md
@ -35,11 +35,12 @@ Goal: two scenario modes on one model.
 This workplan was revised against the calibrated, partly measured start
 (`models/Calibration.md` v0.2, `results/calibrated_starts.json`):

- Discovery and membrane concentration are measured: leading search share
-  0.85–0.94, mobile-OS HHI 0.50–0.78. Both channels are run by foreign
-  operators everywhere except the US.
- Starting domination is US 0.63, EU 0.67, AR 0.71. Argentina starts above
-  the floor.
+- Concentration is measured (Calibration v0.3, 28 of 28 measured targets
+  fitted): leading search share 0.85–0.94, mobile-OS HHI 0.50–0.78, cloud
+  top-3 share 0.63, frontier-model top-3 spend 0.88. Outside the US, all of
+  these channels are run by foreign operators.
+- Starting domination is US 0.69, EU 0.72, AR 0.75: at or above the floor
+  (0.70) everywhere.
 - BR-007: the EU, already at rights-first levers, cannot bring P down. Its
  domination is set abroad. That is the acceptance test for this workplan:
  the coupled model must show whether, and through which channel, EU or US
@ -47,7 +48,7 @@ This workplan was revised against the calibrated, partly measured start
 - Adoption: present between B2 and B3; fitted speed k 1.3–1.8 × B2.

 All runs start from `calibrated_starts.json`. WP-0015 continues to replace
-the remaining priors (Pi, Plin); re-run T05 after each re-calibration.
+the remaining priors (Pm, E, EU/DE attention, AR B and growth); re-run T05 after each re-calibration.

 ## Cross-border channels: inventory and evidence

--- a/workplans/ASSEMBLY-WP-0012-model-matching-flipping.md
+++ b/workplans/ASSEMBLY-WP-0012-model-matching-flipping.md
@ -24,8 +24,8 @@ jurisdiction or society operates under has changed significantly.

 ## Calibrated starting point (set up 2026-10-13, after calibration)

-The calibration (`models/Calibration.md` v0.2) gives the first matching
-evidence:
+The calibration (`models/Calibration.md` v0.3: 39 of 39 targets in range,
+28 measured) gives the first matching evidence:

 - Aligned by measured growth, **B1 cannot reproduce today's adoption growth**
  (US, EU, DE). B2 lands slightly low, B3 fits, and B4 overshoots the level
--- a/workplans/ASSEMBLY-WP-0015-first-checkpoint-review.md
+++ b/workplans/ASSEMBLY-WP-0015-first-checkpoint-review.md
@ -35,7 +35,7 @@ priority: high
 state_hub_task_id: "cb9cc96b-ab9c-5df1-956a-d168efce1fd6"
 ```

-**Partly done 2026-10-13:** L6 US (Pew 2026 and 2021) in `checkpoints/2026/2026-10-13-review.md`. L3, L4, L7 and EU/AR sentiment are open.
+**Partly done 2026-10-13:** L6 US, EU, AR and L3 AR (level) in `checkpoints/2026/2026-10-13-review.md`. L4 and L7 are open, and secondary-source figures need checking against the originals.

 For each indicator, record the value with source, reference period and
 retrieval date in a review record `checkpoints/2026/<date>-review.md`:
@ -63,7 +63,7 @@ priority: high
 state_hub_task_id: "7533595c-21cc-5526-84fc-a33a1d1adc90"
 ```

-**Partly done 2026-10-13:** Pd and Pmem (StatCounter) for US, EU proxy, DE, AR, and US Σ are now `measured`. Pi, Plin, Pm, A, B, E are open.
+**Partly done 2026-10-13:** 28 targets measured (Pd, Pmem, Pi, Plin, B, US A, Σ, AR adoption level). Open: Pm, E, EU/DE attention, AR B, AR adoption growth.

 Replace `prior` targets in `simulations/gadm02/calibration_targets.json` with
 measured ranges, citing source and date in `op`:
@ -83,12 +83,12 @@ level). Do not narrow it because data exists.

 ```task
 id: ASSEMBLY-WP-0015-T03
-status: progress
+status: done
 priority: high
 state_hub_task_id: "0813f5cd-9ba1-5530-a2e5-4a66f3d98610"
 ```

-**Partly done 2026-10-13:** re-calibrated with the measured channels; verdict and R1–R4 unchanged (`models/Calibration.md` § 3.1). Repeat after the remaining T02 values.
+**Done 2026-10-13:** Calibration v0.3, 39 of 39 targets in range; `aext` and `auto` fitted in addition. Verdict and R1–R4 unchanged. Later updates follow `models/Calibration.md` § 5.

 Follow `models/Calibration.md` § 5. Re-run `calibrate_data.py` and
 `calibrated_readings.py`. Compare k, t*, the anchor verdict (today: between