Review fifth uncertain AI cohort and distinguish ARR from service and contract claims

apps/docs/scripts/auth-cli.ts

@@ -182,10 +182,10 @@ async function main() {
   assert.equal(researchResponse.status, 200, "authenticated research API");
   assert.ok(researchResponse.headers.get("cache-control")?.includes("no-store"));
   const research = await researchResponse.json();
-  assert.equal(research.companies.length, 1102);
+  assert.equal(research.companies.length, 1115);
   assert.equal(research.categories.length, 11);
-  assert.equal(research.evidence.length, 2973);
-  assert.equal(research.observations.length, 395);
+  assert.equal(research.evidence.length, 3047);
+  assert.equal(research.observations.length, 413);
   assert.equal((await request("/api/research/ai-companies", { method: "POST", headers })).status, 405);
   const researchPage = await request("/docs/other-research/ai-company-research/overview", { headers });
   assert.equal(researchPage.status, 200);

apps/docs/tests/research.test.ts

@@ -23,9 +23,9 @@ test("research revisions share the authentication database and preserve replay l
     for (let i = 0; i < 20; i++) consumeCode(config, "invalid");
     assert.equal(consumeCode(config, "invalid"), "limited");
     const initial = readResearchSnapshot();
-    assert.equal(initial.companies.length, 1102);
-    assert.equal(initial.evidence.length, 2973);
-    assert.equal(initial.observations.length, 395);
+    assert.equal(initial.companies.length, 1115);
+    assert.equal(initial.evidence.length, 3047);
+    assert.equal(initial.observations.length, 413);
     assert.deepEqual(initial.candidateBatches, seed.candidateBatches, "candidate gaps persist in the same database metadata");
     assert.deepEqual(initial.directoryRechecks, seed.directoryRechecks, "follow-up classifications persist without rewriting the discovery audit");
     assert.deepEqual(readResearchSnapshot(), initial, "reopening does not duplicate records");
@@ -64,8 +64,8 @@ test("research validation rejects broken evidence and unsafe source links", () =
 
 test("reviews link explicit companies without turning global or excluded cases into companies", () => {
   const snapshot = structuredClone(seed) as ResearchSnapshot;
-  assert.equal(snapshot.reviews.filter((r) => r.scope === "companies").length, 182);
-  assert.equal(snapshot.reviews.filter((r) => r.scope === "dataset").length, 16);
+  assert.equal(snapshot.reviews.filter((r) => r.scope === "companies").length, 186);
+  assert.equal(snapshot.reviews.filter((r) => r.scope === "dataset").length, 17);
   assert.equal(snapshot.reviews.filter((r) => r.scope === "excluded-candidate").length, 44);
   const review = snapshot.reviews.find((r) => r.id === "R4")!;
   assert.deepEqual(reviewCompanyIds(snapshot, review), ["AI007", "AI013", "AI021", "AI034"]);
@@ -85,11 +85,11 @@ test("reviews link explicit companies without turning global or excluded cases i
 test("disclosure counts exclude client income, biographies, targets and duplicate milestones", () => {
   const snapshot = seed as ResearchSnapshot;
   const summary = categorySummary(snapshot);
-  assert.equal(summary.reduce((n, c) => n + c.count, 0), 1102);
-  assert.equal(summary.reduce((n, c) => n + c.self, 0), 53);
+  assert.equal(summary.reduce((n, c) => n + c.count, 0), 1115);
+  assert.equal(summary.reduce((n, c) => n + c.self, 0), 56);
   assert.equal(summary.reduce((n, c) => n + c.estimate, 0), 2);
   assert.equal(summary.reduce((n, c) => n + c.platform, 0), 19);
-  assert.equal(summary.reduce((n, c) => n + c.unknown, 0), 1028);
+  assert.equal(summary.reduce((n, c) => n + c.unknown, 0), 1038);
   assert.equal(snapshot.companies.find((c) => c.name === "Rork")?.revenueStatus, "unknown");
   assert.equal(snapshot.companies.find((c) => c.name === "Result")?.revenueStatus, "unknown");
   assert.ok(snapshot.observations.filter((o) => o.selection === "signal").every((o) => o.scope === "公司整体"));
@@ -124,7 +124,7 @@ test("service run rates, legacy freight revenue and paid adoption do not become
 
 test("technical reviews keep accessible materials separate from tested capabilities", () => {
   const snapshot = structuredClone(seed) as ResearchSnapshot;
-  assert.equal(snapshot.companies.filter((c) => c.technicalReview).length, 98);
+  assert.equal(snapshot.companies.filter((c) => c.technicalReview).length, 104);
   assert.equal(snapshot.companies.filter((c) => c.technicalReview?.level === "public-artifact").length, 21);
   assert.equal(snapshot.companies.filter((c) => c.technicalReview?.level === "independent-test").length, 1);
   assert.equal(snapshot.companies.find((c) => c.id === "AI051")?.revenueStatus, "unknown");
@@ -578,6 +578,42 @@ test("the fourth uncertain cohort distinguishes product pivots, managed labor an
   assert.ok(!platformIncomeSample(snapshot, { windowDays: null }).points.some((p) => p.observation.id.startsWith("UNCLEAR4-")));
 });
 
+test("the fifth uncertain cohort separates dated ARR, service run rates, bookings and previous ventures", () => {
+  const snapshot = seed as ResearchSnapshot;
+  const batch = snapshot.candidateBatches!.find((b) => b.id === "YC-AI-UNCLEAR-5")!;
+  const original = coverageAudit.records.filter((r) => r.status === "ai-unclear").slice(96, 120);
+  assert.equal(batch.complete, false);
+  assert.deepEqual(batch.records.map((r) => r.url), original.map((r) => r.source));
+  assert.equal(batch.records.filter((r) => r.status === "linked").length, 15);
+  assert.equal(batch.records.filter((r) => r.status === "pending-ai").length, 9);
+  assert.equal(batch.records.find((r) => r.name === "primitive")!.companyId, "AI002");
+  assert.equal(batch.records.find((r) => r.name === "Result")!.companyId, "AI001");
+  const result = snapshot.companies.find((c) => c.id === "AI001")!;
+  assert.equal(result.cat, "C09");
+  assert.equal(result.revenueStatus, "unknown");
+  assert.match(result.notes, /280万美元ARR属于用户/);
+  assert.ok(result.evidenceIds.includes("UNCLEAR5-31438-TERMS"));
+  const andco = snapshot.companies.find((c) => c.id === "YC31411")!;
+  assert.equal(andco.year, 2026, "a founder's previous venture must not replace the new company's stated year");
+  assert.match(andco.notes, /Layup/);
+  assert.equal(snapshot.evidence.find((e) => e.id === "UNCLEAR5-31411-LAUNCH6")!.companyId, null);
+  const figures = snapshot.observations.filter((o) => o.id.startsWith("UNCLEAR5-"));
+  assert.equal(figures.length, 18);
+  const dated = figures.find((o) => o.id.endsWith("TRELLIS-ARR-AUG"))!;
+  assert.equal(dated.metric, "arr");
+  assert.equal(dated.observedAt, "2026-08-24");
+  assert.equal(dated.bound, "≥");
+  assert.equal(figures.find((o) => o.id.endsWith("INSTA-RUNRATE"))!.metric, "annualized-revenue");
+  assert.equal(figures.find((o) => o.id.endsWith("INSTA-LAUNCH-ARR"))!.metric, "arr");
+  for (const suffix of ["PARROT-CONTRACT", "KINRO-OLD-CONTRACT", "BENTO-OLD-EMPLOYER", "TRELLIS-OLD-MGMT"]) {
+    assert.equal(figures.find((o) => o.id.endsWith(suffix))!.selection, "excluded");
+  }
+  assert.equal(snapshot.companies.find((c) => c.id === "YC31423")!.technicalReview!.level, "description",
+    "an internal benchmark subset is not an independent leaderboard result");
+  assert.ok(figures.every((o) => o.comparisonGroup === null));
+  assert.ok(!platformIncomeSample(snapshot, { windowDays: null }).points.some((p) => p.observation.id.startsWith("UNCLEAR5-")));
+});
+
 test("Chinese discovery entries preserve unread articles and use separate primary evidence", () => {
   const snapshot = structuredClone(seed) as ResearchSnapshot;
   const batch = snapshot.candidateBatches!.find((b) => b.id === "JQ-2026-targeted-1")!;