The user wants one Markdown note synthesizing the filed notes. Let me identify the durable claims and the refinements:

1. Bifurcation claim (Notes 1&2): By 2045, US higher ed bifurcates rather than uniformly shrinks — top ~100 grow, middle hollows into teaching-only, bottom third merges/closes. Regional decline tracks demographic map more than AI.
2. Demography is the load-bearing wall: HS grads peak ~3.9M around 2025–26; post-2008 birth dearth hits college age 2029–2037. Birth data: ~4.32M (2007) → ~3.59M (2023), ~17% decline (CDC/NCHS). HS grad projections (WICHE). Northeast/Midwest/Plains lose 10–20% college-age population; Texas/Florida/Tennessee/Georgia/Carolinas grow.
3. AI is accelerant/sorter, not primary cause. Two-front squeeze: fewer 18-year-olds AND degree as signal weakened (skills directly verifiable). Unbundles campus-credential first at mid-tier.
4. Regional map: hardest hit = Northeast/Midwest small privates + open-enrollment regional publics; most resilient = SEC/Sun Belt flagships + super-selectives (prestige-elastic, not cohort-elastic).
5. Program-mix × region refinement (Notes 5&6 — most evolved): program mix (licensed vs unlicensed share of credit hours) picks winners; regional demography picks which losers can still switch. Licensed expansion takes 7–10 years (clinical placements, faculty pipelines, facilities, state approvals) — longer than declining-region enrollment runway → decline locks in. Projections: Sunbelt/SEC publics +10–25% seats through 2040; Northeast small privates without licensed core −30–50% by 2040; surviving Northeast exceptions are licensed-core with hospital-system/land-grant ties.
6. Tier overlay: middle loses research identity first, becomes teaching-only (most AI-exposed). 2045 university: fewer, bigger, regionalized; research engine with experience-economy shell; credentialing as trust layer; teaching split prestige faculty + model-supervised adjunct tier. Region works through tier.
7. Super-selectives: prestige-elastic, not cohort-elastic.
8. Adaptations/watch: admit-standard relaxation, international recruitment contingent on visa policy, endowment repurposing tell (from Note 3).
9. Kill-switch/falsification: Note 1's — if selective privates AND flagships both show flat/declining yields and international demand despite AI tools, the degree itself is devaluing and bifurcation is wrong (deeper disruption). Note 6's — if Northeast pivot attempts succeed at rates comparable to Sunbelt licensed launches, pivot-lag mechanism is wrong; decline is governance-fixable, not structural.

The instruction says the note should be the current best statement of what is established. The most evolved claim combines bifurcation + program-mix/region interaction. I should keep one title line that captures the sharpest synthesized claim.

Title options: "By 2045 US higher education bifurcates rather than shrinks — and region governs which schools can still pivot: program mix picks the winners, demography picks which losers can switch."

Maybe: "# By 2045 US higher education bifurcates rather than shrinks: program mix (licensed vs. unlicensed) picks the winners, and regional demography picks which losers can still switch — top tier grows, middle hollows to teaching-only, bottom third merges or closes (assessment as of 2025)"

That's long but the notes' titles were long too. Keep it as one line.

Structure:
- Title
- Para 1: demography load-bearing wall with numbers + sources. Births 4.32M (2007) → 3.59M (2023), ~17% (CDC/NCHS). HS grads peak ~3.9M ~2025 (WICHE). Regional: Northeast/Midwest/Plains −10–20%; TX/FL/TN/GA/Carolinas growing. Answer to grow/shrink is regional before AI.
- Para 2: AI's role — accelerant/sorter; two-front squeeze: fewer 18-year-olds + degree as signal weakened (skills directly verifiable). Unbundles mid-tier first. Super-selectives prestige-elastic.
- Para 3: program mix × region mechanism — licensed (nursing/engineering/allied health) is the adaptive path; pivot takes 7–10 years > enrollment runway of declining-region schools → decline locks in. Projections: Sunbelt/SEC publics +10–25% seats through ~2040; Northeast small privates without licensed core −30–50% by 2040; exceptions licensed-core with hospital/land-grant ties. Formulation line.
-