I need to synthesize the durable claims from all notes:

Core answer:
- Both — the system bifurcates and thins; not uniform growth or shrink
- Total enrollment drifts ~19M → ~14–15M by 2045 (Unverified)
- Institution count falls while degrees conferred stay roughly steady — "thins, doesn't shrink"
- Median campus roughly halves
- AI is an accelerant/sorter of existing trends, not the primary cause
- Two-front squeeze: fewer 18-year-olds AND weakened degree signal
- Barbell: system shrinks at the median while top decile grows

Demography (sourced):
- Births 4.32M (2007) → ~3.59M (2023), ~17% decline (CDC/NCHS)
- HS grads peak ~3.9M/yr ~2025 (WICHE "Knocking at the College Door")
- Birth dearth hits college age 2029–2037
- NE/Midwest/Plains lose 10–20% college-age cohorts (WICHE); TX/FL/TN/GA/Carolinas grow
- WICHE ~13% decline in HS grads by 2041
- Net domestic migration to Sun Belt (Census Bureau)

Regional/tier structure:
- SEC-corridor applied flagships grow 30–50% headcount by ~2045
- Sun Belt gains across tiers (+10–25% seats through ~2040)
- Nonselective Northeast/Midwest contracts roughly one-third; closures concentrated PA, OH, IL, New England
- Tier-moderated not geography-absolute
- Real dividing line: endowment depth (privates) + state-subsidy commitment (publics); region sets the demand curve
- Intra-state bifurcation
- By 2045 regional labels matter less than institutional balance sheets
- Casualties in "sufficiency band" below ~60% occupancy
- Program mix: licensed vs unlicensed share of credit hours = dominant predictor; licensed expansion takes 7–10 years
- Three-exposure overlap
- SEC/Sunbelt growth is share-absorption, not sector expansion
- Flagship-selective publics hold longest — until 2035–2045 window

Credential value / signaling:
- Screening vs insurance decomposition
- Degree survives as trust/insurance credential
- Selection-based signals AI-proof
- Signaling migrates upward to selectivity/brand
- Licensed/statute-gated fields hold; licensure pins ~20% of US employment (Unverified)
- Erosion hits by major before institution
- Scissors/youngest edge: new-grad premium flattens first
- Non-selective sell labor-market access → elastic; selective bundle network/sorting → demand rises

Cost structure:
- AI lowers cost of instruction, not cost of the university
- Tuition pass-through asymmetry: savings become margin/discounts at top; deflation at bottom by ~2035
- Labor-layer segmentation: contingent layer first, tenured line last
- Scale machine vs cost-cutting machine — same AI, opposite uses
- Premium tier vs commodity tier bifurcation
- Admin compression + statehouse capture: savings captured by legislatures, not tuition cuts

Research mission:
- Bifurcates along research-vs-teaching axis (also filed as Note 6/10)
- ~100–150 R1 research hubs; teaching-only institutions shrink
- Academic medical centers stealth winners
- Research moat is domain-stratified

Demand shift:
- Field-selective enrollment: AI-complement majors (health professions, engineering, CS) rise from ~25% toward 40%+ of degrees
- Generic content majors fall to minority share at non-selectives
- BA bundle decomposes: sorting signal dies, content delivery free; licensure pathways and elite networks survive as premium
- Program-level pricing: licensure tracks price up 2–3x real, generic majors fall toward delivery cost
- Vendor certification falsifier: if Google/AWS certs win, university enrollment growth becomes a health-workforce story

Signal mechanism:
- Signaling value tracks acquisition cost, not content — cost-of-faking mechanism
- Scissors: elite premium widens, mass premium compresses toward zero; bimodal wage premium distribution
- Degree survives as trust/insurance credential — AI can assay competence but not diligence/trust

The consolidation instruction: "Revise the notes as a set: keep every claim that still holds, drop repetition, sharpen wording. Output one Markdown note that is the current best statement of what is established. Start with a # title, then short paragraphs, then: What