Let me analyze the situation carefully.

The Mind's core question: Do universities grow or shrink due to AI? Current facet: **Cost structure — AI reduces need for faculty/TA, tuition pressure.**

Notes already filed include a comprehensive consolidation (Note 4/7) with a cost-structure inventory:
- "AI cuts teaching payroll first; admin layer captures savings"
- "Marginal-cost collapse: AI tutoring/grading pushes per-student instruction cost toward near zero"
- "Brand-strong institutions scale enrollment without new buildings (Georgia Tech OMSCS model)"
- "Cost-plus pricing dies at the bottom: AI-native providers at ~1/10th price force unselective regionals into price deflation; selective flagships hold or raise sticker because price signals scarcity"
- "Teaching labor only ~25-30% of operating cost; binding costs are fixed (plant, admin, athletics, debt service)"
- "Cost structure bifurcates the system faster than it shrinks it"

The session substance provides:
- Cost map sized: instructional labor ~25-30% of operating spend (already filed!)
- The ~70% untouched by AI — already implied by "binding costs are fixed"
- The NEW claim: **attrition timing** as the one lever that makes untouched costs movable. AI compresses per-student instructional cost only when vacated faculty lines are NOT refilled. This resolves regionally: declining-region (Northeast/Midwest) systems don't refill vacated lines → shed 1-3% instructional headcount/year through the 2030s → genuinely shrink; Sun Belt/SEC systems refill every line and monetize AI savings as expansion capacity. Net: AI never lowers tuition anywhere; same efficiency monetized as headcount shrinkage (declining regions) vs. growth capacity (growth regions). Sector total size roughly flat while geography rotates south.

Now, what's already filed vs. new?

Filed:
- Instructional labor ~25-30% of operating cost; fixed costs bind — FILED (this is restated in the session substance as setup)
- "AI cuts teaching payroll first; admin layer captures savings" — FILED. The session claim "AI never lowers tuition anywhere" is related to "savings can't reach sticker price" — Note 8's analysis mentioned a previous claim established that "savings can't reach sticker price" and "Tuition pressure then comes from competition, not cost cuts." So "AI never lowers tuition anywhere" is essentially already filed as "admin layer captures savings" + "savings can't reach sticker price."
- Regional asymmetry (Northeast/Midwest decline, Sun Belt/SEC growth) — FILED extensively.
- "Cost structure bifurcates the system faster than it shrinks it" — FILED.
- Growth-state flagships get smooth landing, re-architect before pincer — partially filed (Note 8 analysis mentions "Growth-state flagships (SEC footprint, TX, FL, AZ) get smooth landing: demand tailwinds + legislative capital to re-architect before pincer closes.")

NEW in this session:
1. **Attrition timing as the specific mechanism**: AI savings are realized only when vacated faculty lines are NOT refilled. This is a distinct mechanism — it converts a static cost observation into a dynamic hiring-decision claim. Not filed.
2. **The same efficiency monetized two ways**: headcount shrinkage in declining regions vs. growth capacity in growth regions — the sector's total size stays roughly flat while geography rotates south. This is a sharpened synthesis: efficiency doesn't shrink the sector, it relocates it. Related to "Cost structure bifurcates the system faster than it shrinks it" but adds the specific "flat total, rotating geography" claim and the attrition mechanism.
3. **Quantified: declining-region systems shed 1-3% of instructional headcount per year through the 2030s.** New number (Unverified).

Is this durable and non-duplicative? I think yes — the attrition-timing mechanism and the "flat total / rotating geography" resolution is a new, falsifiable, dated claim. It's not merely restating regional asymmetry; it explains HOW AI cost savings express regionally: via refill-or-not hiring decisions.

The tension to acknowledge: "sector's total size stays roughly flat" — but the filed consolidation says total enrollment drifts ~19M → ~14-15M by 2045 and "modest national shrink (~10-15%)". Hmm. The session claim says "the sector's total size stays roughly flat while its geography rotates south." Is this in tension with the filed enrollment decline? Not exactly — the filed decline is driven by demography and credential value, not by AI cost savings. The session claim is specifically about what AI *cost efficiency* does to sector size: it doesn't shrink the sector (that's demography/credential value's job); it reallocates capacity.