Let me parse this carefully.

The structure:
- Core question: Do universities grow or shrink due to AI, over 20 years, regionally dependent?
- Current facet: Enrollment demand shift (credential value vs. AI skills)
- Notes already filed: Notes 1-8 (various — some are meta/truncated planning, some are filed notes)
- Substance from this session: One durable claim on the current facet — the demand shift is not fewer students but different buyers; enrollment migrates from degree-as-credential to degree-as-verified-skill; generic diploma-signal programs (undifferentiated business, communications, generic psychology) lose demand and price tolerance first because AI removes the entry-level tasks those credentials promised; licensure-anchored and applied fields (nursing, engineering, construction/agriculture, data-adjacent trades) hold or gain because they certify scarce, AI-resistant competence; tuition pricing power redistributes to programs showing verified outcomes within ~2 years of entry; flat-tuition, credential-first institutions absorb the demand loss regardless of region — this is the mechanism layer beneath the regional bifurcation already claimed.

Instructions:
- Write one Markdown note: a single dated claim this Mind can stand behind.
- Start with a # title that is the claim in one line.
- Then 1-3 short paragraphs.
- End with: What would change this: …
- Name a source for any load-bearing number or fact, or start that sentence with Unverified:
- Do not mention tools, searches, sessions, passes, or that you are writing a note.
- Do not invent evidence.
- If the session is only planning, output nothing. (The session has substance, so I write.)

Key considerations:

1. The current facet is enrollment demand shift (credential value vs. AI skills). Note 8 in the filed set is already about enrollment bifurcation by region ("grow or shrink resolves as 'both'"). Note 7's substance was about degree-as-insurance. So the new claim must add beyond those: the mechanism is program-level, not institution-level — demand migrates from credential-signal programs to verified-skill programs. It's the mechanism layer beneath the regional bifurcation.

2. Load-bearing facts I might cite:
- The claim mentions entry-level tasks being removed by AI. There's real evidence: the Stanford Digital Economy Lab paper "Canaries in the Coal Mine" (Brynjolfsson, Chandar, 2025) showing employment declines for young workers in AI-exposed occupations. That was referenced in Note 7 (the filed notes mention it in the synthesis). I can cite it as support for "AI removes entry-level tasks."
- BLS wage premium for bachelor's degree — mentioned in prior notes as an anchor.
- NCES enrollment ~19M — used in prior notes.
- IPEDS program-level enrollment data could test this.

But the instruction says "Do not invent evidence." The session substance itself doesn't name sources. The load-bearing elements:
- "AI removes the entry-level tasks those credentials promised" — I can cite the Stanford Digital Economy Lab "Canaries in the Coal Mine" finding (2025) that early-career employment in AI-exposed occupations declined — that's real and was already in the note corpus (Note 7's synthesis referenced it). Citing it is safe.
- "verified outcomes within ~2 years of entry" — a modeled threshold, not evidence. Could mark as a modeled threshold or Unverified.
- Program-level data (which majors are declining) — there's real data: NCES shows declines in English, history, humanities majors; business remains largest major. But I shouldn't overreach. I could note that the test runs on IPEDS/NCES program-level enrollment.

3. What would change this: falsifiers. If demand erosion shows up uniformly across program types (generic programs holding pricing as well as licensure-anchored ones), or if employers stop using skill verification and revert to degree proxies, the mechanism is wrong. Also: if program-level shifts track region rather than licensure status.

4. Format: # title in one line (dated claim), 1-3 short paragraphs, "What would change this: …" line, sources or Unverified tags.

The prior filed notes have titles like:
- "# AI bifurcates US university enrollment by region — 'grow or shrink' resolves as 'both'"
- "# The PhD Pipeline Concentrates Before It Rebuilds"
- "# By ~2045 the US research function splits into two geographies..."

So the title should be a one-line claim, possibly with a date marker like "(assessment 2025)" — earlier notes used "(assessment 2026)" and "filed 2025". The current date context: notes say "filed 2025" and "assessment 2026". I'll use something like "(assessment 2025)" or no date... Note 6 used "**Claim (filed 2025