Let me parse what's being asked. This is a long-running "Mind" exercise about whether US universities grow or shrink due to AI over 20 years, with regional dependence. Many notes have been filed across facets: enrollment bifurcation, demography, program mix, revenue shock, cost structure, research function, PhD pipeline, regional variation (three stacked currents), and labor-market signaling (bimodal BA returns by enforcement mechanism).

The current facet is: Labor market signaling — do degrees lose signaling value?

The substance from this session gives two claims:
- Signaling claim 7 (synthesis for the core): AI destroys the degree's screening monopoly; system sorts into three tiers — (a) research/elite flagships grow, (b) licensed-professional tier stable but capacity-capped, (c) collapsing teaching-only middle. Total degrees conferred likely still rise slightly while the number of institutions falls sharply. "Grow or shrink" is the wrong binary — the system thins, doesn't shrink uniformly. Sun Belt flagship-plus-community-college model outperforms Northeast small-private model.
- Signaling claim 8 (facet): Degrees lose signaling value asymmetrically. Collapse fastest in generalist white-collar fields (business, communications, generic liberal arts) where AI lets employers verify output directly at near-zero cost; slowest where statute enforces the credential (nursing, law, accounting, teaching) or failure costs catastrophic (engineering, medicine). Degree replaced as screening instrument, not as positional good. Portable AI-verified portfolios/skills assessments state "can do the job" more cheaply than a transcript; employers' pre-AI drift away from degree requirements accelerates once verification infrastructure matures. Two brakes: AI-generated application floods push some employers back to crude degree filters; elite degrees retain alumni-network/trust effects. Net by 2035: median degree signals "completed a four-year gauntlet" — luxury positional marker, no longer screening instrument. This split produces the three-tier sorting.

The existing filed notes on this facet: Note 16 (bimodal BA returns by mid-2030s — licensed-profession and top-30 premiums hold, mid-tier unlicensed compress 30-50%, enforcement mechanism). So the increment from this session is:
1. The asymmetry refinement: collapse is fastest in generalist white-collar fields where AI enables direct output verification; slowest where statute or failure-cost enforces the credential. This extends the "enforcement mechanism" framing with a specific field-level ordering.
2. The replacement framing: degree replaced as screening instrument, not as positional good. By 2035 the median degree is a luxury positional marker ("completed a four-year gauntlet"), not a screening device.
3. The two brakes: AI application floods pushing employers back to crude degree filters; elite alumni-network/trust effects.
4. The core synthesis: three-tier sorting (flagships grow, licensed tier stable/capped, teaching-only middle collapses); total degrees conferred rise slightly while institution count falls sharply; "thins, doesn't shrink."
5. Connection: the signaling split is what produces the three-tier sorting.

Prior notes already have: bimodal returns (Note 16), enrollment bifurcation (Note 4/14/15), "protected degree + AI skill stack" demand migration. The new durable claim should center on: degrees lose screening value asymmetrically by verifiability/enforcement of the underlying work, retaining positional value; and this mechanism produces the three-tier institutional sorting — the system thins rather than shrinks.

Load-bearing facts and sources:
- Employers' pre-AI drift away from degree requirements: real trend. Sources: Harvard Business School "Dismissed by Degrees" (2017, Fullerton, Rissing, Cappelli) documented degree inflation and employers dropping degree requirements; Burning Glass Institute / Impellio "The Emerging Degree Reset" (2022) — ~46% of middle-skill job postings dropped degree requirements in 2021-22; states like Maryland, Utah, Pennsylvania, Alaska dropping degree requirements for state jobs (2022-2023); major companies (IBM, Google, Delta, Walmart) dropping degree requirements for many roles. Also Opportunity@Work / Fuller et al. on STARs (skilled through alternative routes). These are real and citable.
- Skills-based hiring / verification infrastructure: real, growing — e.g., credentialing platforms, but the "AI-verified portfolios mature" part is forecast