I’m Going Back 2 School
It’s my Billy Madison year. Time for me to prove to dad that I’m no fool.
I have been on the bench working on recovery for 20 months and now I’m ready to get back in the game with some energy and purpose. I have my goals: Support Natalie in her xc ski race coming up in January, figure out how to live long time with Phi, and help a school that I love.
One year ago, this very day, I was on my second honeymoon with Natalie and celebrating my bday and I penned this idea to help UVU with the Charlie Kirk memorial.
Little came of my idea as it took the year to have the dust settle. Now it’s time to make something happen so that my daughter, a student of the UVU, can have another talking point besides the murder of a father and husband on her campus.
While building a cabin last monthwith one of my best friends and mentors, he told me his idea for UVU, and now I am fully behind him. Here is an option UVU can take for the future of the University, the State, the Country and the World… direct from the markdown file he sent to me:
# Four Million
## A strategy for making UVU software-defined, then AI-defined
*Working draft · September 2026*
---
## 1. Why universities adopt AI as marketing
Every university now has an AI statement, an AI task force, and at least one press release containing the word “transformative.” Almost none has changed anything an auditor could detect: not cost per credential, not who counts as a student, not what faculty do all day, not what a credential is. Universities take this hollow marketing approach to AI because that is the rational choice.
This is not a failure of vision or nerve. It is the predictable output of the incentives every university operates under. It is worth stating the incentives plainly, because any strategy that does not change them will be absorbed by them.
### Legacy incentives work against deep AI adoption
**Revenue is traditionally priced by time, not by learning.** Tuition is charged per credit hour; state funding formulas pay per full-time-equivalent student. A student who learns faster generates less revenue. The most obvious benefit of AI in education, faster and more individualized learning, is revenue-negative under the current model. No institution volunteers for a smaller top line.
**Cost is people, and people hold the veto.** Personnel is 60–70% of a typical university budget. Shared governance gives faculty formal authority over curriculum and practical authority over pedagogy. Any AI deployment that changes the cost structure must be approved by the people whose roles it changes. That is not an AI plan. It’s a stalemate with a press office.
**The rules encode the current process.** Federal aid eligibility is defined in credit hours and seat time. Legacy distance education qualifies for aid only with “regular and substantive interaction” initiated by an instructor who meets the accreditor’s qualifications. Accreditation standards measure faculty sufficiency. Conformity is safe, while innovation is regulatory risk with no regulatory reward.
**Prestige specifically rewards inputs.** Rankings measure selectivity, spending per student, and faculty ratio. An institution that serves more students with fewer inputs falls in the rankings. For most of the sector, doing more with less is a demotion. (UVU stands outside this game, an advantage we’ll return to.)
**Risk is asymmetric.** One student harmed by an AI tutor is a headline. Ten thousand students the status quo never reached are a statistic. Every general counsel and risk committee in the sector prices errors of commission far above errors of omission, and they are not wrong about how the world will treat them.
**The presidency is shorter than the transformation.** The average presidential tenure is about six years. The scorecard is enrollment, fundraising, and the absence of scandal. The costs of a real overhaul land inside one presidency; the returns land in the next. Rational presidents defer.
**Nobody owns the process.** A student’s path traverses admissions, financial aid, advising, the registrar, three or four departments, IT, housing, and a billing department. Each is a fiefdom with its own vendor and its own data. There is no end-to-end owner and no shared data model. You cannot automate what is not instrumented, and nothing in a university is instrumented end to end.
**Faculty met AI as a cheating tool first.** Promotion rewards research, not course redesign. On most campuses the first institutional conversation about AI was about academic integrity. The frontline’s default posture is defensive, and nothing in the current reward structure asks it to be anything else.
**The cheap version of AI satisfies everyone.** A chatbot on the website. An AI minor. A campus license. A Center for AI in Something. Each produces a press release at zero structural cost, and each satisfies the board, the legislature, and prospective students. Marketing spin preserves the current equilibrium.
### A test for telling the difference
An AI strategy is real if, three years after adoption, at least one of the following has changed materially: cost per credential, who counts as a student, what faculty do all day, or what a credential is. If none has, the strategy was marketing, whatever it was called.
### The implication
Incremental goals get absorbed by the current structure of a university. A 10% improvement target is met inside the existing process by effort. Even a 2x target can be met with more adjuncts, larger sections, and a better learning management system. Only a target the existing system cannot reach—under any amount of effort—forces the process itself to be redesigned. That is the argument for a moonshot, and for choosing a number that looks absurd rather than merely ambitious.
---
## 2. The moonshot
> **UVU will serve four million learners a year, at learning outcomes equal to or better than today’s, with its accreditation intact, as a nonprofit public institution.**
Four million is roughly 100x current enrollment (about 40,000) and larger than any university on earth (the largest open universities enroll around three million). The number is chosen because it is the lowest value that cannot be reached by any version of the current institution.
Each term in the moonshot statement matters:
- **Learner** means a person receiving instruction and assessment toward a credential UVU stands behind, not a registrant in a free course.
- **Outcomes equal to or better than today’s** is the quality floor. It makes the number un-gameable, and it requires measuring today’s outcomes first (which few universities do rigorously).
- **Accreditation intact** means NWCCU accreditation is a constraint, and won’t be sacrificed in order to reach scale. Credentials must transfer, stack, and open doors to graduate programs exactly as they do now.
- **Nonprofit public institution** rules out the obvious shortcut: converting to a for-profit and scaling fees with headcount.
**Why 100x, specifically.** Every incentive in Section 1 is a function of headcount. That is what makes a headcount target the right lever. At 100x:
- *Revenue per credit hour cannot survive.* Price must fall by roughly the same factor or the students never come. The revenue model is forced to change from seat time to marginal cost.
- *Personnel cannot scale.* UVU cannot hire 100x faculty or staff. Marginal cost per student is forced toward zero, which means every process that touches a learner must become software. This is the “software-defined” requirement, and it comes before any AI requirement.
- *Price falls below the level at which federal aid matters.* Title IV, and with it the credit-hour definition and the interaction rules, detach by design rather than by exemption. Accreditation remains intact.
- *Prestige incentives do not apply.* UVU is open enrollment and outside the rankings game. This moonshot goes where the selectivity incentive doesn’t reach.
- *The risk asymmetry inverts.* Measured outcomes at scale turn the unserved from a statistic into a number on the same dashboard as the headline metrics.
- *The presidency horizon is addressed* by intermediate deliverables that are valuable in their own right (and within twelve months), whether or not the number is ever reached.
- *The ownership problem is addressed* by a chartered unit that owns the student’s path end to end, application through alumni, because at 100x there is no other way to build it.
- *Faculty authority moves upstream,* from delivering individual sections to owning curriculum, assessment design, and quality across populations of learners, which is where the leverage is.
---
## 3. The logic, step by step
### Software-defined before AI-defined
AI can act only on what is legible: instrumented, stateful, and addressable through an interface. Bolting AI onto email-and-PDF workflows produces demos. The first phase of transformation is, therefore, unglamorous. Every process that touches a learner (exploration, application, placement, enrollment, instruction, assessment, advising, credentialing, billing, alumni relations) gets an API, a data model, and a set of metrics. This is also where the compounding begins—each instrumented process makes the next one cheaper to instrument, and every AI capability built later inherits and leverages the whole stack.
### Why now
[Bloom’s two-sigma problem](https://en.wikipedia.org/wiki/Bloom%27s_2_sigma_problem) has been known since 1984: students tutored one-on-one perform about two standard deviations better than students taught conventionally, and the constraint was always cost. The cost of individualized instruction and feedback has fallen by orders of magnitude in three years and is still falling. This is the only new aspect of tutoring today. Everything else has been true for thirty years and was simply unaffordable.
### Price is the regulatory strategy
Below a few hundred dollars a year, federal financial aid becomes irrelevant to the learner (the FAFSA is no longer worth filling out). Without Title IV, the interaction rules, credit-hour definitions, and satisfactory-progress requirements vanish. What remains is the accreditor, whose question is whether qualified humans own curriculum, assessment, and the quality system, not whether one owns each learner’s loop. That is the difference between a faculty ratio of 1:30 and one of 1:3,000.
Two consequences. First, the price target should be set at the ignore-the-FAFSA threshold, not at the cost floor. The gap between the two prices is the financial margin that funds human time, and human time is what makes “full value” true rather than cosplay. Second, the existing campus population keeps federal aid and everything that comes with it. The moonshot doesn’t touch it.
### Funds and governance are separate
Declining state appropriations does not remove the trustees, the Board of Higher Education, the legislature, or the state’s ownership of the name and the campus. Self-funding buys real autonomy, since the budget is the legislature’s main lever, but not independence. Accreditation, meanwhile, is institutional—it belongs to UVU and doesn’t get externalized or exported.
The structure that resolves this already exists within the state. BYU-Pathway grants no degrees and accepts no federal aid. It is a service organization that delivers courses on behalf of BYU-Idaho and Ensign College, which award the credentials and hold the accreditation. Learners who need federal aid switch providers (e.g., from BYU-P to BYU-I) and stay in the same degree.
The UVU version: UVU awards the credential, holds the accreditation, and remains a state-governed, Title IV institution for its campus population. A separate nonprofit builds the software, runs delivery, holds the intellectual property and the revenue, and never touches federal money. The state’s relationship with the nonprofit is contractual rather than statutory. The accreditor reviews a substantive-change filing rather than a new institution. The nonprofit needs no accreditation of its own—UVU’s is the one that counts.
### How a student earns the same degree
The degree is defined by what a learner can demonstrate, not by how they were taught. That is the guarantee, and it is the same one UVU already makes when its transcript does not distinguish online sections from face-to-face ones. _The nonprofit is a delivery channel into UVU’s programs, not a parallel or competing institution._
#### A learner’s path
1. **Enrollment.** The learner enrolls with the nonprofit: open enrollment, identity verified, at the nonprofit’s price. From the first day _the learner is pursuing a UVU program of study_, with the same catalog outcomes and degree requirements as a campus student. The nonprofit cannot alter those requirements.
2. **Placement.** Placement is by assessment, not transcript. A learner who can already demonstrate proficiency gets credit and moves on to their next course.
3. **Instruction.** Delivered on the nonprofit’s platform: AI tutoring and feedback, adaptive sequencing, projects, and a cohort or gathering group. The content is faculty-designed. Faculty don’t teach sections for the nonprofit. They design courses, define outcomes, set the assessment bar, and monitor results across the whole population.
4. **Assessment.** Every outcome is assessed against instruments UVU faculty own, _with the same standards as campus._ AI handles routine feedback; high-stakes judgments are made by faculty or faculty-certified evaluators, under identity verification and secure assessment conditions. Where the accreditor permits direct assessment, credit is awarded on demonstrated competency rather than seat time.
5. **Record.** Verified results flow through the API into UVU’s student record. The learner earns UVU credit on a UVU transcript, and _the transcript does not report the delivery channel._
6. **Credential.** _UVU confers the degree._ Same diploma, same accreditation, same transfer and graduate-school standing as a campus graduate. There is no second-tier credential, and the quality floor is what makes that claim defensible: outcomes are measured for both channels and published side by side.
7. **Mobility.** Because the credit is UVU credit, a student can move between channels at any time with nothing lost: to campus for federal aid, a residential experience, or a lab-based program; from campus to the nonprofit for price or flexibility. This is how BYU-Pathway learners switch providers today.
#### The division of authority
- **UVU owns the program:** outcomes, curriculum, assessment instruments and bar, faculty qualifications, the student record, and the credential. Faculty governance applies to all of it.
- **The nonprofit owns the channel:** the platform, AI instruction and feedback, community, advising triage, support, pricing, and the alumni and mentoring network described above.
- **A written agreement defines the boundary,** and NWCCU reviews it as a substantive change. Federal rules limit how much of an aid-eligible program an unaccredited partner may deliver; the nonprofit channel sits outside those rules because it carries no aid, and _the campus channel is untouched._
#### Two records, one boundary
“The student record” becomes two different things, and the separation of responsibilities is what makes the division of authority possible.
The *learning record* is everything the platform generates: attempts, feedback, tutoring transcripts, mastery signals, engagement data. It is large, fast-changing, and operational. The nonprofit designs it, holds it, and is free to change it.
The *academic record* is the small, slow ledger of facts UVU vouches for: this person was admitted to this program; this outcome was assessed against this faculty-owned instrument and passed on this date; this credit was awarded; this degree was conferred. It is what a transcript is printed from and what an accreditor audits.
The boundary between the two records is the attestation. The platform may report that a learner has demonstrated an outcome; only UVU can attest it. The nonprofit emits the verified result, UVU’s ledger records the credit, and the registrar remains the owner of the schema and the sole issuer of transcripts. The nonprofit builds the ledger to UVU’s specification as UVU’s property, UVU can export and run it without the nonprofit at any time, and the nonprofit holds learner data as a school official under UVU’s direct control. Nothing about who runs the servers changes who owns the degree.
### Full value, decomposed
“The full value of a university education” is a bromide until it is broken into parts and each part is assigned a delivery mechanism.
- **Instruction and feedback.** Software and AI, with faculty owning design and quality. This is the part that scales without limit.
- **Assessment and credential.** Competency-based and direct-assessment where the accreditor permits; faculty-owned; the credential is UVU’s and transfers as UVU credit.
- **Peer community.** Cohorts, studios, peer review, and gathering groups. This scales because learners are the supply. Georgia Tech’s online master’s program and BYU-Pathway’s gathering model are both proof.
- **Mentorship and advising.** The scarce resource. Human time is rationed by AI triage and extended through alumni and employers. Alumni relations therefore lives in the same stack as everything else, not in a separate advancement database: the alumni record is the student record continued, and mentor pairings, employer introductions, and re-entry as a learner are matched and managed from it. At four million, the alumni base is the mentoring supply, and it scales the same way community does.
- **Place.** The campus remains, and becomes the lab and studio for those who come to it.
The honest claim is not that every learner gets the residential experience. It is that every learner gets the components that matter for their goal, at a quality UVU measures and publishes.
### Why UVU
- Open enrollment, so there is no selectivity to protect.
- Outside the rankings, so there is no prestige to lose.
- Dual-mission, so certificates, associate degrees, and bachelor’s degrees already form one pathway.
- Already working the seams, through Fast Track 1-2-3 and the K-16 concurrent-enrollment alliance.
- Accredited by NWCCU, which already [accredits WGU’s competency-based model](https://cm.wgu.edu/t5/Information-Resources/Accreditation/ta-p/178) at six-figure scale.
- Sitting between Silicon Slopes and the organization that built BYU-Pathway.
- Four million learners is Vision 2030 without a geographic boundary.
### Bounded downside
Fall short by 10x and UVU serves 400,000 learners as a software-defined institution where “AI-forward” is a description. Fall short by 100x and UVU still has an instrumented process map, a measured outcomes baseline, and one direct-assessment program, which is more than any AI task force in the country will deliver.
### The first twelve months
This document plans twelve months and no further. Anything targeting a later date will change as the first year teaches what nobody can know today, and committing to it now would be false precision.
The ask is not that UVU adopt the number (and certainly not that UVU advertise the number publicly). It is that UVU charter a unit with its own rules and four deliverables:
1. **An outcomes baseline.** What today’s 40,000 learn, measured well enough to be compared against.
2. **A process map** in which every workflow, from application through alumni relations, has an API and a set of metrics.
3. **A leverage model** that states, as a number, how many learners one qualified faculty member can own when software does everything else.
4. **One direct-assessment program** filed with NWCCU and the Department of Education.
Those four deliverables become the forcing function. The moonshot number is what makes them non-optional.
---
## References
**BYU-Pathway Worldwide**
- [Main site](https://www.byupathway.edu/)
- [Tuition and financial aid policy](https://www.byupathway.edu/2-4-tuition-financial-aid-edu)
- [Independent program review](https://www.collegetransitions.com/blog/online-college-review-byu-pathway-worldwide/)
- [Background (Wikipedia)](https://en.wikipedia.org/wiki/BYU%E2%80%93Pathway_Worldwide)
**WGU accreditation**
- [WGU’s accreditation statement](https://cm.wgu.edu/t5/Information-Resources/Accreditation/ta-p/178)
- [WGU’s story](https://www.wgu.edu/about2/story.html)
- [2024 NWCCU reaffirmation press release](https://www.wgu.edu/newsroom/press-release/2024/03/fortifying-excellence-wgu-nwccu-accreditation-reaffirmed.html)
- [WGU archive of IRAC accreditation letters](https://archive.wgu.edu/wgu-archive-structure/accreditation)
**WGU’s Title IV audit**
- [2017 OIG audit coverage (Inside Higher Ed)](https://www.insidehighered.com/news/2017/09/22/education-depts-inspector-general-calls-western-governors-repay-713-million-federal)
- [NPR on the faculty-role question](https://www.npr.org/sections/ed/2017/09/28/553753020/who-is-a-college-teacher-anyway-audit-of-online-university-raises-questions)
- [2019 Department decision not to enforce (Higher Ed Dive)](https://www.highereddive.com/news/western-governors-u-does-not-have-to-pay-back-713m-in-title-iv-funds/545908/)
**Regular and substantive interaction (RSI)**
- [WCET’s interpretation of the 2021 rule](https://wcet.wiche.edu/frontiers/2021/08/26/rsi-refresh-sharing-our-best-interpretation-guidance-requirements/)
- [NC-SARA summary of the final rule](https://nc-sara.org/news-events/us-department-education-issues-final-rules-distance-education-and-innovation/)
**NWCCU**
- [Accreditation process](https://nwccu.org/accreditation-process/)
- [Substantive change policy](https://nwccu.org/substantive-change/)
- [Substantive Change Manual (2023)](https://facsen.wsu.edu/documents/2024/09/2023-nwccu-substantive-change-manual.pdf/)
**Precedents**
- [ASU charter](https://www.asu.edu/about/charter-mission)
- [Georgia Tech OMSCS 2025 annual report](https://omscs.gatech.edu/2025-omscs-annual-report)
- [OMSCS origins](https://pe.gatech.edu/blog/georgia-tech-celebrates-40-years-learning-distance)
**UVU**
- [Vision 2030](https://www.uvu.edu/vision2030/achieve.html)
- [Fast Track 1-2-3 announcement (Deseret News)](https://www.deseret.com/education/2026/01/27/uvu-announces-initiative-to-help-students-enter-work-force-sooner/)
- [UVU Online strategic plan](https://www.uvu.edu/academicaffairs/docs/academic-programs/uvu-online-strategic-plan.pdf)
Brent’s idea
Published by Daddy x happy 420