The Sanctioned Hack of Higher Education and AIM

A Washington Post reporter clicked his way to eleven college credits in a few afternoons; add in that the federal rulebook was being rewritten to make that the approved path, and well...

(If you find yourself in a hurry—this one is pretty long, but there’s a link to a video explainer that’s about 8 minutes that sums it up pretty well down at the bottom of the piece.)

A reporter named Todd Wallack just earned eleven college credits for $83.74.

He did it on a platform called Sophia Learning, mostly by opening the electronic textbook in one browser window, opening the open-book test in another, and using CTRL-F to find the answers. Two classes in a single day. Environmental science, three credits, five and a half hours, a subject he’d never studied. He skipped the lectures. He skipped the readings. When a test asked about the Kyoto Protocol, he searched the course materials for “Kyoto,” found the passage, and moved on. The one writing assignment he submitted ran a third of the required length, didn’t fully answer the questions, and passed anyway on partial credit. Everything is pass-fail. You need a 70. You can retake what you miss.

Those eleven credits transfer to more than fifty accredited colleges. At more than a hundred institutions nationally, the courses will actually count. Of course, Wallack didn’t bother transferring them because he already has a degree and a job writing for the Washington Post. But, a working adult with no degree and a scholarship deadline could stack these toward most of a bachelor’s in a matter of a few weeks, for (a lot) less than the cost of a single three-credit course at the university where I teach.

Wallack’s story is really worth reading. Sophia told him that people “with extensive professional work experience and degrees,” like him, move fastest, and some of his speed was just that. But the companion piece he reported the same week is what turns the afternoon-long course from one journalist’s stunt into a system fact: hundreds of thousands of paying users, the fastest-growing segment of the entire transfer economy. He’s not an outlier. He’s a clean demonstration of something already happening at scale.

He closed the piece with a doozy: “I did learn just how quickly you can earn college credit using CTRL-F.”

It’s not that pay-for-play private education is new. It’s not that easy courses are new. It’s the speed and hackability and the potential AI-ness that is. And, I’ve written about this before, sort of. Back in the spring, when Wallack ran his first degree-hacking story (the one about Christie Williams finishing a bachelor’s in three months and a master’s in five weeks for four thousand dollars total), the temptation was to read it as a story about clever students gaming a sleepy system.

That’s the wrong read, and it was the wrong read then. The students aren’t really the story. The students are doing the sector’s diagnostic work for it, out loud, with a receipt. What they’ve figured out is that for a large slice of American higher education, the credential and the learning it used to certify have already come apart. The hack works because the product was already hollow. They just noticed first.

What’s new in June isn’t the hack. It’s two things that arrived in the same few weeks and have not yet been put together. So, here you go.

The first is that the hack is no longer something a reporter has to go find. He can just do it, in an afternoon, and write down what it felt like. The second is that while he was doing it, the federal government was finishing a rewrite of the rulebook that turns the whole arrangement from an exploit into the mostly approved architecture. Not closing the loophole. Paving it with nice asphalt.

Let’s take those in order, because the second one is the part that should make higher ed admins (I’m looking at you provosts, but a lot of others too) need some good sleep meds, and it doesn’t make sense without the first.

The platforms aren’t colleges, which turns out not to matter

Sophia is owned by Strategic Education, a publicly traded company in Herndon, Virginia, that also owns two for-profit colleges, Strayer and Capella. Both Strayer and Capella accept Sophia credits. That’s an interesting business model: one company runs the place that manufactures the credits and two of the places that honor them. The credit and the degree, vertically integrated under a single corporate parent.

Sophia’s own numbers, via Strategic Education’s filings, show why this is a growth business and not a curiosity. Average users in the first quarter of 2026 were up 40 percent year over year. Over four years, 215,000 paying subscribers, 65,000 of them new last year. Two of the three big platforms, Sophia and Study.com, let you take as many classes as you can stomach for under a hundred dollars a month. Study.com used to proctor its exams. It dropped the proctors early last year to match Sophia’s open-book, nobody-watching model. The third platform, StraighterLine, kept its remote proctors, and its CEO explained why in a sentence that tells you these companies know exactly what they are: “I don’t want to be known as the cheat path for hacking college.”

So one platform competes on rigor and the other two on frictionlessness, and the frictionless ones are the ones growing 40 percent a year. The incentives here aren’t subtle. When the product is a transferable credit and the buyer is a time-strapped adult, the platform that asks the least of you wins. That’s not a moral failing on anyone’s part. It’s a market doing what markets do once you’ve defined what’s being sold as the credit rather than the learning and when you have multiple sources of disruption flying around your industry all at once.

And the buyers are clear-eyed about what they’re getting. “You won’t learn anything,” one wrote on a Sophia forum, “but you will have it done.” That isn’t a confession. It’s a product review. Check the box.

But the platforms aren’t colleges. They don’t grant degrees, they don’t touch federal student aid, and so they sit outside the reach of both the Education Department and the accreditors. A former vice president of the Council for Higher Education Accreditation put it plainly: “They are unregulated.” So how does a course you finish in an afternoon become credit at an accredited university?

The bridge is an outfit called the American Council on Education, which represents something like 1,600 colleges and universities. None of what follows is secret or new. ACE has recommended courses for college credit since the 1970s, and evaluated military training for it since the 1940s. The machinery is transparent, routine, and decades old, which is exactly what makes it worth looking at: this is how the sector learned to accept credit it didn’t generate without asking what the credit represents. ACE reviews courses on these platforms and issues recommendations, and the recommendations carry so much weight that students, platforms, and registrars all just call them “ACE credits,” as though ACE were doing the teaching. More than a hundred colleges lean on those recommendations to decide what to accept. ACE charges the platforms fees for the reviews, which brings in a couple hundred thousand dollars a year. The group says the fees cover costs and don’t sway its judgments, and I have no reason to doubt the people involved are acting in good faith. But ask ACE’s president whether the council studies how much students actually learn in these courses, or how they compare to a traditional class, and the answer is: “That’s not our role.”

ACE’s position deserves its due. I mean, at least it’s internally coherent: they’re not an accreditor, they’re a recommending body, they review courses the way faculty review a colleague’s syllabus and they leave the accept-or-reject call to the colleges. Fair enough. But follow the chain all the way down. A platform builds a course. ACE reviews it for a fee and recommends it, without measuring whether anyone learns anything, because measuring that isn’t its job. A college accepts the recommendation, because vetting every outside course itself is expensive and ACE’s imprimatur is the cheap substitute. The student gets the credit. The credit goes on a transcript that looks identical to one earned the slow way. And at no point in that chain does anyone whose job it is to certify learning actually certify that learning happened. Everyone has discharged their narrow responsibility correctly, and the system as a whole has quietly stopped doing the one thing it exists to do.

Then there’s the newer wrinkle, from March. Sophia got an endorsement from an actual accreditor, the Higher Learning Commission. Sophia paid a $10,000 application fee and makes smaller annual payments to keep the endorsement. The commission’s spokeswoman was admirably candid about what the endorsement does and doesn’t mean: HLC “did not review any of Sophia’s courses,” and the endorsement “applies to the provider, not to individual offerings.” So an accreditor has put its name on a platform, for a fee, without looking at a single course, and the platform can now wave that endorsement at any college still on the fence.

The fee isn’t really the scandal either. Accreditors charge fees; institutions pay their accreditors; that’s how the system has always run, and a $10,000 application fee is rounding error. The fee bought an accreditor’s name on a provider it never reviewed, at the course level, which is where the actual teaching either happens or doesn’t. That’s a new product for an accreditor to sell, provider-level endorsement decoupled from any look at what gets taught.

Hold that thought, because there’s about to be a federal rule that rewards exactly this kind of lean, fast recognition.

What’s being sold

The obvious objection to everything above is: “so what?”

Adults with life experience testing out of intro courses they’d ace anyway is arguably a good thing. Forty-three million Americans started college and never finished. If a single mother working full time can convert what she already knows into a credential for a few hundred dollars instead of forty thousand, that’s mobility, not fraud.

The objection holds mostly. For a knowledgeable adult, testing out of an intro course in an afternoon isn’t fraud; it’s competency-based education doing what CLEP and AP have done respectably for fifty years. Seat-time was always a lousy proxy for learning. One of the speedrunners in Wallack’s reporting, a 47-year-old marketing manager, put it cleanly: “Sophia essentially just validated what I already knew and turned it into college credit.” If that’s what’s happening, the hack isn’t hollowing the degree; it’s pricing a product that was overpriced all along.

So speed isn’t the real problem here; the checkpoint is. It can’t tell the difference between the adult who already knew the material and the kid (or reporter) who found the answer with CTRL-F. On an open-book test with no proctor and a 70 percent floor you can retake, the thing certifying the course has stopped being able to distinguish mastery from search.

Which throws the weight onto a question everyone keeps stepping around: what is the degree actually selling?

As my readers likely know, I’ve built a map of American higher education: 1,556 four-year institutions plotted on two axes, institutional resilience and post-college market position, sorted into four quadrants. I won’t re-litigate all of it here. The relevant finding is this: 491 institutions, just under a third of the total, sit in what I’ve called High Stress, below the median on both axes at once. Shrinking and structurally fragile. Disproportionately the small, tuition-dependent, less-selective schools and a big chunk of the master’s-comprehensive middle. And 82 percent of R1 research universities sit in the opposite corner, High Capacity, strong on both.

That distribution is important, but it’s not a verdict. The map measures resilience and market position. It doesn’t measure what a school sells, and hackability is a property of what you sell, not of how solvent you are. A flush R1 can run a thoroughly hackable online master’s; a broke little college can sell intense formation it can barely afford. So the quadrants aren’t pathway labels. What they give you is a tendency: the institutions under the most financial pressure are the likeliest to have already thinned the product toward a transferable transcript, because that’s the cheapest product to sell.

The pathways I've proposed are the complementary tool, I think. I laid out four of them a few months back, the directions institutions are sorting into as the old all-things-to-everyone bundle comes apart, and they sort by a single question: how much of what you’re selling survives a CTRL-F?

Start at the hackable end. A Credential Factory sells a certified transcript, efficiently. That’s the value proposition, and there’s nothing wrong with stating it plainly: fast, cheap, outcome-anchored, get the working adult the piece of paper the labor market still asks for. Sophia is the purest form of it, but plenty of accredited online completers run the same play: the schools that let you transfer in up to three-quarters of your coursework from the platforms. For this pathway, the hack just is the product, working as designed. You cannot “cheat” your way through a Credential Factory any more than you can cheat your way through a vending machine. You put in the money, you get the credit. That’s the deal, openly.

At the other end sits Elite Mentorship, and it’s essentially unhackable. Not because it’s better policed. Because there’s nothing separable to extract. What a degree from this pathway certifies is that you were in the room: the seminar where the argument actually happened, the lab where the technique lives in someone’s hands, the network, the years of formation. There is no browser tab for “I was mentored by the person who wrote the field.” This is most of the R1s, the top doctoral cohorts, the residential liberal arts colleges with the endowments to keep class sizes small. CTRL-F finds you nothing here, because what’s being sold was never written down in a place you could search.

None of which makes Elite Mentorship honest and everything below it crooked. The formation layer gets gamed constantly, just by different people and different means: legacy preferences, the strategic donation, the ghostwritten thesis, the Columbia sophomore who ran ChatGPT on eighty percent of his essays and turned the suspension into a venture round. What changes down the gradient is the cost of entry. Gaming Elite Mentorship takes money, access, or outright fraud. Gaming a Credential Factory takes a discount code. Hackability is a continuum, not a wall, and the thing that moves along it is how much you have to spend to beat the checkpoint.

Experience University is the odd case, and I should be honest that it doesn’t really sit on the hackability axis at all. It’s immune for a reason that has nothing to do with whether its product resists a CTRL-F. What it sells is four years of residential life, the football Saturdays, the Greek system, the assortative sorting, the alumni network you join for good. The credential rides along, but it was never the point. I made this argument about Harvard a few weeks ago: that Harvard and the big state flagships land in the same high-resilience corner by completely different machinery, and that neither one fails undergraduates in any real number because failing the customer is hostile to the product, whatever the product happens to be. The hack doesn’t threaten Experience University because you can’t CTRL-F your way into the network. You had to be there, in the dorm, at the game, for four years. The transcript is a souvenir.

And then there’s the Hybrid Knowledge Platform, which is where the hard fight to come lives. These are the public research universities, the land-grants, the regional comprehensives, the institutions trying to sell research-driven teaching and applied learning at scale, to a lot of students, many of them commuting, most of them price-sensitive. This pathway can’t ban-and-supervise its way to safety like Elite Mentorship; it’s too big and too distributed. It can’t accept the hollowing like Credential Factory without dissolving the exact quality that makes it worth more than Credential Factory. It has to do the hard work: defend the cognitive labor, redesign assessment, build the kind of teaching a chatbot can’t shortcut. I’ve called this defending the first mile, and it’s genuinely hard, and most of the institutions that need to do it are sitting in that contested middle of the map with the thinnest resources to pull it off.

So three of the four sort cleanly by how much survives a CTRL-F. Credential Factory hacks because the product is the transcript; Elite Mentorship is the hardest to reach because the product is the formation; Hybrid sits between them, the one pathway still undecided. Experience University is the exception that won't stand on the line at all, immune not because its product resists the hack but because its product was never the credential in the first place.

More hackable on the left, harder to reach on the right.

You can watch the line getting drawn inside single institutions right now, in the June reporting, if you know to look. The University of Maryland’s flagship campus at College Park won’t accept platform credits. The University of Maryland Global Campus, its own online sibling, will. Same brand, two answers. Arizona’s main campus says no to Sophia; Arizona Global says yes. The credential is splitting from the degree, and you can see the seam running right down the middle of a shared name. A school’s flagship arm protects the old product while its online arm sells the new one, and they both put the same word on the diploma.

The rule that paves the road

Which brings me to the part I actually sat down to write about, and the reason the timing of Wallack’s June pieces matters so much.

In May, the Accreditation, Innovation, and Modernization committee (AIM, the Education Department’s negotiated rulemaking on accreditation) reached consensus. I wrote about it at the time as one of four converging pressures reshaping the credential, and I promised to track what happened as the rule moved from negotiated text toward an actual regulation. This is me keeping that promise, I guess. Take the full consensus redline, and read it against Wallack’s degree-hacking package, and well, it’s a strange experience. It’s like reading the user manual for what the reporter just did.

Three provisions, in particular.

The first is a transfer-of-credit rule. Under the consensus text, accreditors have to ensure that institutions award transfer credit for comparable undergraduate coursework completed at any school accredited by a recognized agency, unless the institution writes down a specific reason to refuse. The presumption flips from “we’ll evaluate your credits” to “we accept them unless we can justify saying no.” Read it next to the Maryland and Arizona examples and you can see what it does: it leans on the College Park end of every split-brand institution to behave more like the Global end. And notice the boundary. The award requirement is explicitly undergraduate. It doesn’t reach graduate coursework. The negotiators protected the master’s market’s closed, high-margin economics while prying the undergraduate market further open. You can read that boundary two ways. The charitable read is that graduate credits are more specialized and genuinely harder to equate across institutions, so the presumption shouldn’t apply. The cynical read is that the undergraduate credit is where the volume and the political pressure sit, so that’s the market that got opened. Whichever read you take, the effect is the same: the commodity-transfer machine is being bolted onto undergraduate education specifically, which is exactly where Sophia sells.

Hollis Robbins has been circling this a lot longer than I have, and she gave the mechanism its name. In an essay back in April she called it the Transfer Industrial Complex: the mostly invisible apparatus of learning-outcome rubrics, common course numbering, and articulation grids that exists so a credit earned at one school can move to another. Her rough estimate for what the sector spends every year to make credits portable runs between two and three billion dollars, almost all of it buried in operating budgets where nobody can add it up. And the point of the spending is sameness. Comparable data needs comparable units. A sociology course here has to mean the same thing as a sociology course there, or the transfer math breaks. So the system files every course down until the courses are interchangeable.

Robbins has a metaphor for the finished product that I really can’t improve on: each course becomes a kind of QR code. Scan it and the receiving institution has everything it needs to grant the credit. What the code doesn’t carry, she notes, is “anything about quality (or ability) of instructor or whether the student learned anything.” A course with the right code can be taught by a senior scholar or a first-year grad student or an adjunct running six sections, in a seminar or a thousand-seat hall or an asynchronous shell, and it scans the same regardless. This will be increasingly painful to the egos of many in the higher ed landscape, and will be a source of identity threat for sure.

Then line that up against Wallack’s afternoon on Sophia. Robbins did, as it happens: the day his piece ran, she posted that the transfer industrial complex “needs to end.” The reason eleven CTRL-F credits move to a hundred colleges is that the sector spent two decades and a couple billion dollars a year building a machine whose job is to strip out the information that would let a receiving school tell a real course from a hacked one. The hack isn’t beating the transfer system; it’s riding the rails the system laid down. And, let’s be very clear, the AIM rule invents none of this. It’s the federal capstone on a structure that was already most of the way built.

Robbins is careful here, and I should be too. The portability machine is not pure waste. Before these mandates, fewer students could move at all, and the ones who did ate the full friction themselves, re-proving every course from scratch under rules that put the burden on them. Making credits portable was a real gift to the forty-three million people who start college and have to stop. The trouble isn’t portability itself; it’s that the same machine that lets a genuine community-college calculus course count at a university also lets an afternoon of CTRL-F count, because it was built to read the QR code rather than the course. It can’t tell a good transfer from a hollow one. Telling them apart is precisely the work it was designed to skip.

The second provision is the one that deserves some more thinking. In applying standards about program length, an accreditor “must not categorically prohibit or unreasonably restrict” the accreditation of programs that run shorter than the traditional time to a credential, as long as the program produces “comparable academic, professional, and employment outcomes.”

That last clause matters, and it cuts against the easy version of my point. The rule doesn’t strip the accreditor of every tool. An accreditor can still move against a fast program on the basis of bad outcomes, weak placement, poor earnings, low completion. What it can no longer do is object to the speed itself. The traditional floor was a proxy: a “real” course takes a semester, fifteen weeks of seat-time, because that’s how long learning was assumed to take. That proxy is exactly what the provision retires. An accreditor that looked at the eleven-credits-in-an-afternoon model and said “no, that’s too fast to be real” would now be out of compliance, because “too fast” is no longer a permissible objection. It would have to show the outcomes are worse. And given that the outcome usually being measured is whether the credit transferred and the graduate got a job, not whether anyone learned anything, that’s a much harder case to build than “this isn’t a real course.” The rule doesn’t gag the referees. It takes away the one call they used to be able to make without a replay review, and leaves them the calls that are hardest to prove.

The third is structural. AIM lowers the barriers for new accreditors to win federal recognition and pushes the system toward outcomes-based standards (completion, placement, earnings) and away from inputs like faculty credentials or library holdings. On its own merits there’s a real argument for this; inputs are a crude proxy and outcomes are what students care about. But put it next to the Higher Learning Commission quietly endorsing Sophia for a $10,000 fee without reviewing a course, and you can see where the gradient points. A recognition regime that rewards measurable throughput and welcomes new, leaner accreditors is a regime in which the HLC-endorses-a-platform move stops being an experiment and becomes the template. The credit-recognition machinery Wallack documented (ACE recommending for fees, an accreditor endorsing a provider without reviewing its courses) is precisely the architecture AIM is about to bless and scale.

The rule isn’t a cartoon. Outcomes-based standards cut both ways: they squeeze the Credential Factory programs with weak placement data just as surely as they reward the ones with strong numbers. The transfer presumption really does help the forty-three million Americans with stranded credits. The conflict-of-interest provisions buried elsewhere in the text go after some real accreditor capture. None of this is the work of people who hate education.

And I think there is a sharper version of the case than the rule’s own defenders usually bother to make for some reason. For a thirty-five-year-old warehouse supervisor with sixty stranded credits and a promotion that requires a degree she already half-earned a decade ago, the fifteen-week semester and the four-year residential ideal aren’t sacred. They’re a tax she can’t pay, defended most loudly by people whose own degrees were bought and paid for long ago. Gatekeeping has victims too, and they are rarely the ones writing essays about the sanctity of the credit hour. A professor or an admin mounting a defense of slow, expensive, in-person education should say plainly whose interests that defense also happens to serve, mine among them.

But read alongside what’s happening on the platforms, the rule has a clear net direction: it makes the structural economics easier for exactly the credentialing functions that CTRL-F can already complete, and it tells the accreditors they may not stand in the way.

So here’s the convergence…from below, the hack: a reporter clicking through eleven credits and the credit-recognition machinery underwriting it. From above, the rule: a federal rewrite that turns the transfer pipeline into a presumption, protects the speedrun from accreditor interference, and welcomes the pay-for-endorsement model into the mainstream. It’s one story: the same gradient, hardened from both ends at once.

What’s also interesting is that I don’t think somebody sat in a room and planned this mishmashmeshing. The hack and the rule were built by different people solving different problems. A platform chasing subscriber growth. A negotiating committee chasing a deregulatory mandate and a stranded-credit constituency. A sector chasing the cheapest product it can still call a degree. They converged anyway, because the incentives all point the same direction, and that’s what makes it hard to stop. There’s no villain to name and no switch to flip. Just a lot of rational actors building, brick by brick, a thing none of them would defend if you made them describe it whole.

What Wallack did with a keyboard shortcut, the rule is on track to make the default by 2027, if it survives the comment period and the lawsuits and all that. (That’s not a small if.) But the litigation is going to land on the culture-war provision riding alongside these, the intellectual-diversity language, not on the transfer and acceleration rules.

Those are the part nobody’s going to march about. The pieces that turn the hack into the default are the pieces almost no one is fighting. Not an exploit of the system. The blueprint for it.

Where this leaves the institution in the middle

If you run a Credential Factory, none of this is a threat. It’s an awesome tailwind. The rule runs in your favor, the platforms are your supply chain, and the only risk is that your outcomes data isn’t good enough to survive the new scrutiny, which is a fixable problem and an honest one.

If you run an Elite Mentorship institution or an Experience University, you’re insulated, for now, by a product the hack can’t touch. The seminar and the network and the football Saturday don’t transfer in from Sophia. Your exposure is at the margin, where the BA still works as a generic employer signal, and at the level of public trust, where “even Harvard’s grades don’t mean anything anymore” corrodes the whole category. But your core product survives CTRL-F, because your core product was never a searchable document.

It’s the Hybrid pathway that’s really caught, and it’s caught precisely because it’s where most of American higher education lives. The contested middle of my map (the regional comprehensives, the master’s institutions, the public R1s without the endowment cushion) is full of schools selling a product that’s better than a Credential Factory’s but that they haven’t figured out how to defend as better. They’re charging Hybrid prices for programs that are, in practice, hackable, and the rule is about to expose the gap. A student can now assemble three-quarters of a degree for a few hundred dollars and transfer it in under a presumption the accreditor is required to honor. If the last quarter the institution adds doesn’t visibly do something the first three-quarters didn’t, if it’s just more of the same searchable coursework at fifty times the price, then the institution is selling a markup it can no longer justify on the merits.

Whether students actually stop paying is a separate question, and the answer isn’t obvious yet. Credentials are quite sticky. The BA still works as a labor-market filter regardless of how it was earned, and a degree can be hackable and still worth buying as long as employers keep rewarding it. The repricing could take a decade, or stall out entirely behind the inertia of a signal that still works. But “the market hasn’t punished us yet” is a thin foundation to build a strategy on, especially when the price gap between your product and the hacked version is widening and increasingly visible to the eighteen-year-olds doing the math. The pressure is real even if the reckoning is slow.

The way out likely isn’t to fight the transfer rule or pretend the platforms don’t exist. That fight is pretty lost and probably shouldn’t be won. The way out is to be honestly, demonstrably what a Credential Factory isn’t: to put the applied research, the clinical hours, the lab residencies, the mentorship, the assessment a chatbot can’t complete, at the center of the product and make it legible to students and employers. To defend the first mile. Most institutions in the contested middle haven’t done this, because it’s expensive and politically hard and the budget office would rather not, and because for thirty years they got away with not having to. The rule ends the getting-away-with-it.

I’ve let CTRL-F stand in for AI this whole essay, and that was deliberate. CTRL-F is the hand-tool version of the problem, the one a reporter can demonstrate without quite looking like he cheated. The students coming up behind him don’t need it. They have a model that reads the textbook, sits the test, and writes the touchstone faster and cleaner than search, for the price of a subscription. Swap the keyboard shortcut for the chatbot and everything in this essay gets worse, and none of the structure changes.

What AI changes is the clock. An unproctored, open-book, retakeable test was always gameable; AI makes it gameable perfectly, invisibly, and at scale, which turns a chronic weakness into a terminal one. And it keeps moving the line: every model release pushes a little more of what used to take a person (reading, writing, analyzing) into the searchable zone, so the boundary between hackable and unhackable slides rightward, toward the formation layer, a bit more each year. AI is the fourth of the four pressures I keep naming, and it’s the one that sets how fast the other three arrive. Which is the real reason defending the first mile isn’t optional.

Steven Mintz, the historian at UT Austin, spelled out the operational version of this, and he frames it as a fork I recognize: two paths, one that makes the degree cheaper and faster, the other that makes the education deeper. The throughput path is the one I’ve spent this whole essay describing, and he names the same state reforms feeding it, Maine’s ninety-credit degrees, Florida’s thinned-out general education, California’s simplified transfer. The other path is what defending the first mile looks like once you get specific, and the piece of it that bears directly on the checkpoint problem is his account of assessment.

Mintz’s claim is that the take-home paper was never a good measure of what a student knows, and AI only made that impossible to keep overlooking. Detection software is the wrong fight; the right one is assessment an instructor can watch happen: oral examinations, staged projects followed from proposal through revision, in-class work where a student is handed an unfamiliar document and has to interpret it and defend the reading under questioning. You can’t CTRL-F your way through an oral exam. Which is the one hopeful note in all of this. The defense against the hack and the defense against the chatbot turn out to be the same move, and it happens to be a description of good teaching: put the work back in the room, where someone can see it being done.

---

When Wallack finished his environmental science course, he noticed the platform let him mark a “favorite instructor,” with a little photo and a name at the top of the page. So he tracked the instructor down. A person named Jensen Morgan, in Colorado, who’d recorded some narrated videos for the course more than a decade ago and had nothing to do with it since. “I have no contact with any students,” Morgan told him. “I never did.”

That’s the whole thing, right there, isn’t it? At the hackable end of the gradient, the credential certifies your passage through a course that has no teacher. Not a bad teacher, not an overworked teacher. No teacher. An avatar with a smile and a favorite-instructor star, attached to videos nobody watches, in a course you finish by searching the textbook you didn’t read. And the rule now being finalized in Washington bends, in its flat bureaucratic way, toward making that transcript count: presuming those credits, protecting their speed, welcoming the providers who sell them.

The students who figured this out first weren’t cynics. They were reading the signal accurately and following incentives.

A slice of what we call higher education has become a document you can fill in an afternoon, and the people writing the rules have decided that’s fine, or at least that it’s not their job to say otherwise. The rest of higher education, the part where someone is actually on the other end, still exists. It’s just going to have to prove it now, in a way it didn’t have to before.

The hollowing of the product isn’t new, and I’ve never said it was; it’s been slowly, secularly underway for thirty years, through adjunctification and online conversion and enrollment-management math.

What’s new is that the last formal backstop is going. The accreditor’s review was supposed to be the floor under the word “bachelor’s,” the institutional promise that somebody checked. The rule doesn’t lower that floor so much as confirm it was already gone and write the absence into federal code. We’re not voting away a floor that was holding. We’re recording, officially, that there wasn’t one.

The next move to watch is the AIM formal proposal and the comment period, and the legal challenges that’ll come with them. But the strategic clock started the day consensus was reached, and the institutions acting like they have until 2027 are reading a different signal than their students are.

---

Sacred Cow BBQ covers higher ed, politics, institutions, and the trust questions underneath. The university map is the working tool behind a lot of this: 1,556 four-year institutions, four quadrants, the underlying data open. If you’re at one of them, find your dot. And if you’re a provost or dean staring at a transfer-credit policy this summer and wondering what it commits you to, I’d genuinely like to hear from you.

Video explainer link: