Healthcare, Government, HR, and Education Innovation Keeps Failing for One Reason; The Sectors Th...
In 2016, the U.S. Government Accountability Office reported that the Department of Defense was coordinating America’s nuclear forces on a 1970s IBM computer that stored data on 8-inch floppy disks. The same review found the federal government spent about 75 percent of an IT budget exceeding $80 billion on operations and maintenance, and that this “maintain the status quo” spending kept rising while modernization spending fell. Years later, the number had moved in the wrong direction; GAO now reports that agencies typically spend about 80 percent on operating and maintaining existing IT, including legacy systems.
If you have faxed a medical record (something which my family experiencing cancer recently had to do, in 2026), retyped your entire résumé into an application system immediately after uploading your résumé, or watched your kid sit through a class that peer-reviewed research proved wrong, you have met my other three observations.
Healthcare, government, workforce and HR, and education are the four sectors that most decide the quality of a person’s life, and they are the four most resistant to real innovation.
Each one now claims to be a “tech” sector, and each one makes that claim with little more than a digitized record of the paperwork it already had. I will applaud anyone who takes a serious swing at disrupting any of them, because nobody can know in advance which attempt is right, and we can already measure that what most people experience today is bad.
These four sectors were built to make people readable to whoever pays, and the person being read (you) has almost never been the one paying but it is your data and this is a service for you; I think that single fact explains why their “innovation” keeps ending at better record-keeping.
A Paperwork Trap Explains Why Control and Care Look Identical
An interesting version of this argument says these institutions resist change because they exist to control their market.
Political scientist James C. Scott spent a career studying why grand state schemes fail, and in Seeing Like a State he argued that states have always needed to make people countable from a desk. Scott described government efforts to arrange a population so taxation, conscription, and the prevention of rebellion became simpler to administer. Permanent surnames, standardized measures, land surveys; every one made people processable by someone who would never meet them.
I call the failure pattern in these four sectors the Paperwork Trap; an institution measures what makes a person fit its paperwork (the chart, the case file, the personnel record, the transcript), the paperwork is filled out for someone other than the person (an insurer, a legislature, an employer, an accreditor), and so every technology purchase gets judged by whether it makes the paperwork better.
Control and care look identical from the outside because both depend on the same readable person. A hospital needs your diagnosis coded for the same reason the IRS needs your W-2 (the form employers file to report your wages); a paying party needs you standardized before it can process you. Nobody had to conspire; the institutions optimized for their actual customer, and the patient, citizen, employee, and student became the inventory.
I spent the early part of my career in search marketing, and this is the 2005 SEO problem wearing a lab coat. Pages back then were stuffed with keywords for Googlebot, and the human who clicked through got promotional junk that ranked well only because the creator know how to use the system. The page performed for its real audience; Google fixed it only when people got pissed off with all the spam.
These four sectors have never had their Google moment.
Healthcare Digitized the Bill and Called It Patient Care
A 2019, William Shrank, Teresa (Terry) Rogstad, and Natasha Parekh estimated U.S. health care waste at $760 billion to $935 billion a year, roughly a quarter of all spending, with administrative complexity (inefficient rules, forms, and requirements) the single largest category at $265.6 billion.
They found no studies exploring innovation aimed at administrative complexity. Where are you founders?? The biggest pile of waste in the system had not even been studied as a thing to fix. Shrank called the administrative figure “striking,” in a word that says to a market, here is an outstanding opportunity and it is shocking no one is doing anything.
The electronic health record is the sector’s proof that it “adopted tech.” If like me and you’ve had to use MyChart, across a half dozen different providers, you have to be wondering why it’s stuck in 2005 and can’t even function as one app for YOUR care. Ask more rigorously and you get the answer; the coder, the auditor, and the insurer. Your experience of trying to see a doctor, get a price, or move your own records barely changed, because the record improved for its reader.
The symptom: cost, wait times, and paperwork.
The cause: the patient is the subject of the transaction, and the payer is its customer.
A disruption worth applauding: anything that makes the patient the buyer, including direct primary care, price transparency, and tools that hand the record to the person it describes.
Government Modernizes by Maintaining
Let me clarify first that I don’t want to explore here how government is a cause or path to fixing healthcare, I want you to consider how government as a sector suffers from the same thing. This is a call for government innovation and for us to stop pretending GovTech has fixed much of anything.
When four of every five IT dollars go to keeping existing systems running, “modernization” means preserving the current way of processing citizens on newer hardware. GAO’s IT lead Dave Powner told the House oversight committee that “clearly, there are billions wasted.”
The cost to entrepreneurship is measured in years.
I recently walked through how the DEAL Act took eight years to edit a single definition in securities regulation, and in that piece, I described the Permission Layer; the legal definitions, licenses, and funding rules that decide whether a customer may buy from you at all. Government shapes every transaction a founder makes before the first hire, and when the Office of Management and Budget floated the ability to cancel federal grants for any reason, it demonstrated how little the system values the person on the receiving end.
Policy professionals should stop treating digital services as the innovation goal. A beautiful portal sitting on an 80 percent maintenance budget is a new storefront on the same warehouse.
Workforce and HR Records Every Door and Ignores the Hallway
Gallup’s State of the Global Workplace 2026 report found global employee engagement fell to 20 percent in 2025, its lowest level since 2020, at an estimated cost of $10 trillion in lost productivity. The United States does better at 32 percent engaged, which means roughly two of every three workers are not.
HR software is the paperwork layer for employment. In The Intellectual History of HR that helped lead me this article, I read fifty years of peer-reviewed research and found that HR systems were designed to record events (hire date, pay change, termination) and almost nothing that happens tying them together. That is why I wrote that HR Tech behaves like a coroner when it should behave like a doctor (ironic in my assessment of healthcare); it documents an employee exit but the person, or the company, already decided to part, and nothing helps either avoid that outcome.
Dartmouth’s Vijay Govindarajan (VG) said of HR teams in 2007, “They create processes. They are viewed as a nuisance.” The processes exist because the employer, the auditor, and the regulator need to read the workforce, and every hire starts a compliance clock that rewards documentation.
Hiring is where the Paperwork Trap does its most expensive damage to the economy.
Today’s systems are screening for keywords, titles, and credentials, which, like a person’s health, are subjective circumstances; people are not datapoints. Just as Google worked through evolving from a directory of webpages to an algorithm of better results, we should be demanding an explanation of why workforce tech is stuck being a data record, and we should be fearful of why everyone hopes AI will change it, when the technical capability to actually innovate employment exists.
The Pipeline That Feeds the Other Three: Education
Scott Freeman analyzed 225 studies in 2014 and found students in traditional lecture courses were 1.5 times more likely to fail than students in active-learning courses (classes where students solve problems and get feedback during the session).
The authors wrote that the results “raise questions about the continued use of traditional lecturing.”
A decade later, walk into almost any lecture hall. I’ve been in nearly 100 startup accelerator classes this year. Sit in a k-12 class.
Education’s “tech” is the learning management system, which digitized the gradebook. The gradebook feeds the transcript, and the transcript is paperwork written for the next reader; the admissions office. Which is then… an employer’s ATS.
We spend twelve years training entrepreneurship out of people and then sell classes and incubators claiming to put it back, because mass schooling was built to produce punctual, rule-following, interchangeable employees, and it does this extremely well. That essay traces the divergent-thinking decline George Land reported and Sir Ken Robinson popularized by exploring how schools kill wonder. The system rewards the single correct answer because single correct answers are the easiest thing to grade.
1942 Wired Three of These Sectors Together, and Education Supplies the Inputs
The four sectors resist change individually, but they also hold each other in place.
During World War II, the Stabilization Act of 1942 imposed wage controls, and the National War Labor Board let companies use fringe benefits such as health insurance to recruit and keep scarce workers. By 1957 more than three quarters of Americans had health insurance, mostly through employers.
A government wage rule (government) turned the employer (HR) into the gatekeeper of medical coverage (healthcare), and the school system (education) produced the credentialed employee who would walk through that gate.
Society now suffers with job lock; the reluctance to leave a job because leaving means losing coverage.
The interlock is the reason a single disruptive company in one sector so often fails.
A founder who wants to leave employment to build something loses health coverage, a would-be patient who wants a better care model is tied to the employer’s plan, and a student who skips the credential becomes unreadable to the hiring filter.
Leaving any one system costs you standing in the other three.
Venture Capital Avoids These Sectors Because Founders and Investors Believe the Buyer Is the Institution
Innovation requires recognizing that the buyer is a person.
Mercatus Center research by Liya Palagashvili found that venture firms tend to avoid startups in heavily regulated industries, and that regulation can push startups to pivot out of fields such as medical technology entirely.
Missing capital here is a symptom. The root cause is a belief, shared by founders and investors, that the only viable customer in these sectors is the institution itself; the hospital system, the government agency, the HR department, the school district – the very thing that WON’T CHANGE. Founders who hold that belief build tools that make the institution better at processing people (billing software, case management, compliance dashboards, gradebooks), because that is what the institution will buy, and investors fund those companies because procurement cycles are at least predictable.
That belief is why so much “HealthTech,” “GovTech,” “HR tech,” and “EdTech” amounts to the Paperwork Trap with a subscription fee. The market shows up for paperwork tools, which reinforces the belief, which sends the next founder toward the next paperwork tool. Venture capital avoids ecosystems where real solutions can’t win; no one is interested in another data platform, even when it does something as competitively irrelevant as adding “AI” to the game (which is what all four sectors are doing now; pretending that is innovative).
Innovation is Change, Not Solutions
Founders: build for the person being processed and treat the institution as a channel at most; direct-to-patient, direct-to-citizen, direct-to-worker, and direct-to-learner models are where the disruption lives. A policy layer belongs in your go-to-market plan because fixing these sectors is going to require changing policy.
Investors: stop discounting these sectors as “regulated, therefore slow,” and start appreciating again what Silicon Valley (frankly) taught us (and you), that disruption is a key word in innovation that means changing things to fix them.
Policymakers: decouple inter-dependencies because every link in the 1942 chain raises the cost of trying something new. Machine learning and predictive intelligence (not AI!) can provide infrastructure that enables new business models and solutions but you have to let that happen.
HR and education leaders: buy tools that tell you something about people and judge by how the people actually performed as individuals, not standards; workforce innovation in harder markets shows what happens when constraints refuse lazy assumptions.
Applauding the Attempt Is the Only Defensible Position When the Baseline Already Fails
Floppy disks coordinating nuclear forces, a quarter of health spending lost to waste in its biggest category never even studied for a fix, four of five federal IT dollars spent standing still, two of every three American workers not engaged, and classrooms producing failure rates that research is actively indicting.
Every one of those “tech” solutions is code, not innovation.
Google cleaned up search when it changed what it rewarded, and these four sectors will change when we change who they serve. That is why I will cheer the founder who fails trying to make the patient, the citizen, the worker, or the student the customer; the valuable attempt is a test of the right hypothesis because the status quo is failing in public.