If you are around my age, 56, you may remember the Looney Tunes cartoon with Ralph Wolf and Sam Sheepdog. Each morning, Ralph and Sam meet at the base of a hill, exchange pleasant greetings, “Morning, Ralph.” “Morning, Sam.” and then punch a time clock mounted on a fence post. The moment the whistle blows, Ralph charges the flock with everything he has, Sam beats him senseless, and the mayhem continues until the end-of-day whistle sounds. Then they clock out. “Night, Ralph.” “Night, Sam.” Same time tomorrow. No hard feelings.
I have thought about that cartoon while watching Philadelphia’s education wars play out over the past three decades. On one side: the teachers unions and their Democratic allies, thundering about underfunding, equity, and the sanctity of public schools. On the other: the school choice advocates and their billionaire backers, thundering about failing schools, parental rights, and the moral imperative of escape hatches. Both sides are perpetually outraged, perpetually organized, and perpetually well-funded. And the children of Kensington and Strawberry Mansion and West Philadelphia are, year after year, perpetually left behind.
To an observer, like me, watching from the peanut gallery, the drama seems intentional, structurally designed, and not a coincidence. The fighters need each other. The litigation requires an adversary. The fundraising letter requires a villain. The rally requires an opposition. Ralph and Sam clock in every morning, perform their choreographed battle, and the flock, the children, the actual object of the exercise, remains right where it was.
Meanwhile, online learning materials are available for free like an unending buffet waiting to be consumed. Organizations like MIT have put their entire curriculum online for free. Khan Academy has 189.6 million registered users. But almost nobody in Philadelphia’s education establishment mentions these free online resources to the families who need them most. Or, teaches our neighbors to organize themselves around learning, building cultures of intellectual aspiration and strong household academic culture, or developing neighborhood mentorship networks.
Performance politics
The School District of Philadelphia now spends approximately $42,621 per student annually in total expenditures, a figure the district’s own financial reports have documented, though district officials prefer the lower $28,736 number that excludes debt service and capital outlays. By either measure, per-pupil spending has more than doubled since 2016, when the district spent $17,040 per student. Pennsylvania taxpayers statewide are spending $21,985 per pupil in the 2024–25 school year, the highest in state history, seventh-highest nationally, while serving 39,000 fewer students than in 2019.
The return on this investment is not great. PSSA scores released for the 2024–25 school year show that roughly one quarter of Philadelphia students in grades 3–8 score proficient or above in math, and approximately one third in reading. That is the “best” recent year. The state-mandated Basic Education Funding Commission determined that Philadelphia is underfunded by at least $1.4 billion annually, a genuine grievance, and yet the district ranks 454th out of 499 Pennsylvania school districts in math proficiency. The statewide NAEP results show no measurable improvement since 2003 in closing achievement gaps exceeding 25 percentage points for Black students, Hispanic students, and economically disadvantaged students. None. In twenty years.
READ: The Review by Ajay Raju | America’s future looks more like a rice paddy than a frontier (July 2, 2026)
Eighty-five percent of Philadelphia eighth graders are not proficient in math. Eighty-two percent are not proficient in reading. These are not projections or estimates, but are the actual scores.
In the same period, the district added staff while losing students. Total non-charter enrollment dropped from 134,227 in 2016 to 117,956 in 2025, a 12 percent decrease. Class sizes shrank from 16.4 students per teacher to 13.9. Meanwhile, the district, with more than 20,500 employees, is now one of the largest employers in Philadelphia, surpassing Temple University, Comcast, SEPTA, and Drexel University.
The money has been spent. The outcomes have not followed. This is the fact that both sides in the education war treat as an inconvenient footnote rather than the central question.
The Union
The Philadelphia Federation of Teachers is a serious political organization. They represent teachers, most of whom are overworked, underpaid relative to their suburban counterparts, and genuinely dedicated. PFT, its state affiliate the PSEA, and their national parents the NEA and AFT are political powerhouses.
During the 2023–24 election cycle, the four largest public-sector unions, NEA, AFT, SEIU, and AFSCME, spent a combined $915 million on elections and progressive political activism, 86 percent of which came from member dues. In 2024 alone, the NEA spent $432 million in total, but only 9 percent, not even $40 million, on teacher representation, the core purpose of the organization. Its spending on politics and other contributions was more than four times higher than its spending on representing the teachers who paid its dues.
The PSEA, Pennsylvania’s largest teachers union, spent just one dollar in five of member dues on representing teachers and staff. The rest went to administration, politics, and lobbying. From 2021 to 2023, PSEA lost nearly 2,000 members while its revenue grew from $76 million to $81 million and its net assets expanded from $125 million to $161 million. The union is thriving, but the members are leaving. Its political influence is vast. And the students in the districts it serves are still mostly not proficient in math or reading.
The Choice Advocates
The school choice side of this war is no less interesting as a study in motivated reasoning.
Jeffrey Yass is the wealthiest person in Pennsylvania, a brilliant billionaire options trader who founded Susquehanna International Group and has made school choice the defining cause of his political spending. Yass has given nearly $47 million to his Students First PAC, which flows through the Commonwealth Children’s Choice Fund into the Commonwealth Leaders Fund, which in turn has become one of the most powerful forces in Pennsylvania Republican politics. He spent $35 million on Pennsylvania state elections in the 2024 cycle alone, more than the top ten corporate PACs in the state combined. His PACs contributed $13.4 million during the 2024 cycle to Republican candidates up and down the ballot, including $1.4 million to Senate President Pro Tem Kim Ward, who ran unopposed.
The argument Yass and his allies make is sincere in important respects. Philadelphia’s public schools are failing. Families in those schools deserve alternatives. Children assigned to a failing school because of their zip code are victims of a system that prioritizes institutional preservation over their welfare. These propositions are defensible.
READ: The Review by Ajay Raju | America’s future looks more like a rice paddy than a frontier (July 2, 2026)
But what has the school choice movement actually produced in Philadelphia? The School District of Philadelphia has not approved a new charter school since 2018. The average payment to the district’s 79,357 charter students represents $19,211 per student, 60 cents for every dollar the district spends. Cyber charter schools were, until recently, charging the district between $8,639 and $26,564 per student per year, prompting Governor Shapiro to cap cyber tuition at $8,000, a reform estimated to save school districts $262 million annually. For years, the choice ecosystem was not primarily an education reform movement, but a billing operation.
More fundamentally: neither the union nor the choice advocates has a compelling answer to the question of what happens to the child who is too poor, too isolated, too unsupported at home to navigate either system well. The union’s answer is more money to the district. The choice advocate’s answer is a voucher. Neither answer includes: “Let me show you Khan Academy.”
The Nonprofit and Political Insider Economy
Philadelphia has one of the most elaborate nonprofit education advocacy ecosystems of any American city. Children First, formerly Public Citizens for Children and Youth, has published annual reports on Philadelphia children since 1982. The Education Law Center, which brought the landmark 2023 Commonwealth Court case that declared Pennsylvania’s school funding system unconstitutional, is staffed with experienced lawyers and advocates. Dozens of community organizations, civic engagement groups, and research nonprofits populate the landscape.
These organizations do genuine and important work. The 2023 Commonwealth Court ruling, finding that Pennsylvania’s funding system was unconstitutional, was a genuine legal and civic achievement. The adequacy gap, the $4.5 billion shortfall between what Philadelphia schools receive and what the state’s own commission says they need, is real.
But there’s a mismatch with the incentive structure in this ecosystem. The advocacy organizations exist to advocate. They measure their success by legislation passed, lawsuits won, budget lines expanded. Their fundraising letters are built around the persistence of the crisis. Their relationships with legislators are cultivated through the machinery of the budget fight. If the crisis were resolved, if Philadelphia children were actually learning at high levels, the advocacy apparatus would have to reconstitute itself around a different purpose. The crisis is not in anyone’s professional interest to end. It is, however, in many people’s professional interest to manage.
This is the insider extraction pattern that defines Philadelphia’s civic economy across multiple sectors, and education is its most consequential theater. A $4.5 billion adequacy gap generates conferences, reports, rallies, lawsuits, legislative testimony, media appearances, campaign contributions, and endorsed candidates. It does not generate a single parent being told about MIT OpenCourseWare.
Free Buffet
Neither side of this war is focused on certain solutions that could make an impact.
For example, MIT’s OpenCourseWare platform has put more than 2,500 courses online for free, the full curriculum of one of the world’s most elite universities, including lecture notes, problem sets, exams, and video lectures from world-renowned professors, with no registration, no fees, and no start dates. The platform has reached over 500 million learners worldwide in its twenty-five-year history.
Meanwhile, Khan Academy logged 66.8 billion learning minutes in the 2024–25 school year across 104.9 million active learners globally. Students who use it for 30 minutes a week show learning gains 26 to 38 percent larger than expected. These FREE platforms work.
READ: The Review by Ajay Raju | Two wars, one nation, citizens divided (June 25, 2026)
And Sal Khan is not standing still. This spring, at TED2026, Khan Academy joined with TED and the testing organization ETS to unveil the Khan TED Institute, a new attempt to rebuild higher education itself around what a person can actually do rather than how many semesters they have sat in a chair. The pitch is a full bachelor’s degree, in applied AI to start, for under $10,000, built on three legs: core academic knowledge, applied AI skill, and the “durable” human skills, communication, collaboration, judgment, that no chatbot can fake its way through.
Students advance by demonstrating mastery, not by accumulating seat time, which means a disciplined student could plausibly finish in half the time a traditional degree demands. Khan is once again asking why access to serious learning should be gated by cost or geography at all, and he is asking it just as loudly at the level of the university as he has for two decades at the level of the free video.
And in Philadelphia, the city with a $4.5 billion adequacy gap, 85 percent eighth-grade math non-proficiency, and a budget that now exceeds $42,000 per enrolled student, when was the last time a union leader or a school choice advocate stood in front of a room of parents in Strawberry Mansion and said: “Khan Academy: free, it works, here’s how”? When was the last time a nonprofit education advocate spent an afternoon in a Kensington community center showing families how to navigate the platform?
The answer is almost never. Because there is no budget line for pointing people toward a free website. There is no foundation grant for sending a text message. There is no rally permit for showing a parent a URL.
Learning how to learn
If the buffet is free, the harder question is how a child should eat from it, and this is where the emerging debate over AI in the classroom actually matters, far more than the funding fights that consume the front pages.
Not every problem a student faces belongs on a screen. Some things are still best worked out with pen and paper, the long division that builds number sense, the proof that teaches a student to hold an argument in her head from premise to conclusion, the handwritten essay draft that forces her to think in complete sentences before a machine can smooth them for her. Other problems are precisely suited to the newest tools: a simulation that lets a student watch a gas law unfold instead of memorizing it, a tutor available at 9pm when the only adult in the house is exhausted from a second shift. The mistake both the techno-optimists and the techno-skeptics make is treating this as one question. It is two questions, and the answer to each is different. The real work of modern pedagogy is deciding, subject by subject, skill by skill, which problems still belong to the pencil and which belong to the machine.
Where AI does belong in a classroom, the evidence increasingly points to one design principle above all others: it should be Socratic, not declarative. An AI that hands a struggling eighth grader the answer to a math problem is not a tutor, it is a shortcut, and a shortcut is just cognitive offloading dressed up in a friendly interface, one vessel pouring its contents into another without either vessel doing any of the work that builds capacity. Khan Academy’s own Khanmigo was built around the opposite premise, an AI that answers a question with a question, that nudges a student to explain her own reasoning before it will confirm or correct it. Whether or not any single product gets the balance right, the design instinct is correct: the AI tools that will actually help a generation of students are the ones built to provoke thinking and invite collaboration between students, not the ones built to make thinking unnecessary.
This is also why the Khan TED Institute’s emphasis on “durable skills,” collaboration, persuasion, the give and take of peer feedback, is worth watching closely. Its architects have said explicitly that they want students working in teams, tutoring each other, and building things together across time zones, precisely because those are the human capacities that no model can substitute for and that most factory-model classrooms have never systematically taught at all. The ultimate goal of any education, in an age when a phone can generate an essay or solve an equation in seconds, is not the accumulation of facts. It is learning how to learn: knowing what kind of problem is in front of you, knowing which tool it calls for, and knowing how to work alongside both other people and intelligent machines without surrendering your own capacity to think. A teacher, in this model, becomes something closer to a proctor of that experience, someone who is guiding the encounter between student, peer, and machine, not someone who has been replaced by it, and certainly not someone who allows the machine to become a second, more efficient version of the copied homework that has plagued every generation of students before this one.
READ: The Review by Ajay Raju | The future beneath our feet — reimagining Philly’s subway systems (July 6, 2026)
None of this closes the funding gap or resolves the charter dispute, and it should not be mistaken for a substitute for either. But it is, unlike most of what passes for reform in Philadelphia’s education wars, actually free, actually available tonight, and actually aimed at the only outcome that has ever mattered: whether a child in Kensington or Strawberry Mansion learns to think for herself.
Why the hard problem is hard
It would be unfair to end the indictment without acknowledging the genuine difficulty of what is being asked.
The digital divide remains real: up to 12 million American students lack adequate internet access. But this is a $50-a-month problem, not a $42,000-per-student problem. More fundamentally, access is not the core barrier. The core barrier is cultural and social.
Consider what separates the communities that thrive academically from those that do not. It is not, primarily, spending per pupil, the highest-performing and lowest-performing districts in Pennsylvania both spend above the state average. It is the invisible architecture of expectation: the family organized around intellectual achievement, the peer group that treats academic striving as normal rather than deviant, the neighborhood subculture that reproduces ambition the way athletic communities reproduce athletic identity.
Research on social and cultural capital is consistent on this point. High-income families transmit not just money but navigational knowledge, the habits, networks, and institutional literacy that convert aspiration into attainment. The problem is not that low-income parents fail to want their children to succeed. The desire is there and it is fierce. The problem is that they lack what the literature calls the skills, habits, and knowledge to translate that desire into effective action. It is harder to ask a friend to join you at the library than at the bar. The communities that organize themselves around learning are the communities that learn.
This is the structural condition that no funding increase and no voucher program directly addresses. It requires a different kind of intervention: the patient, unglamorous work of building cultures of intellectual aspiration in communities where they do not yet exist. Study circles. Neighborhood mentorship networks. Adults who have navigated the system modeling that navigation for young people who haven’t. The Germination Project’s model, cultivating civic leaders from within communities, applied to the micro-level of household academic culture.
Same time tomorrow?
The PA Budget drama is over and legislators are now posting victory lap videos. There is some good news. School districts statewide will receive $11.8 billion in the coming year, an increase of roughly $670 million over last year. Basic education funding rises by $58 million to $8.32 billion; special education climbs $55 million to nearly $1.6 billion; career and technical education gets a modest bump as well. PSEA’s president called it a reflection of the state’s remarkable commitment to public education. House Majority Leader Matt Bradford, standing on the floor of a chamber that has spent the better part of a decade in some form of dysfunction, called the package “pro-growth” and said Pennsylvania was proving it could get things done even under divided government.
Meanwhile, the lawyers who won the 2023 case that started this whole multiyear “adequacy” ritual, the Education Law Center and the Public Interest Law Center, issued the same statement they issue every year: progress, but schools remain billions of dollars short of what the state’s own commission says is required, and some historically underfunded districts had to cut programs even this year. The Philadelphia Federation of Teachers welcomed the money while registering its disappointment that lawmakers couldn’t manage a statewide cellphone ban. On the other side of the aisle, the Commonwealth Foundation, Jeffrey Yass’s favored policy shop, complained that the deal did nothing to meaningfully expand school choice, and noted that for the first time in years the budget didn’t even grow the tax-credit scholarship programs choice advocates depend on.
Of course, nobody, on either side, mentioned online resources like Khan Academy. Nobody mentioned the scaffolding needed to develop a culture of learning in neighborhoods that need them most. The budget did find room, notably, for a new tax credit program to fund grants to life sciences and biotech startups, proof that when a cause has the right lobbyists, Harrisburg can move quickly and creatively. It found no such creativity for the only intervention that has ever been shown to move a proficiency number without first winning an election.
To be fair, an additional $670 million into the state’s public schools is not nothing. The Legislators are managing a genuine structural deficit, a state Supreme Court ruling on skill games, a rainy day fund they were determined to protect, and a constitutional mandate from a 2023 court decision that is not going away. I hope, genuinely, that this year’s dollars land differently than last year’s, and the year’s before that, and that five years from now some Philadelphia eighth grader’s math proficiency has actually moved off the number it has been stuck at for two decades.
But I have read enough of these press releases now, across enough budget cycles, to recognize the tune. Every side declares vindication. The union thanks the “pro-public-education leaders” it helped elect. The advocacy nonprofits mark the day as a step forward while carefully preserving the shortfall that justifies their next campaign. The choice lobby laments what it didn’t get, which keeps its own donors engaged for the next cycle. Everyone is right, in their own way, about their own slice of the fight. And the PSSA scores, when they’re released next spring, will very likely look almost exactly like the ones that came before them, because none of this year’s $670 million was aimed at the actual mechanism, the cultural architecture of aspiration in a household, that determines whether a child learns. It was aimed, as it is every year, at the fight itself.
Ralph and Sam clocked in again this July. They will clock in again next July. I hope, sincerely, to be proven wrong by the next set of test scores. I have not yet had occasion to be.
“Night, Ralph.” “Night, Sam.”