I Told the Department of Education Their Graduates Were Useless
The room went silent. But I'd been watching this crisis unfold for decades—starting on a factory floor in 1981
Nothing like telling the Assistant Deputy Secretary of Education that what America’s colleges were producing was useless to me—to get a room excited.
It was 2009, at the Stanford Research Institute in Palo Alto. I was the only person from industry at a long conference table surrounded by about two dozen educators—Jim Shelton (Department of Education), presidents of leading Silicon Valley and California universities, heads of charter school associations, teachers union leaders, and state regulators. Jim had invited us to discuss innovation in K-12 education.
We went around the table. One by one, they shared their innovations: using computers for “better testing,” optimizing assessment tools, and digitizing standardized exams. Everyone was playing nice, telling the Department of Education what they wanted to hear. I was sitting near Jim, one of the last to speak.
When my turn came, I didn’t hold back. “I’m not an educator. I’m a buyer. And what you’re producing, I can’t use it.”
The room went silent. Not a brief pause—a long, uncomfortable silence. Everyone turned to look at Jim.
He leaned forward. “Explain.”
“You’re creating the world’s best test-takers,” I continued. “Not thinkers. Your graduates arrive at HP looking for the one correct answer—because that’s what you’ve trained them to do for sixteen years. But there is no one correct answer in the real world. There are constraints, trade-offs, and ambiguity. And your system has conditioned them to panic in the face of it.”
Some university presidents tried to push back. I made it clear: “I’m not sharing education theory. I’m sharing my lived experience. This is ‘as is,’ not ‘what should be.’”
The reactions split. Half the room looked to Jim for guidance. The other half looked down at their papers, avoiding eye contact with anyone.
No one agreed with me. Not even partially. It was either silence or defensiveness.
At the end, I was politely thanked as we left. No one pulled me aside. But word got back to HP’s Government Affairs team—my guess is Jim reached out to them directly. I was then “asked” to do a few meetings with the Department of Education teams and with Jim himself.
That’s when the idea came up: could I modify my FIRE Innovation Framework for use at the local school level—attracting collaboration between parents, teachers, and principals? I created the toolset pro bono and gave it away. We tested it. It worked.
The US education system never adopted it. Educators in other countries did.
I wasn’t speculating when I told Jim the system was broken. I’d written about this exact problem many times, including in my book, Beyond The Obvious. In Chapter 10, I’d warned: “The old adage that ‘what was good enough for you when you were a kid is good enough for today’ gives comfort and justification that everything is fine.”
Everything is not fine.
That was 2009. Thirteen years before ChatGPT launched, and made everything exponentially worse.
How I Learned to Think (By Accident)
I didn’t learn thinking skills in a “Thinking Skills 101” class. I learned them by wrestling with problems woven throughout my education.
Math word problems forced me to translate real-world scenarios into solvable equations. Every problem was a constraint puzzle: “If Train A leaves Chicago at 60 mph and Train B leaves Denver at 45 mph...” You couldn’t just memorize a formula. You had to map the variables, identify the relationships, and reason your way to an answer. It was woven into every subject.
Then came four years of architecture in high school. Every project was an exercise in constraint-based problem solving. You’d get a plot of land with specific dimensions. A family with a particular number of kids. A budget that didn’t care about your creativity. Your job: design a house that worked within all those constraints simultaneously.
For my senior project in 1979, my then-girlfriend (who became my wife) defined the requirements: parents plus five kids. It had to fit on a real lot available in our area. The budget had to align with a theoretical salary. The project took the full nine months of my senior year. She still has those blueprints in her hope chest, pulling them out occasionally to tease me about my design skills at the time.
Every blueprint was a forcing function. Your ideas had to survive contact with reality. Bad reasoning meant an unbuildable house. Good reasoning meant a family got what they needed within what was possible. I was learning to think in constraints without knowing that’s what I was doing.
When I got to college and took my first formal logic course, something clicked. I loved it. The course gave me vocabulary and frameworks to understand what I’d already been doing intuitively through architecture. I wasn’t learning logic from scratch—I was naming and systematizing patterns I’d been using for years.
This is what education should do: weave thinking into every subject. Math taught me to reason through constraints. Architecture taught me to prototype and iterate. Logic taught me to name the patterns. Not one “thinking skills” class—but thinking embedded in everything.
I didn’t know it then, but I was building cognitive infrastructure that would change my life.
Phil McKinney
The Factory Floor
In 1980, I got married. A year later, my wife—a nursing student at the University of Evansville—was thrown by a patient during her shift as a nurse’s aide and suffered a severe back injury. I suspended my college studies and moved to Evansville to take care of her and work to cover the bills.
I landed at Anchor Industries, a canvas manufacturing company that had been around since the late 1800s. Many of the staff were related, and their families had worked there for generations. My architecture background fit naturally into my role: solving design problems for the canvas products they produced, documenting solutions with full blueprints.
John Daus, the CEO and owner (whose grandfather had founded the company), saw something in me that I didn’t fully understand myself: confidence. But it wasn’t confidence in the traditional sense—it was comfort with uncertainty. I’d learned early that there’s no one correct answer to design or architecture problems. You had to be willing to test and experiment with your ideas.
John recognized that Anchor needed to constantly push boundaries or die. He was willing to take risks and experiment. He and his wife became like second parents to us, newlyweds struggling to make it. I’m not sure what would have happened if Anchor hadn’t given me a job.
What I didn’t know at the time was that I was continuing to learn and refine thinking skills—divergent thinking (generating lots of ideas), convergent thinking (narrowing down choices), and many others. The recognition I received from John was a key motivator. I loved conceiving, designing, testing, and then producing ideas.
My drafting table was on the second floor of the main building, with big windows overlooking the production floor. I could see what was being done, but I was curious. So I went down, pulled up a chair, and literally watched them work.
The soft goods industry operates on “piece rate” metrics—workers are paid based on the number of finished units they produce, rather than the hours they work. So having someone from engineering sitting there watching felt natural. For all they knew, I was doing a time study.
I watched for days.
Anchor had a growing business in custom pool covers. Back in the ‘60s and ‘70s, nobody had simple oval or square pools—they were all custom-shaped with rocks, waterfalls, and irregular curves. Pool builders from around the world would send blueprints and order forms. Those orders would queue into production, and I watched the same process repeat:
A team would clear a spot on the production floor and use chalk to draw the pool’s outline full-size—sometimes 20, 30, 40 feet across. Once the outline was chalked, they’d roll big spools of material over the chalk lines, get on their knees, and cut the pool cover by hand. Then they’d bundle the material in the correct sequence, send it to sewers or welders who would hand-feed it through their machines, then to finishing for grommets, then to QA, and finally to boxing and shipping.
They could do two pool covers a day. Maybe.
The bottleneck was obvious: chalking, cutting, and bundling. Each cover was labor-intensive and cognitively exhausting. Translating a 2D blueprint into a full-scale chalk drawing, then figuring out how to cut material efficiently to minimize waste while accounting for seams, webbing, stress points, and assembly—it required intense spatial reasoning every single time.
After watching for days, I went to the cutting table and started asking questions. What were their constraints? Width of material, waste allowances for seams and welds, stress points, and structural requirements. I needed to understand the boundaries of what was actually possible.
Then I saw it. This wasn’t an experienced labor problem—it was a computational geometry problem.
I was a computer science major and had worked as an undergraduate research assistant to a computer graphics professor. So I went home and had what I thought was a “crazy idea.” I wrote a program on an Apple II in Applesoft BASIC. You’d type in the information from the blueprint along with the material constraints. Within minutes, the program would output cut instructions that the cutting table could use directly—no chalking, no full-scale drawings, no mental gymnastics.
I built a quick prototype, then “borrowed” university computer time to create a more robust version using better tools than what PCs had at the time. I worked on it for weeks, iterating and refining. I didn’t know any better—this was my first production program. I just assumed that if I worked hard enough, I could solve any problems that came up.
When I had something worth showing, I printed out the first cut sheet instructions on a dot-matrix printer and brought them to John Daus, our CEO.
“I’ve got this crazy idea,” I told him. He knew I was a computer science major, so he was curious.
“Interesting,” he said, studying the printout. “Let’s do five covers with the craziest shapes and make them up. See how they turn out.”
We walked down to the cutting table together to get feedback from the cutting manager. They gave it a big thumbs up.
The engineering team—four people total, including my manager—couldn’t believe the program could account for all the constraints. But once we showed them the output and they started producing real covers, everyone jumped on it.
The secretary of the engineering group was assigned to input the pool cover information. She was thrilled. For many on the production floor, it was the first time they’d seen a PC, much less seen it do something like this. During breaks, workers would come up to see what I was doing. Some brought their kids in to see it.
The production team had initially feared computers would replace them. Instead, the program allowed Anchor to grow, creating more opportunities for them to work on other specialized projects that required their unique skills—such as circus tents, custom industrial covers, and complex fabrication work.
Within ninety days, production went from two covers per day to eighteen. Eventually, it hit thirty per day—a fifteenfold improvement.
More than the numbers, though, was what it meant. Anchor built a reputation as having the fastest turnaround for pool covers—especially in the critical fall season when everyone was shutting down their pools for winter. They became the go-to supplier worldwide.
The engineering secretary? She eventually went back to college, got a degree in computer science, and became the “CIO” of Anchor Industries.
The program became my undergraduate senior project—one of the early uses of PC graphics in solving design constraint problems, way before Adobe or any other graphic design systems existed.
Here’s what still amazes me: Anchor reached out ten years later, asking for help. Apple hadn’t made Apple IIs in years, but Anchor had been stockpiling parts to keep the system running. They finally realized they had to move to something new. My solution had been so effective that they’d kept it alive for a decade through sheer determination.
And here’s the kicker: articles were written about the pool cover program. One of those articles was seen by a man named Bob Davis. That’s how he reached out to hire me when I moved back to Chicago in 1982. The pool cover program opened the door to what came next.
I didn’t just make them faster. I changed what was possible. And I’d done it by applying thinking skills I’d been developing since high school—observation, constraint mapping, domain transfer, prototyping, iteration.
The Decision We Made in 1988
What I didn’t tell Jim Shelton at the Department of Education—or anyone else in that room in 2009—was that my wife and I had made a similar calculation twenty-one years earlier.
In 1988, when our oldest daughter was four and already reading, we decided to homeschool. She was reading everything—books, signs, anything with words. When we brought her to the school, they had no ability to accommodate her. They were designed for standardization, not thinking.
Not because we were anti-education, but because we’d already seen what I was now telling them: the system was optimizing for the wrong thing.
We chose a different path.
That was thirty-seven years ago. We’ve been watching the education system decline ever since.
The Lunch That Changed Everything
In 1987, I started my first consulting company, Millennium Partners, Inc, after a number of years in Silicon Valley. I’d been applying thinking skills naturally to my work, but I needed to articulate what I actually did. So I brought together a handful of friends who worked at Apple, Intel, HP, Roam, and 3Com for lunch at St. John’s in Sunnyvale.
I walked them through all the projects I’d worked on—Anchor, and others. I explained how I wanted to do more work like this. Then I asked: “What do I call it? How should I describe it?”
What followed was two hours of them asking questions, me answering, testing ideas, and then another round of questions. It was thinking skills in action—divergent exploration, convergent refinement, iterative testing.
That lunch produced the first time I used the word “innovation” to describe what I did. It was the birth of “innovation consulting” as my service offering.
But I didn’t stop there. I continued hosting lunches with my friends to break down the components underlying “innovation.” What were the actual skills? We identified brainstorming, constraint-based thinking, logical thinking, problem solving, design constraints, and more.
I treated “What is innovation?” as a design problem. I observed my own work, mapped the patterns, prototyped explanations, and iterated until I had a teachable framework.
My clients—Ashton-Tate, Roam, Intel, HP, IBM Research, and Naval Research—were paying me to teach thinking skills, even if they called it innovation consulting. These weren’t companies that needed motivation. They needed methodology.
I’d used thinking skills to understand thinking skills. The same process I’d used to solve the pool cover problem, I now used to deconstruct innovation itself.
Two Hundred Interns, One Impossible Problem
By the time I became CTO at HP, I’d spent thirty years building, refining, and teaching thinking skills. And I watched the best students in the world arrive at our Cupertino campus completely unprepared for problems without answers.
Every summer, we brought in approximately two hundred interns from elite universities around the world. Brilliant kids: perfect GPAs, top-tier schools, glowing recommendations. HP had a reputation for running one of the best internship programs in tech—we hired many of the interns after their summer with us.
During the first week, we’d break them into smaller teams and give them a product design problem.
As I watched, I saw them attacking it like a math test. They were searching for the one correct answer. Everything about the problem sounded solvable—the specs were clear, the constraints seemed reasonable. But there was a catch: buried in the problem was a constraint that made it impossible. You wouldn’t discover this by thinking about it. You’d only find it by building a prototype and testing it.
We gave them a week. They Googled. They whiteboarded. They debated. They pulled all-nighters searching for “the answer.” They produced beautiful slide decks explaining their solutions.
On day seven, when we told them there was no answer—that we’d deliberately given them an unsolvable problem—the room filled with whispers and murmuring. They were furious.
I’d tested this approach before rolling it out widely. Each summer, I invited two interns to live with me for the summer—a concept I called “reverse mentoring.” I learned as much from them as they learned from me. Through those experiences, I refined how to break their conditioning without breaking their spirit.
“Why would you do that?” one intern demanded. “We wasted a week!”
No, I told them. You learned something more valuable than any answer we could have given you. You learned that life isn’t a math test. There isn’t always one correct answer. Sometimes the problem itself is wrong. And if you can’t think your way through ambiguity, you’ll spend your career chasing solutions that don’t exist.
Some had egos and thought they knew everything—oh, to be young again. But we didn’t coddle. We invested time, energy, and mentorship.
After the unsolvable problem broke their mental model, we spent the summer rebuilding it. We taught them divergent thinking (generate possibilities), convergent thinking (narrow choices), logical thinking (reason through constraints), systems thinking (see connections), and design thinking (prototype and iterate).
But we didn’t just lecture. We pushed them to become experimenters. We paired them with teams in HP Labs working on real R&D problems. We forced them to build, test, fail, and refine.
Our culture was all about collaboration, not hierarchy. Titles didn’t matter—everyone was welcome to contribute. By the end of the summer, they got it. They were totally different people. Week one, they were seeking answers. In week twelve, they were comfortable with uncertainty. They all wanted to work at HP after the experience. Many did. Many are still there.
One intern who came back multiple summers is now one of the leaders in Silicon Valley’s gaming industry. Another is now an executive at a major Hollywood studio. There are dozens of similar stories.
One intern later wrote: “Phil taught me that the struggle to find the answer is more valuable than the answer itself. That summer changed how I approach every problem in my career.”
The ones who got it—who leaned into the uncertainty instead of resisting it—transformed. By the end of the summer, the anger was gone. In its place: confidence. Not the fake confidence of having “the answer,” but the real confidence of knowing how to think when there is no answer.
Years later, some would tell me: “That summer changed my career. I didn’t learn what to think—I learned how to think.”
But most struggled. And each year, I watched the incoming class get worse. More reliant on Google. Less comfortable sitting with a problem. More conditioned to seek the answer rather than develop the thinking.
This went on for years. Each summer, a new batch of two hundred interns. Each year, the problem got worse.
That was the early 2000s—before ChatGPT, before large language models, before AI could write your code and design your products and answer your questions in seconds.
Then Everything Accelerated
Now? The friction is almost entirely gone.
I watch students use AI to skip every step of the thinking process. Don’t observe—ask AI what the problem is. Don’t map constraints—ask AI for solutions. Don’t prototype—ask AI to generate the design. Don’t iterate—ask AI to optimize.
They get answers. But they don’t build thinking skills.
And here’s the brutal truth: In a world where AI can generate answers instantly, the only sustainable competitive advantage is your ability to ask better questions, see hidden constraints, and think in ways machines can’t replicate.
The interns I worked with in the 2000s could at least be un-taught. We had a summer. We could break their conditioning and rebuild their cognitive foundation.
Today’s generation? They’re not just conditioned by schools—they’re conditioned by tools that make thinking optional. And most of them are opting out.
When I sat across from Jim Shelton in 2009 and told him the education system was producing graduates I couldn’t use, I wasn’t exaggerating. I had decades of evidence. I’d run the experiments. I’d tried to fix it—through the HP intern program, through the book I published in 2012, through the brainstorming process I created for parents and teachers.
The US education system ignored every solution. Other countries adopted them.
Thirteen years after that meeting, ChatGPT launched. The crisis I’d been warning about for decades became undeniable—and exponentially worse.
The Path Forward
Here’s what people miss: We don’t need thinking skills classes in K-12. We need to weave thinking into every subject. Make math about problem-solving, not formula memorization. Make history about analyzing causation, not memorizing dates. Make science about experimentation, not reciting facts.
Eliminate the test culture that prioritizes memorization over critical thinking. Bring back the struggle.
But here’s the reality: If you’re reading this, your K-12 education already happened. You can’t go back and spend four years in architecture classes wrestling with constraints. You can’t redo your formal logic course. You can’t sit on a factory floor for weeks observing problems.
So here’s what I’m offering: the compressed version. The patterns I spent decades discovering, you can learn in hours. The frameworks I had to reverse-engineer from my own work, you get handed to you explicitly. The ten thousand hours of constraint-based problem-solving? I’m giving you the cognitive shortcuts.
That’s what the YouTube series is. It’s thinking skills on an accelerated timeline.
Here is Episode 1: Why Thinking Skills Matter Now More Than Ever
I’ll teach you the frameworks—divergent thinking, systems thinking, constraint mapping, all the components I spent decades deconstructing. But here’s the catch: frameworks are just vocabulary. The skill comes from use. You still have to wrestle with real problems. You still have to sit with uncertainty. You still have to prototype, fail, and iterate.
Subscribe to the YouTube channel so you don't miss a single episode in the series.
I’m giving you the map and showing you the shortcuts. But you still have to walk the terrain.
The difference? Instead of wandering for years, figuring out the patterns yourself, you’ll know what to look for. Instead of ten thousand hours of trial and error, you’ll practice deliberately. Instead of accidentally stumbling into thinking skills like I did, you’ll build them systematically.
Because here’s what’s coming: AI will get better. The tools will get more powerful. And the gap between people who can think and people who can only prompt will become a chasm.
Those who can think—who can observe what others miss, question what others accept, and navigate ambiguity—will thrive.
Those who can’t? They’ll be automated, manipulated, or left behind.
I didn’t learn thinking skills from a textbook. I learned them by sitting on a factory floor for days, watching, questioning, and prototyping. By spending nine months designing a house that would never be built. By working through constraints I couldn’t shortcut.
The HP interns didn’t learn thinking skills from a lecture. They learned them by failing at an impossible problem, then spending a summer experimenting with real R&D teams.
The struggle wasn’t wasted time. The struggle was the skill.
In 2009, I told Jim Shelton the system was broken. He knew. Nothing changed.
Thirteen years later, ChatGPT launched.
Now it’s not just broken—it’s obsolete.
Don’t wait for the system to fix itself. It won’t.
You can spend years figuring this out the hard way. Or you can learn the patterns now and start applying them immediately.
Your choice. But choose quickly.
Because the interns I worked with in the 2000s at least knew they had a problem. Today’s generation doesn’t even realize what they’re losing.
Phil McKinney
