I wrote the list thirty years ago and was mostly right. The AI brought them current with today — in public, with numbers.
I wrote the list thirty years ago and was mostly right. The machine corrected the rest—in public, with numbers.
I created a list once.
Eight critical skills, drawn not from a survey but from the work itself—from nearly a thousand executive searches, where companies told me in plain terms the kind of skills they would pay real money to find.
The list was solid. The research validated. They held up for a long time.
I believed in it.
I even wrote a book about it: WANTED: Eight Critical Skills You Need to Succeed. Created a successful blog, too, with nearly 1,000 articles at criticalskillsblog.com.
The skills have remained solid and reliable over the years.
But the world the list was built for evolved, and a man who cannot say that his own work needs to keep up with the times has no business writing about skills at all.
So it’s time for an overhaul.
Understand that I’m not here to take anything back. I’m here to celebrate that set of skills—and align it with the present—and the future.
What the old lists got right
Start with the thing nobody wants to admit:
The frameworks have changed far less than the noise around them suggests.
The Department of Labor put out the SCANS report in 1991. It named five competencies and a foundation of basic skills, thinking skills, and personal qualities.
It was built the right way—not from a questionnaire mailed to executives, but from commissions that talked to business owners, supervisors, and workers, and watched the jobs being done. Read it today and the bones are familiar. Communication. Problem-solving. Working with people. Handling information. Using technology. Knowing how to learn.
Those bones never left. One framework after another has named them again across the decades—the World Economic Forum does it every two years, and acts surprised each time.
The vocabulary shifts. The substance sits still.
The proof is in their own pages. The Forum’s 2018 list of the skills that would matter most named analytical thinking, critical thinking, creativity, complex problem-solving, emotional intelligence.
Read its list from years later and you find the same words, announced again as discovery!
And even in 2018 it marked one skill as fading: the routine installation and maintenance of technology. Operating the machine was already losing its worth. The machine had not yet learned to talk back, and the writing was on the wall.
And it sits still for a reason. The economist David Deming showed it with hard numbers, not opinion: between 1980 and 2012, jobs demanding high social skill grew by almost twelve points as a share of the American workforce, while the math-heavy jobs that needed no people skills shrank. The strongest growth, the best wages, went to the work that demanded both a sharp mind and the ability to stand next to another human being and get something done together. That was true before I wrote my list. It is true now. Some things do not move.
So when a new report announces that communication and critical thinking are the skills of the future, I am not impressed. They were the skills of the past, too. The honest headline is duller and truer: most of what mattered still matters.
That is what the old lists got right. Mine included.
What changed underneath them
Here is what they could not have seen.
For thirty years, the “technology skill” meant: learn the tool. Learn the spreadsheet. Learn the database. Learn the syntax.
The skill was operating the machine, and the machine did exactly what you told it, no more, no less. If the answer was wrong, you had told it the wrong thing.
That world is gone forever. The machine no longer waits for instructions. It produces. It drafts your memo, writes your code, builds your analysis, and hands it back polished and confident and sometimes completely wrong.
I want to show you the moment the ground moved, because it is not a theory.
It happened in a study and the numbers are exact.
Harvard and Boston Consulting Group ran an experiment with 758 consultants. Real ones. They gave them real tasks and gave some of them an AI to help. On the tasks the AI was good at, the consultants soared—faster, better, more. But on the tasks that fell outside what the AI could actually do, the ones using it scored nineteen points worse than the ones working alone. The machine did not say “this is beyond me.” It answered with the same confidence it brought to everything else.
And the experts, trusting the polish, followed it off the edge.
It gets worse, and this is the part that should hold your attention.
A later study went back and watched what happened when those consultants pushed back—when they fact-checked the machine and told it that it was wrong.
The machine did not concede. It held its ground. It restated, it defended, it found new support for the same wrong answer. It was wrong the whole time.
Read that twice. The tool will look you in the eye and defend a false answer.
Why this splits one skill into two
When I wrote my list, I kept Information and Analysis as separate skills, and over the years people asked me why. Were they not the same thing? Gathering facts, reasoning from facts—why two?
I could not have given you the answer I can give you now. The machine has proven me right by accident.
I had done this once before, and the world taught me to.
The original list did not separate them. Then the internet arrived, and everything changed. When any fool could publish anything, and did, the burden moved onto the reader. Finding facts was suddenly easy. Sorting them for relevance and knowing whether they were true became the hard part, a separate labor all its own.
So, I split the skill in two.
Gathering and judging what is true became one thing. Reasoning from it became another. The internet pried them apart, and I named the seam.
The machine has now pried them further apart.
Gathering information and judging whether it is true is one skill. Reasoning soundly from what is true is another. They were always different. But they used to fail together—if your facts were bad, your reasoning rarely saved you, and a careful thinker usually had careful sources. The two traveled as a pair.
Now you can be handed a flawless chain of reasoning built on “facts” that was more the product of AI hallucination and trying to please, but never true. Your analytical logic might be sound; but your conclusions are poison – maybe true – maybe untrue. The analysis is perfect, presented professionally, but the information underneath it was invented somewhere in the machine’s confident AI world.
So the first question is no longer “what does this mean?” It is “is this even real?”—and that is a separate act of the mind, exercised before reasoning begins.
Information asks: is this relevant and real? Analysis asks: what follows if it is? A generation of new frameworks has merged the two into a single box called “critical thinking.” They merged them at the exact moment the world was prying them apart.
I will keep them separate. The machine insists on it.
The eight, brought current
So here is the list now. Not torn down—evolved.
What survived, survived.
What the world changed, I changed.
The Critical Skills of the future emerged.
- Communications: Getting ideas out of your head and into another’s, and their ideas into yours. It was first on the old list. It is first here. Nothing the machine does has touched it. It’s still the most powerful skill on the planet.
- Information: Gathering what you need, making sure it is relevant, judging whether it is true, and judging what the machine hands you. This is the skill the machine has made dangerous and indispensable at once. In a world that floods you with plausible falsehoods, knowing what is real and exercising judgment are no longer the easy part of the work. They ARE the work.
- Analysis: Reasoning from verified fact to sound conclusion. Findings from facts, conclusions from findings, recommendations from conclusions. All based on solid information and competent analysis. The machine can imitate this and sometimes does it well. It cannot be held responsible for it. You can.
- Technology and AI Fluency. Choosing the right tool, using it with confidence and competence, and judging what it gives back. The old skill was operating the machine and learning how to operate new machines. The new skill is directing the machine—and never letting it forget who is in charge. The fluent worker is not the one who uses AI the most. She is the one who knows where its frontier ends and refuses to follow it past the edge. And the market has already seen this. In a single month in the spring of 2026 the largest pools of capital on earth spent billions—not on smarter machines, but on the scarce people who can put them to work inside a real organization. The model was rarely the real bottleneck. The human who can direct it is.
- Collaboration: Working with people so that they finish better for having worked with you. This was “interpersonal” on the old list. Deming’s numbers say this is the most durable skill of all, the one the market has rewarded harder every decade for forty years. The machine cannot do it. It has no one to be in a room with.
- Adaptability: Staying useful and relevant when the ground moves. This was “continued education” on the old list—keep learning. That was too gentle. Learning is not enough now. You must be able to throw out what you knew and rebuild while the work is still moving under your feet. The skill is not the diploma. It is the willingness to become a beginner again, on purpose, at fifty.
- Execution: This was the old “production” skill. Turning an idea into a thing, on time, at a level you would put your name on. The machine has flooded the world with cheap, fast, mediocre output. The premium has gone, hard, to the person who can be trusted to deliver something finished and true.
And then there is the one that matters most.
- Creative Problem-Solving. Facing a problem with no stored answer and building one anyway. This is the work the machine cannot reach, because currently the machine only knows what already exists. When the problem is genuinely new or unique, there is currently no pattern to retrieve it, and the human who can frame it and solve it is worth more than ever. This might change as AI evolves, but it is where the next thirty years will be won.
What I am not telling you
I am not going to tell you what to do with this.
You are an adult and the conclusion is yours to draw.
But I will tell you what the evidence keeps insisting, against every prediction of human obsolescence. When that same AI was put in front of five thousand customer-service workers, it lifted the weakest among them by a third and barely moved the best. The machine does not replace judgment. It substitutes for the lack of it. Where there was little skill, it adds some. Where there was real skill, it gets in the way.
And across seventy-six million job postings, more than three-quarters still ask for the human skills—communication, collaboration, the durable things. Eight of the ten skills the forecasters say will matter most in 2030 are skills that would have been recognizable to a foreman in 1985.
The machine produces. It cannot judge—not whether its facts are true, not whether its tool fits, not whether its reasoning holds. Every one of those judgments stays with you.
That is the shape of it. The tools changed completely. What they cannot do did not change at all—it only became more valuable, and harder to fake.
I wrote a list once, and I was mostly right.
The part I got wrong the machine has now corrected for me, in public, with numbers.
The work of the next thirty years is not to compete with the machine at what it does well.
The machine will keep getting better.
So must you.
And you must remain the one in charge.
* * *
Charles C. Jett is a Professional Certified Coach (PCC), author, and civic educator. A graduate of the U.S. Naval Academy (Class of 1964) and Harvard Business School, he is the developer of the Field Studies methodology — endorsed by the U.S. Department of Labor and Harvard — and the author of six books on critical skills, leadership, and civic education. He writes at criticalskillsblog.com and civicsage.com, and hosts three podcast series including Making a Great America (all 85 Federalist Papers) and the Jefferson-Adams Letters (nearly 300 letters, 1812–1826).