Key Takeaway
The real edge in the AI era is the ability to learn fast and let go fast. Kids raised on external rewards — gold stars, allowance for grades, prize-for-performance loops — grow into adults who can’t put down their tools. Psychology has a name for it: the overjustification effect. A 1973 Stanford drawing study showed that the mere expectation of a reward cut kids’ intrinsic interest by about half. Meanwhile, the World Economic Forum projects that 39% of core job skills will change by 2030. The adults who’ll struggle most are the ones whose identity is welded to a skill they’re afraid to drop. I’m a father in my thirties, and this post is what I’ve been thinking about while watching colleagues at my company resist AI — and what I want my kids to grow up understanding.
Two Camps in My Office
The company I work at is aggressively pro-AI. Executives, middle managers, individual contributors — almost everyone uses AI in some form. And by “uses,” I don’t mean “asks ChatGPT a question once a week.” I mean building automations, restructuring workflows, treating AI as part of the toolchain.
In that environment, the holdouts stand out. They’re the people who reject AI integration, push back on AX (AI transformation) projects, and keep doing the work the old way.
Here’s the thing nobody says out loud: the holdouts are usually the company’s strongest individual contributors. The Excel wizard who automated half of finance. The deck-builder whose slides made the room go quiet. The researcher who could outpace anyone on background work. Before AI, those skills were their moat.
When AI shows up, they hesitate. They argue: “AI can’t do that.” “I’m faster doing it myself.” “The verification overhead cancels the speed gain.” All true, technically. Spending two hours on prompts and reviewing outputs to replace one hour of work you can do in your sleep is not a productivity win.
But underneath the technical objections, there’s a different feeling. If I hand my best skill to AI, anyone can copy it. If anyone can copy it, my position erodes. I’ve watched colleagues quietly look for jobs at companies that haven’t gone all-in on AI yet — somewhere their old moat still holds.
I get it. Up until late last year, I was that guy too. I told myself, “AI is too unreliable for legal work — at best a sidecar tool, never a real transformation lever.” That story let me protect what made me valuable. So I have a lot of empathy for the holdouts. They’re not being lazy. They’re protecting a reward structure their identity is built on.
Why “I Earned It” Quietly Becomes a Trap
We all carry a deep instinct: effort should pay off. I studied hard, so I deserve the grade. I put in late nights, so I deserve the promotion. I spent years getting this certification, so I deserve to keep working in this field.
Nothing wrong with the instinct in moderation. People won’t sustain effort without some promise of reward. The problem starts when the instinct hardens into an unwillingness to let go of any reward structure that’s already paying out.
If you’ve spent a decade sharpening a single skill, that skill is no longer just a skill — it’s part of how you describe yourself. It’s the proof of who you are without needing to explain. Handing it to AI isn’t “switching tools.” It feels like dismantling the structure you used to stand on.
So the holdouts find reasons to refuse. They emphasize AI’s limits, double down on their own accuracy, and run the math on retraining costs. The reasons sound rational — many of them genuinely are. Underneath the rationality, though, is a quieter sentence: “I worked too hard to build this to give it away.”
This is the emotional root of what behavioral economics calls the sunk cost fallacy — the irrational pull to keep investing in something because of what’s already been invested. Sunk cost is the bookkeeping. The reward instinct is what makes it feel personal.
Where the Reward Instinct Comes From — The 1973 Stanford Drawing Study
In 1973, three Stanford psychologists — Mark Lepper, David Greene, and Richard Nisbett — ran an experiment with preschool kids that’s still cited in education and psychology today.
They started with kids who already loved drawing on their own. Then they split them into three groups:
- Expected reward: told in advance they’d get a “Good Player” certificate for drawing
- Unexpected reward: drew without any promise, then surprised with a certificate at the end
- No reward: drew, no certificate, no mention of one
A few days later, during free play, the researchers put paper and crayons out and watched. The result was sharper than anyone expected. The expected-reward group spent roughly half as much free-play time drawing as the other two groups.
The researchers called it the overjustification effect. When an external reward is layered on top of an activity a child already enjoys, the child quietly rewrites the story: “Oh — I’m doing this because of the prize.” The moment the prize disappears, so does the motivation.
“Children showing intrinsic interest in a target activity during base-line observations were exposed to one of three conditions… subjects in the expected-award condition would show less subsequent intrinsic interest in the target activity than subjects in either of the other two conditions.”
— Lepper, M. R., Greene, D., & Nisbett, R. E. (1973). Journal of Personality and Social Psychology, 28(1), 129–137.
External rewards pull behavior out in the short term. But they corrode the intrinsic motivation — the curiosity, the play instinct, the “I just want to” — that was carrying the behavior in the first place. Once that’s been rewired, a person who finds themselves in a no-reward environment loses the ability to start moving on their own.
Self-Determination Theory and the External-Motivation Trap
Lepper’s finding was extended by Edward Deci and Richard Ryan in Self-Determination Theory (SDT). They split human motivation into intrinsic and extrinsic strands and identified three psychological needs that fuel intrinsic motivation:
- Autonomy — the sense that I’m choosing this
- Competence — the sense that I can do this
- Relatedness — the sense that I’m connected to others through it
When all three are met, people can sustain behavior with no external prize attached. When external rewards start crowding out autonomy, the story flips: behavior becomes “I’m doing this for the reward,” not “I’m doing this because I want to.” Autonomy and competence weaken at the same time. The willingness to step into something new — a new tool, a new field, a new way of working — drops with them.
The piece that matters here: the pattern is laid down in childhood and runs the rest of someone’s life. A kid raised on “study hard and you get pocket money,” “get a 100 and we’ll buy you a game,” “don’t skip the academy and we’ll go out to eat” learns to source their behavioral justification from outside themselves. As an adult, that becomes performance reviews, promotions, comp bands. And anything threatening those external rewards triggers a reflex: defend, resist, push back.
The AI Era Rewards Fast Unlearning
Times have shifted.
Learning new things has always mattered. What’s different in the AI era is that letting go of what you’ve already learned matters just as much. A skill that was your personal moat yesterday becomes a feature anyone can use today, after a single AI update.
About four years ago, I built a wildly complex Excel sheet for credit management. It took a week of late nights to put together, drew on years of experience, and was — without exaggeration — the best single deliverable in my career. Nothing comparable existed in my industry or adjacent ones. The structure and formulas only lived in my head. I’d been getting mileage out of it for years.
AI rebuilt that same sheet in thirty minutes. I had to frame the problem and steer the work, sure, but watching years of effort get reproduced with one prompt was — I don’t have a clean word for it. Hollow comes close.
That’s the point. Skills that took years to grind out stop being skills the moment technology catches up. Cling to them and the pace of change leaves you behind.
The World Economic Forum’s Future of Jobs Report 2025 (January 2025) puts numbers on it.
- 39% of core job skills will change by 2030. (Down slightly from the 44% projected in the 2023 report. The drop is attributed to corporate investment in reskilling absorbing some of the change.)
- The fastest-rising skills are AI and big data, networks and cybersecurity, and technological literacy.
- Alongside the technical skills, creative thinking, resilience, flexibility, agility, curiosity, and lifelong learning are projected to rise sharply on the human side.
- Analytical thinking remains the single most-cited core skill, named by 70% of surveyed companies.
The report says one thing in different ways: the people who survive are the ones who can absorb new skills fast and shed old ones just as fast. Anyone whose reward instinct keeps them welded to their existing toolkit will struggle on both ends.
What Happens When a Reward-Driven Child Grows Up
Back to my office.
The colleagues who lean into AI have one thing in common, and it’s simpler than I expected. Their identity isn’t tied to a specific skill. Ask them to describe what they do and the answer sounds like “I solve whatever problem lands in front of me, and I want to be good at solving it.” The mode and the tooling can scramble underneath them — the goal stays put.
The holdouts, by contrast, have an identity that’s almost fused with their signature skill. That skill is what made them visible at the company. It’s also the reward source that defined who they were. Letting it go feels like losing a limb.
What’s probably playing out underneath: a learned reflex to recover whatever reward they’ve already earned, refined over decades of external-reward conditioning. The possibility that something new might pay off doesn’t outweigh the certainty that something they already have could be lost.
This connects to loss aversion, a finding made famous by Daniel Kahneman: people feel a loss roughly twice as strongly as a gain of the same size. Future rewards from picking up a new tool are abstract. Losing the moat you already have is concrete. The stronger the reward instinct, the harder loss aversion bites.
The depressing part is that by the time you notice this pattern in yourself in your thirties or forties, it’s already late. Realizing you’re being threatened at work and trying to rewire your motivation structure on the fly is not impossible, but it’s painful. The foundations of how you handle reward and change get poured in childhood.
How Do We Raise Kids Differently
The kids entering the workforce in the 2040s will see a faster pace of change than we have seen. So how do we raise them so they’re not welded to their first reward source — so they can actually move when the ground does?
1. Praise the process, not the outcome
Carol Dweck’s growth mindset research is famous for one finding: praise like “You got 100 — well done!” trains a child to chase outcomes. Praise like “You tried a new approach this time” or “You stuck with the hard problem” trains them to keep showing up for the next attempt.
Outcome-praised kids tend to identify as “the smart one” and dodge new fields where they might not be smart yet. Process-praised kids identify as “the kid who tries”, and that identity treats every attempt as a self-worth deposit, regardless of result.
2. Be careful about external rewards
Pocket money, snacks, treats, screen time, sticker charts — none of it is bad in isolation. There are moments when external rewards genuinely help. But don’t bolt rewards onto activities your child already enjoys.
Once a reward enters an already-fun activity, the child’s motivation rewires around the reward. Save external rewards for activities the child has low intrinsic interest in, treat them as temporary scaffolding, and fade them out as soon as the behavior takes hold.
3. Model “letting go” yourself — especially in front of them
Hard one to swallow, but important. If we want our kids to drop tools fast, we have to drop tools fast in their line of sight.
Whenever we change something at home — a tool, a routine, the way we handle a chore — explain the change out loud. Why was the old way uncomfortable? How did I notice? What did I try? Did it work? When it didn’t, what did I do? The cycle of “try, drop, try again” is what we’re modeling, not the specific change.
A few months ago, our oldest (age four, six in the Korean count) asked why we had a new long-handled bathroom brush. I told him the truth: I’d been using a small handheld brush, kept scraping my hand on the floor, bought the long one to fix that, then discovered I couldn’t get enough pressure on the floor with it, so I switched back to the small one. He thought for a second and suggested I wear thick gloves next time. Conversations like that — “this is why I changed it; here’s what I learned; here’s what I might try next” — have become routine, and his resistance to change at home has dropped noticeably.
A parent who’s risk-averse with change can tell a kid “you have to learn new things” all day. Kids follow what we do, not what we say.
4. Ask “Why do you want to do that?”
This one is straight from Self-Determination Theory’s autonomy axis. When your child wants to start something, ask why before you say good job. The question gets them to surface their own motivation source.
If the answer is “because mom will be happy” or “because the teacher praised me last time,” external motivation is doing the driving. If the answer is “because it’s fun” or “because I want to see what happens” or just “because I felt like it,” intrinsic motivation is alive.
Asking the question regularly trains them to separate the two. Over time, they get better at noticing when their motivation is coming from outside themselves — and that awareness is what fuels self-starting in environments where no reward is attached.
Closing — We Have to Model It First
If I’m being honest, the audience for this post isn’t really our kids. It’s us.
I’ve watched colleagues resist AX with a mix of empathy and frustration, and somewhere along the way I realized: the AI era will eventually be displaced by something else, and I’ll be standing in their shoes when it happens. Whatever I’m building skill in right now — whatever I’m proud of, whatever feels like my moat — the moment I can’t put it down, I’m the holdout in someone else’s office.
If we want kids who aren’t trapped by reward instinct, we have to drop our own reward instinct first. The kind of adult our kids need to grow into is one who can put down yesterday’s tool today, who can move off a hard-won skill the moment a better one shows up. They learn that by watching us do it casually, not by being told to do it.
The WEF’s list of rising human skills — curiosity and lifelong learning — is the same idea in different clothes. Knowing the answer matters less than being able to change the answer. The capacity to question your own toolkit, and discard it when you should, is what the AI era is asking for from grown-ups.
AI took the problem-solving. Tools took the moat-building. What’s left is the judgment of what to pick up and what to drop. The foundation for that judgment is the motivation structure parents pour in childhood.
So I’m thinking again, today, about the weight of one praise sticker.
Further Reading
- Lepper, M. R., Greene, D., & Nisbett, R. E. (1973). Undermining children’s intrinsic interest with extrinsic reward: A test of the “overjustification” hypothesis. Journal of Personality and Social Psychology, 28(1), 129–137. — APA PsycNet record
- World Economic Forum, Future of Jobs Report 2025 (Jan 2025) — Report homepage
- World Economic Forum, Future of Jobs 2025 — Chapter 3: Skills Outlook — Skills outlook chapter
- Deci, E. L., & Ryan, R. M. — Self-Determination Theory: selfdeterminationtheory.org
- Dweck, C. S. (2006). Mindset: The New Psychology of Success. Random House