The Most Expensive Instinct

Pew released their big "Americans and AI" study last month, and the age breakdown contains a finding I can't stop chewing on. It's not that older adults are afraid of AI — it's stranger than that. The most worried people in America are using AI the most. The least engaged people aren't particularly worried at all. They've simply left the field.

Last week, I wrote about the most expensive sentence in business — "AI is just not there yet" — and what it costs the organizations that say it. Today I want to talk about the personal version of that bill. Call it the most expensive instinct. It's quieter than the sentence because it doesn't announce itself. It doesn't even feel like a decision.

Most of the time, it doesn't feel like anything at all.

The Worried Are Leaning In. The Unworried Have Checked Out.

Pew surveyed over 5,000 American adults in February. The age breakdown is stark. About two-thirds of adults under 50 use AI chatbots; among 50-to-64-year-olds it's 42%, and among those 65 and older, 23%. Daily use falls off the same cliff — a third of adults under 50, versus 19% and 7% for the older groups. Confidence? Just 6% of adults 65 and older feel very confident with these tools. 77% don't use them at all.

But here's the finding that rattled me: adults under 30 are actually the most pessimistic about AI (nearly half think it will be negative for society) and they're also the heaviest users. Older adults are less alarmed, more unsure, and mostly... absent.

Read that again. The worried are leaning in. The unworried are checking out.

Which means this isn't a fear problem. It's a checkout problem. Nobody slammed a door. They just quietly declined the invitation.

Against the Natural Order of Things

Douglas Adams saw this coming decades ago. He proposed a set of rules describing how humans react to technology. Rule One: everything that exists when you're born feels like a normal part of how the world works. Rule Two: everything invented between fifteen and thirty-five feels new and exciting. You can probably even build a career on it! And then the kicker, Rule Three: "Anything invented after you're thirty-five is against the natural order of things."

That third rule is doing a lot of work in the Pew data. Notice what it explains: opting out never feels like opting out. It doesn't feel like anything. The new technology simply reads as not-for-you. A toy, a fad, someone else's department. There's no moment of decision, which is exactly why nobody audits it.

(I'm not throwing stones, by the way. I'm north of thirty-five myself. Rule three is coming for all of us. The only question is whether we notice when it arrives.)

The Zero Problem

I spent last week filming a LinkedIn Learning course, and one segment kept circling back to this exact data. Here's the idea I couldn't stop hammering: if AI is a multiplier, then we've got to talk about what I call the zero problem.

Not to mansplain math, but zero times any number is zero. Zero times ten is zero. Zero times a thousand is zero. In the age of AI, if your baseline skill in a domain is weak — or zero — AI simply multiplies that weakness and hands you back zero.

Think about it like a weight room. If you can bench press 100 pounds of creative force on your own, AI is like a 10X exoskeleton: now you're benching 1,000. But if I could train you to bench 250 unwired? That same exoskeleton takes you to 2,500. The gap between those with expertise and those without isn't shrinking in the age of AI. It's widening — because multiplication amplifies differences in the baseline.

I made the basic version of this case two years ago in Mind the (Generational) Gap: your experience isn't a liability with AI; it's the raw material. I still believe that. But I'd sharpen it now. Experience doesn't automatically create an AI advantage. It creates the possibility of one. Multiplication requires contact. Judgment that never touches the model multiplies nothing.

Which is what makes the Pew data so tragic. The people with the most baseline expertise to multiply (the seasoned executives, the thirty-year operators, the ones holding all the 250-pound baselines) are the very people opting out. They're not declining a tool. They're converting a lifetime of judgment into an unmultiplied asset.

The most special thing about AI is you. Not the model. Not the interface. Not whatever release made everyone briefly lose their minds this week. You — your context, your taste, your sense that "this isn't quite right" before you can explain why. AI multiplies whatever you bring. The experienced bring the most.

In Defense of Sitting One Out

Let me steel-man the skeptics, because they've earned it. If you've been working for thirty years, you've watched a parade of "revolutionary" technologies come and go: the Segway, Second Life, crypto, NFTs, the metaverse. Sitting those out cost you nothing. In several cases, the skeptics were flat-out vindicated. Pattern-matching this moment to those moments isn't stupidity — it's induction.

But watch how the pattern-match curdles into a loop. You try ChatGPT once. You give it a lazy prompt, because you don't expect much. You get a mediocre answer, because lazy prompts produce mediocre answers. You feel quietly vindicated. Low expectations produce lazy collaboration; lazy collaboration produces mediocre output; mediocre output confirms low expectations. And then you call the loop "discernment."

There's a massive difference between concern after practice and concern instead of practice. One is wisdom. The other is avoidance wearing a sport coat. Some concerns about AI are entirely legitimate: accuracy, dependence, what happens when people outsource thinking they should be developing. But the people whose skepticism I take seriously earned it through contact with the thing itself. That's different from standing outside the gym announcing that weightlifting seems overrated.

And this pattern-match fails on two counts anyway. Those past technologies asked you to bet money. This one asks you to bet twenty minutes. And those technologies didn't compound against you while you waited. AI proficiency does. As I wrote in It's a Skill, Not a Pill, access isn't the inflection point; reps are. Every week on the sidelines isn't a week of safety. It's a week of widening gap.

The Opt-Out Is Reversible

If this sounds permanent, it isn't. I once watched it reverse in a single afternoon: a senior board member of a $10 billion company, sixty minutes into an AI immersion, confessing "This is overwhelming for me... you use tools as I use my tool set for mechanics." Four hours later, the same man was declaring "No information should get to me that wasn't reviewed by an AI," and asking for help implementing it that day. Nothing about his age changed in between. Only his exposure did.

Don't Research. Rep.

So why is the first step so hard? Honestly, I think it's because experienced folks have forgotten how to learn something new. The learning sciences are abundantly clear: it's not the consumption of information that builds skill — it's the application of information. Learning scientists call this the difference between passive learning (which isn't really learning at all) and active learning (which builds skill).

Which means the answer is not to read twelve more articles about AI. (I tell podcast listeners the same thing: better to stop an episode halfway and do the thing that resonated than to finish it with a warm sense of accomplishment and zero action.)

The answer is one rep. Pick something you actually know something about — a client situation, a hiring decision, a negotiation, a sermon, a syllabus. Then open a frontier model and paste this:

"I want you to be my thinking partner. Interview me about the following situation, one question at a time, drawing out my context, assumptions, and hard-won instincts one question at a time before you give any advice. Then summarize what you heard, identify what I might be missing, and offer a few obvious and non-obvious possible paths forward. Please start by understanding my situation and recommending the “subject matter expertise” I should be counting on you to provide."

Answer like a human being. Ramble. Add the weird context. Say the thing that "probably doesn't matter" (it often matters). And when it responds, don't grade it like a vending machine — manage it. Tell it what it missed. Ask it to try again from the perspective of someone who's been in your industry thirty years.

That's the rep. Twenty minutes. That's the whole bet — and it's the first time you'll feel what it's like when a technology finally has room for everything you know.

Rule Three Is Not a Law

Adams was joking, of course. Things invented later in our lives aren’t “against the natural order of things.” Far from it: the natural order is progress! But the joke lands because the feeling is real: past a certain age, new technology genuinely registers as a violation of the natural order. The feeling isn't the problem. Treating the feeling as unquestionable wisdom is.

Rule three isn't a law of nature. It's a default setting. And a default can be overridden by anyone willing to spend twenty minutes finding out what's on the other side of the instinct.

The most experienced people in the workforce are one honest rep away from discovering that this technology was built for them. Don't let the most expensive instinct win.

Don't opt out.

Related: Mind the (Generational) Gap
Related: The Most Expensive Sentence in Business
Related: It's a Skill, Not a Pill

Join over 52,147 creators & leaders who read Methods of the Masters each week.