This week is a long-form piece on a topic I have been ruminating on for a while. And after completing it, I am proud of the piece. If you enjoy it, and know someone who could use hearing the message, give it a share. Thanks as always for reading -

There is a temptation with every transformative technology to ask, "Will this replace us?". Workers had these same thoughts during the industrial revolution. People thought the internet was going to be the death of civilization. Obviously both turned out not to be true.

And it is the same today with AI - I still think we're asking the wrong question.

AI is extraordinary. I use it every day. It has helped me become a better writer, a faster researcher, a stronger leader, and a more effective learner. In many ways, it feels like having an endlessly patient expert sitting beside you, ready to help solve almost any problem.

It isn't going away, and it shouldn't.

But over the last few months, especially as I have used it more, I've started wondering about something else.

What happens long-term when we stop thinking because something else can do it for us?

Not overnight, but one prompt at a time. I wanted to write about that this week. To offer an alternative, non-doomsday outlook on AI. And to provide a prescription and some thoughts to how to live the best of both worlds. Hope you enjoy -

A recent podcast episode with Ezra Klein (ironically about the rapidly declining state of our politics) branched off into an interesting discussion on AI, and it is ultimately what got my wheels spinning for this piece, and is mentioned frequently below. You can listen here.

When Google arrived, it changed how we found information. When ChatGPT arrived, it changed something much deeper.

Google still required effort. Which seems silly looking back on because in the moment it felt so…easy. Kind of like the difference between broadband internet after years of AOL dial up tones as a kid.

You had to decide what to search for. Open multiple tabs. Compare sources. Ignore weak arguments. Connect ideas across different perspectives. Live with uncertainty while you figured things out.

There was friction. And that friction was doing something important. It wasn't just helping us find answers, it was building the mental machinery that allowed us to think.

AI removes almost all of that friction. Ask a question, receive a polished answer. No searching. No synthesizing. No wrestling with competing viewpoints.

It is, as Ezra Klein said on the episode - "the most powerful uncertainty killer ever built."

That's incredible. It's also worth paying attention to. Because uncertainty isn't always the enemy. As a matter of fact, sometimes it's the gym where better thinking gets built.

One of my favorite ideas from that conversation had nothing to do with AI.

It was about books. You know how big of a fan of books we are around here.

Klein described books not simply as containers of information, but as a technology of thinking - a “scaffold that helps us develop our own ideas rather than borrowing someone else's.”

He said:

That sentence will stop anyone who reads often or occasionally.

For years I've encouraged readers of The Superset to spend more time with physical books. I usually talk about knowledge or learning.

But maybe those aren't even the biggest benefits. Maybe reading is one of the last places where we still practice sustained thinking.

When you read a physical book, questions naturally arise. You don't immediately reach for your phone. You don't ask ChatGPT. You sit with them. You wonder.

You connect today's chapter with something you read six months ago. You disagree with the author. You replay an idea during a workout. You notice a contradiction.

Your mind keeps working. That process isn't a bug, it’s the foundational point of picking up the book in the first place.

Klein argues that one of the greatest benefits of reading on paper is cultivating "the form of sustained attention without reaching to resolve every question that occurs in your mind."

In other words, books don't just teach us. They train us.

This concern isn't unique to AI.

Back in 2010, Nicolas Carr asked a question that seemed radical at the time in The Shallows: “What if the internet isn't just changing what we know, but changing how we think?”

His argument wasn't that technology makes us less intelligent. More that it changes the habits of attention our brains repeatedly practice.

We become incredibly good at scanning, and with that less practiced at contemplating.

Cal Newport has spent years making a similar point.

The ability to focus deeply isn't merely a productivity hack. It is becoming a competitive advantage precisely because it is becoming rare.

The people who can sustain attention while everyone else fragments theirs will create better ideas, make better decisions, and produce more original work.

AI doesn't eliminate the need for deep thinking. Ironically, it may increase its value.

And this is where we find the paradox. The same paradox that will benefit those who understand it over the rapid evolution of these technologies in the years to come.

The smartest people I know often feel almost superhuman using AI. They're faster. More informed. More productive.

Klein offers a fascinating counterpoint.

He worries that what feels like becoming superhuman may actually be making us "less human."

Not because AI is bad. But because we slowly stop exercising the very abilities that make us uniquely valuable.

These aren't skills you will find on a spreadsheet. They are deeply human ones. And they develop through experience, not instant answers.

As Klein put it:

So maybe our goal shouldn't be to compete with AI at what it does best.

Maybe our goal should be, in his words, "trying to be as little like a machine as possible."

That extends beyond books. Klein talks about "the void."

Moments where nothing is happening externally, but something important is happening internally.

We spend so much of our lives trying to eliminate boredom, yet boredom has always been one of the gateways to insight.

Some of my best newsletter ideas haven't appeared while staring at a screen. They've shown up halfway through a run. Or walking Tatum around the block. Or sitting quietly in the wee hours of the morning after finishing a chapter of a book.

The world calls those empty moments. They're anything but.

My overarching message is that none of this means we should reject AI.

Far from it. Instead:

Just don't let it replace the thinking that only you can do.

Technology has always promised convenience. The challenge is making sure convenience doesn't quietly become dependence.

As Klein warns, if we aren't intentional, we risk allowing these technologies to "colonize our minds" by outsourcing our attention, our curiosity, and eventually our ability to think independently.

That future isn't inevitable. It's a choice.

As AI becomes more capable, I'm trying to practice a few habits that keep my own thinking sharp.

Read physical books. Not because they're nostalgic, but because they remain one of the best technologies ever invented for building sustained attention. Read slowly enough that your own thoughts have room to interrupt the author.

Let questions breathe. Resist solving every curiosity in thirty seconds. Live with uncertainty for a little while. Your brain often produces better answers than your first search result.

Protect the void. Take walks without headphones. Drive in silence. Leave your phone behind occasionally. Your best ideas rarely arrive when your attention is completely occupied.

Use AI as a sparring partner, not a substitute. Write your first draft before asking for feedback. Form an opinion before requesting a summary. Solve the problem yourself before asking for the answer. Let AI sharpen your thinking, not replace it.

Cultivate things machines struggle to imitate. Taste. Empathy. Wisdom. Storytelling. Relationships. Judgment. These aren't becoming less valuable in the age of AI.

They're becoming the differentiators.

I don't think the future belongs to the people who use AI the most. I think it belongs to the people who know when not to.

AI will almost certainly become smarter than any of us. That doesn't make deep thinking obsolete. It makes it one of the last truly human advantages worth defending.