This one is for me, to laugh at or learn from later. AI will change everything? Or nothing? Or something in between? Here are ten things I’ve thought about often and said rarely, and how they’re changing the way I try to show up.

(This comes on the heels of refactoring this site, having more fun with code, and building GTM tools in minutes and hours instead of weeks and months. I’m bullish but these thoughts keep running in my mind and writing them down helps.)

1. Outside the model

AI cannot feel cold water.

A couple of years ago, before AI went mainstream, Zach Ware published an exceptional and thoughtful post (read it here) that’s been rattling around my head ever since. His argument, roughly: as information processing becomes cheap and everywhere, mastering yourself becomes more valuable. Physical strength is the foundation for mental resilience, and the people not taking the AI revolution seriously are usually the same people not taking their bodies seriously. Zach writes,

“Physical health is crucial in our increasingly virtual world. There’s really no excuse to neglect building strength, staying mobile, and maintaining your vitality when your physical presence is what sets you apart from digital entities.

Your body, with all its systems of movement, feeling, and response, connects you to realities that algorithms simply can’t access. As distractions multiply around us, only the disciplined development of physical robustness—through training, movement, and good nutrition—provides the mental clarity needed to navigate wisely.”

That reframed the whole thing for me. AI cannot feel ice water on a hot day. It cannot experience a runner’s high. It cannot know what it’s like to be tired or hungry or strong. If AI is about to do the things we once considered uniquely human (coding, writing, strategizing), then why not double down on the thing it cannot do at all? Get fit so you can experience more of yourself, more adventure, more of what a screen cannot offer. Lift something heavy. Go for a long walk. Sit with your own body for an hour a day and see what shows up.

The gym is a hedge against the terminal.

2. Welcome charisma

You won’t need people the way you used to. But you’ll want them.

In an AI world, you may not need people for what you used to need them for. But you’ll want them to want you around, and that’s charisma. Connection and rapport get richer when you step away from the keyboard and stand awkwardly in a group of strangers. The person who can welcome others and make a room feel at ease has a large role to play in this era. I’ve been hosting dinner parties with strangers for years now, partly because I like it and partly because it’s a practice. It turns out to have been good training.

Be easy to be near.

3. Read more, not less

Reading, and thinking about what you read, is the edge.

For three decades we’ve been reading less and less even as the volume of words, articles, and books has exploded. More noise, less signal. And yet, in my day-to-day work with AI, I’m reading more. I’m reading the response. I’m reading the work someone else delegated to AI.

When creating becomes a commodity, consumption becomes the differentiator. You don’t need to read the 100-page research paper. You do need to read, analyze, and process the 3-page summary and actually understand it.

If AI output is 100x, your input had better be 10x.

4. There’s just so much to do

There is always more to do.

This one will age well or poorly. While we worry about what AI will do to our jobs, the more useful possibility is that there’s simply a lot of work left to do. In the 1950s researchers predicted a 10-hour workweek because computers would make most labor unnecessary. Instead we work more than ever.

The future looks similar. If coding a page gets automated, you’ll talk to a customer. When that gets automated too, you’ll think of something else worth doing. (Yes, all of us. Maybe you need a three-month sabbatical first. At some point, though, doing is more interesting than not doing.) There’s far more work to be done than we’ve even considered possible.

We don’t run out of work. We run out of patience to notice it.

5. Focus will always matter

Focus is still the whole game.

Henrik Karlsson wrote a good essay on the multi-armed bandit and how focus unlocks nearly everything (henrikkarlsson.xyz). Half contradictory, half complementary to the previous point: the less we do, the more we’ll accomplish, because we’ll concentrate. And if we don’t concentrate, well, that’s not good, but it’s also not new.

Focus doesn’t mean “focus on the task at hand.” It means focus on what matters to you in this one life you have. Maybe we’ll stop worrying about notification counts and start worrying about outdoor hours with our kids. That’s the metric I’m building my life around.

Pull one arm harder. Not three arms lightly.

6. Novelty for the win

The new hire has an advantage right now.

Starting a new job right now is an unusual kind of privilege. When you’ve been in the same role for eight years, as I had at Apptegy, you carry a decade of context. You know who to call about what. You know the why behind every decision and the effort it took to arrive there. It’s almost impossible to write it all down.

When you start fresh, you know very little. That’s precisely the point. AI is most useful to the person still collecting data, organizing information, and building context. Every meeting, every document, every loose end becomes material for a growing context layer. It’s a little like spinning up a new Claude project or a new repo. You and your AI pal grow the institutional knowledge together, day by day. So cute.

Fresh is an advantage now. It didn’t used to be.

7. Redefining our expectations

Stop thinking in 2x. Start thinking in 10x.

AI is already doing this for people who never considered it. Instead of improving 9-10% each month and reaching roughly 100% growth in a year, can AI 10x that? Could I do a year’s worth of work in a single month?

That used to be a superpower available only to a handful of people, most of whom made fortunes with it. Now it’s available to the rest of us. But it demands a different posture. You have to stop thinking in increments and start thinking in transformations.

Most people won’t. That’s the opening.

8. Thinking about our thinking

When ChatGPT was about a year old, everyone asked it the same party trick: “Based on what you know about me, tell me about myself.” The result was a digital horoscope. Charming, mostly wrong, easy to screenshot.

The better use is the rabbit hole underneath it. AI is a remarkably generous teacher on the subject of how you think, if you’re willing to read the response. You get a tool for zooming out and seeing yourself from a bit of distance, which is a rare thing to have on demand. You can ask it how you argue. What you avoid. Which patterns you repeat across different decisions. The answers aren’t truth. They’re mirrors that reflect what’s real but aren’t real in the same way.

Good, objective mirrors are rare. Don’t waste the one you have.

9. Thinking about our learning

Learn how you learn. Then learn faster.

Adam Robinson’s old book What Smart Students Know looks like a cheat sheet and isn’t. Its core move is to pause before studying and ask three questions: What do I already know about this topic? How do I know it’s true? What don’t I know? The idea is deliberate learning instead of rushed studying.

AI is ideal for this. Before you open a new book, ask Claude: “What are the core concepts, and what should I already know going in?” Answer the questions yourself first, without the tech. Then read the book. You’ll be surprised by how much sticks and how quickly you connect it to everything else you’ve read.

Pause before you study. Study less. Remember more.

10. Skin in the game

If building is free, persistence is the new currency.

This is the one I can’t stop thinking about. If building, coding, and writing take minimal effort now, where does resilience come from?

Five years ago, building a web app meant sunk cost. You suffered through ideation, hired a technical co-founder, raised money, dealt with HR issues, absorbed rough client feedback. You didn’t quit, because you couldn’t afford to. If the same web app takes a long weekend and minimal human interaction, who cares if you stick it out? You only lost a weekend.

If distribution and sales become far more important in this new era (and I think they will), the question is who’s going to do the unglamorous grind to figure it out. Maybe AI solves that too. But the human decision to persist is often bound to the blood and sweat already spent.

Which is why #1 matters so much here. Physical fortitude is the foundation for mental resilience. When the sunk cost disappears, the only thing left to carry you through the hard middle is the body and mind you’ve trained to keep going. The gym isn’t just a hedge against the terminal. It’s a hedge against giving up.

What happens when we don’t spend any?

+1. Ten thousand choices

Freedom is choosing. Convenience is the opposite.

AI opens a set of uncomfortable questions about freedom. Ted Chiang’s New Yorker essay from August 2024 makes the case better than I can (newyorker.com). His claim, as I read it: AI deprives us of choice, and freedom is the capacity to choose.

When you are writing fiction, you are — consciously or unconsciously — making a choice about almost every word you type; to oversimplify, we can imagine that a ten-thousand-word short story requires something on the order of ten thousand choices. When you give a generative-A.I. program a prompt, you are making very few choices; if you supply a hundred-word prompt, you have made on the order of a hundred choices.

Pseudo-philosophical AI conversations tend to land at one extreme or the other. Either AI gives us back time and freedom to do what we want (even though there’s plenty left to do, see #4), or we lose all agency to a handful of trillionaires and their models. As with most things, it’s probably some of both, and mostly somewhere in the middle.

Less writing. Less choosing. Less you.

I wrote all of this without AI, by the way, until right here at the end. Then I handed it to Claude to clean up my typos and make it a slightly better read for you.

That’s choice and freedom, right there. :)