The 6 Skills AI Will Never Replace

You've seen the headlines. If these predictions are even half right, artificial intelligence may soon outthink most of us. So in this video, I'm going to give you a survival plan, a way to stay valuable, relevant, and hard to replace. I'll walk you through the 6 human skills, the ones where we still beat the machines, that will matter most in the age of AI. You know, I started working on this topic 20 years ago with this book, A Whole New Mind, and my perspective has been shaped by several other books and 2 decades of research. What I'm going to share isn't a list of technical abilities. It's 6 ways of thinking, of behaving, of being that will set you apart. And I'm also going to give you some simple, practical techniques for building these muscles. So let's go. The first human skill: questioning. Right answers still matter, but smart questions now matter a hell of a lot more. Think about it. We used to have search engines. Now we have answer engines. Large language models and other AI tools produce answers on command. Lots of answers, sometimes dazzling answers. But here's the thing. When answers are everywhere, questions become the scarce resource. When answers get cheap, curiosity becomes priceless. In a world of answer engines, curiosity is your killer app. Because every breakthrough starts with a question. Sometimes a weird, unlikely question. What if light behaves like a particle? What if I can carry 1,000 songs in my pocket? What if people actually want to sleep in someone else's home? Great scientists, great founders, great writers— they aren't vending machines for right answers. They're unstoppable generators of interesting questions. They begin their sentences with phrases like these: Why does? What if? Why not? How about? And they often drop the most powerful question of all: What are we actually trying to solve here? So how do you strengthen the question muscle? If you want a simple starting point, grab The Book of Beautiful Questions by Warren Berger. It's the best practical guide to questioning I've ever found. And here's a dead simple technique for sharpening this ability. The 5 Whys. Toyota used it in the 1950s. You can use it today, and it works absurdly well. Let me give you an example. Say you need a contractor for a big project— redesigning your website, building a studio, renovating your office. So you fire up Claude or ChatGPT and ask for something like the best contractor in your area, and you get a long, confident list of answers. But as you investigate, none of them seem right. Ask why. Maybe it's because none of them is a great fit for your situation. That's your first why. Now ask why 4 more times. Why are none of them a great fit? Because their proposals are all over the map. Why? Because each contractor is making different assumptions about what you want. Why? Because you never clearly defined the scope, timeline, or success criteria. Why? Because you and your team don't actually agree internally on what success looks like. Boom! That's the real answer, the one that comes after the 5th why. AI may be better at delivering answers, but for now at least, you are better at asking questions. And once you've identified the right problem through questions, the next human edge kicks in. Something machines still struggle to fake. Taste. In a world drowning in mediocrity and slop, knowing what's good is a superpower. Taste. Discernment. Judgment. The ability to tap your experience, your intuition, and your values and look at a pile of options and say with confidence, that one. That's it. Remember, AI is really good at generating stuff, but as it pumps out endless drafts, scripts, images, and ideas, taste becomes the filter that separates the marvelously meaningful from the merely meh. Let me give you an example from my own work. I've got an email newsletter. Before AI, I wrote the subject lines myself. Now I hand the draft newsletter to Claude or Gemini and ask for suggestions, and they deliver. 50 subject lines in just a few seconds. It's astonishing. But here's the thing. Most of them stink. And when I say most of them, I mean 47 out of 50 are usually awful. 2 may be solid and 1 might be genuinely good. But to know which is which, I have to apply my 25 years of experience as a writer, my knowledge of who our audience is and what they care about, my own comfort with style and word choice. That's taste. And what usually happens is that I take one of AI's suggestions, tweak it based on that taste, and come up with something even better. That's the secret. Not human or machine. Human plus machine. AI provides the raw material. Taste shapes it into something real. Here's a way to put this idea into action, to get serious about developing and understanding your own taste. Create your own Hall of Fame. In a physical folder, in Notes, Notion, Dropbox, whatever. When you see an example of great writing, great design, great solutions, great innovations, anything that makes your neurons catch fire, capture it, save it, study it. Over time, your Hall of Fame becomes a map of your taste. Patterns emerge, standards rise, judgment sharpens. And that's how you turn taste from something vague and mystical into something concrete And powerful. The future doesn't belong to people with the most ideas. It belongs to the people with the best taste. And that sets the stage for the next human advantage, the one that turns good taste into great solutions. Iteration. Your first version won't be your best version. Your 10th, or maybe your 110th might be. If questioning frames the problem and taste sets the standard, iteration is how you close the gap. Think of James Dyson building more than 5,000 prototypes, or the great abstract expressionist Willem de Kooning working on a single canvas for 2 years, endlessly scraping off paint and starting again until he got it right. Here's the part we don't like admitting: most good things start out bad. The magic isn't in the first spark. The magic is in the relentless revision. AI can help you generate variations at astonishing speed, and that's great, but it still takes a human to refine, redirect, discard, and polish. AI accelerates the quantity. Iteration delivers the quality. Here are a few tactics that have helped me. Simple, unglamorous habits that compound fast. First, adopt Anne Lamott's principle of the shitty first draft. Don't aim for perfection. Just get it done. Then refine, iterate, and refine and iterate some more. Second, space out your iterations. Sometimes I'll write a draft, make a few quick passes, then deliberately leave it alone for a week. When I come back, the flaws pop, the fixes are obvious, and the iteration speeds up. That works for anything— pitch decks, designs, wedding toasts, you name it. Third, adopt what some folks call the version 0.8 rule. Share your work when it's at 80%, not 100%. Now, I'll admit, this is really, really hard for me. I want my stuff to be great. But I found that many times waiting until I'm fully ready blocks progress. Shipping at 0.8 forces me to iterate, to learn, to improve fast. You don't have to get it right the first time. You just have to get it right over time. AI gives you the options. Iteration gives you excellence. And once you have those excellent pieces, you need to know how to assemble them. That's the human skill of composition. 4. Composition. AI is excellent at delivering ingredients. Humans are better at serving meals. Composition is the art of assembling pieces— ideas, scenes, arguments, visuals— into something coherent, meaningful, and emotionally resonant. A composer does it with sound. A filmmaker does it with cuts and pacing. A painter does it with color. Composition is the ability to synthesize rather than analyze, to see relationships between things that might not at first seem related, to combine elements in a way that makes the whole larger and more powerful than the sum of the parts. It's a fundamentally artistic skill, but now even non-artists must master it because you compose every time you create and make a presentation. You compose every time you assemble a team. You compose every time you put on an event or even throw a party. I wrote about an early version of this in A Whole New Mind when I described my experience learning how to draw, seeing the negative space in the FedEx logo, and trying to understand the great symphonies. Here are 3 simple ways to sharpen your compositional skills for the age of AI, when large language models can flood you with components. First, use the rule of 3. Whenever you explain something— a point, an idea, a story— organize it into 3 beats. That will make your thinking clearer and your communication cleaner. A second composition technique: learn to see the structure. Here's one easy trick. Go into your phone camera settings and turn on grid. It's usually a 3x3 layout. This is a super easy way to learn the rule of thirds. And once you understand this rule, the next time you take a picture, instead of putting the subject at the center, try placing the subject on one of the intersecting lines. Congratulations, you're now a composer. And third, play the movie pause game. When you're watching a visually striking movie, hit pause during a scene that doesn't have much action. Where are the actors standing? Are they framed by a doorway? How are they lit? Is there a leading line like a road or a railing pointing at them? AI can make lots and lots of parts, but you don't win with more pieces. You win with better arrangement. If composition is the musical score, the next one is the orchestra. Number 5, allocation. Dan Schipper is the CEO of the media and software company Every, and he recently said something really profound. He said, In a knowledge economy, you're compensated based on what you know. In an allocation economy, you're compensated based on how well you allocate the resources of intelligence. We used to idolize the individual hero, the person who could do everything themselves. But the future belongs to people who can coordinate humans and machines. The new superstars will be the people who can orchestrate and allocate tools, teams, AI systems, timelines, constraints, and bringing them together toward a clear outcome. Now, this isn't entirely new. Of course, a great director doesn't operate the camera. A great coach doesn't play the game. But as AI moves from novelty to collaborator, allocation becomes a core skill for all of us. Allocation is knowing which tools to use, which people to involve, which systems to engage, and at what moment. But it's also deeply human. It requires empathy. Emotional intelligence, and an honest understanding of what people are actually good at. This is what Wharton professor Ethan Mallek calls centaur thinking— combining human and machine intelligence to get results that neither can achieve alone. Want to become a better allocator? Here are some simple ways to begin. Do a team inventory, even if it's just you and AI. List who or what does things best, then delegate accordingly. Stop treating every task like it's yours. A related idea, the 2-pile technique. Every project is really just a collection of tasks to be done. Organize them into 2 piles, one for AI, things that involve speed, quantity, and generation, and one for you, things that involve real thinking, creativity, and taste. Your job isn't to hand everything to AI. That's a huge mistake. Your job is to give AI that first pile so you can excel At the second. Or try the time as talent audit. Look at your calendar for the past 2 weeks. Circle every task that didn't require your taste or creative skills. Ask yourself, could AI do this? Could a tool do this? Could someone else do this? Then reallocate those tasks going forward. We all need to stop managing time and start reallocating talent, including your own. Now, if AI multiplies intelligence, allocation decides where to aim it. But that aim must be true. And that's where the final Human advantage comes in. Number 6, integrity. Technology amplifies your power. Ethics determines how you use it. And in an age of AI, power is scaling faster than character. Every technological revolution forces a moral reckoning. We are in one now, and that makes integrity the most important skill of all. You've seen the news, you've seen the social media posts. AI can hallucinate, fabricate, and confidently spin out of control. It has no conscience, no responsibility, no moral compass. That's where you come in. When intelligence becomes abundant, wisdom becomes even more valuable. And wisdom is rooted in integrity, in making choices based on honesty, fairness, responsibility, and accountability. Integrity isn't abstract. It's practical. It shows up in moments like telling the truth when lying is easier, Pushing back when a large language model says something that compromises your values. Protecting privacy when invading it is more profitable. And asking, who does this decision affect? And would I make it if I were them? When you have more power, more speed, and more leverage, your character becomes your fate. And when AI can scale your impact instantly, integrity isn't just a virtue, it's a leadership skill. Now, practicing integrity is the work of a lifetime, not a single video. We've got teachers, parents, and clergy to help us with that. But here are 2 tips that might help deepen this quality in yourself and others. First, run the Washington Post test. When I worked in politics, including my years as a White House speechwriter, we used something that we called the Washington Post test. Before you took an action or wrote an email, you'd ask yourself, would I be okay if this showed up on the front page of a major newspaper? If not, hit the brakes. Second, run an integrity inversion. Take any questionable decision and flip it. If someone did this to me or to someone I love, would I think it was fair? If the answer is no, don't do it. This simple inversion cuts through rationalization and brings morality to the surface. AI may reshape everything we do, but only integrity and wisdom determine who we become. So those are the 6 human abilities that could matter most in the age of AI. Questioning: asking the sharp, original questions that machines can't. Taste: knowing what's good when everything is possible. Iteration: improving your work version after version. Composition: assembling pieces into something meaningful. Allocation: orchestrating humans and machines toward a clear goal. Integrity: choosing what's right when everything around you is moving fast. These aren't luxuries. They're success skills, maybe survival skills. for the next decade. If there's one thing I know after studying human behavior for 25 years, it's this: when the world gets more artificial, we need to get more human. Hey, what human skill do you think belongs on this list that I didn't include? Add it to the comments. I'm always curious to hear what you see that I miss.