La paradoja de la IA: 5 habilidades humanas más valiosas que nunca

[0:00] You've been sold this picture on where AI is going and the tide coming in and every few months the water takes another one of your skills off the beach and things are sort of doomfilled. I think that picture is wrong and I have receipts. It is tempting to say that the only people that can make AI product that matters are people who are geniuses and people who are on super teams and that is the usual story we hear about Alexander Wang, Natt Freriedman and the team at Muse. But I'm here to tell you that the skills that they use to build
[0:30] muse are absolutely skills that one they talk about publicly and two skills that I have talked about in this video channel for the last year and they are learnable and practicable by you and I know that because I've had conversations with hundreds of you and I've seen the skills in practice and so we are going to talk about five skills that have increased in leverage in the age of AI in the last year even as AI has gotten so much more capable. That's the opposite of the story that you're hearing. And if you plan next year off
[1:00] of the I'm losing skills story, you're betting on the wrong horse. So, let's fix that. If you're making plans for 2027, I would take another look at the projects you've already ruled out. The service you couldn't launch without hiring. The problem your team understood but never had time to fix. The idea that needed a technical person you couldn't afford because those decisions were based on what the work required then. And a lot of those requirements have changed with AI, right? That's what we talk about on this channel. My bet for
[1:30] 2027 is that more of us will be able to carry a whole piece of work that used to require a bunch of people's help. And that means that we need to develop the skills that allow us to exercise that degree of practical power. I'm going to give you five takeaways, one for each skill. Taste, agency, focus, resourcefulness, and follow-through. For each, we'll connect what this year has taught us to something you can work on for next year. And by the end, you're going to have a way to decide which of
[2:01] the projects that have been sitting in your backlog deserve a second look and what it would take to make something really useful. So, takeaway one, taste. It gets defined away a lot, but it matters a ton. Taste is acting on something you care about. And by the way, not many people define it that way, right? A lot of people say it's the ability to recognize what good work looks like, what a situation calls for. And I think that's true, but there's more to it than that. When I wrote about taste in September of last year, one of
[2:32] the things I emphasized was that people early in their careers can have exceptional taste in a particular area. They've been paying close attention because they're passionate. They've tried things. They've compared them and they've developed a sense of what good looks like. Along the way, since then, I've met a lot of those people. Years of broad experience don't automatically mean you have good taste. Last November, I also wrote about judgment, which I'm including with taste here. That included defending what you depp prioritize,
[3:02] deciding which actions ought to come first. This year, the opportunity to act on that understanding has grown. Think about a part of your work where you keep noticing the same problem again and again and again. Maybe new customers are getting confused at the same spot on the site or the explanation makes sense to the team that wrote it, but you've had enough conversations with customers and you know it's not going to make sense to them. You can almost predict the question they're going to ask. That kind of attention is where taste starts. You
[3:36] care enough to notice distinctions other people they're just going to skip over. Take the customer example. You might previously have written a memo explaining the problem and you waited for another team to do something about it and you're like, "Ah, throw it over the wall, right? Waggle will fall." Now, you might be able to use AI to build an interactive guide or a rough version of a better onboarding experience. You can do something about it. You have something a colleague or customer can try. You discover details that weren't obvious from the complaint. Perhaps the
[4:07] customer understands the explanation but doesn't know which choice applies to them. Perhaps the sequence is wrong. Getting something in front of them gives your understanding a chance to become more precise. It gives you a chance to effectively take action against the taste and intent you have and see what comes out. And the AI can be helpful on that process, right? It might suggest an approach you hadn't considered. It might explain why a design you admire works. your own taste might improve by
[4:37] collaboration with the AI. I think that that bit gets underplayed when we talk about taste. You're able to learn from the work that you're making faster and from the intelligence that helps you make it. So for 2027, I expect more of the people closest to a problem to be able to just say, you know what, screw it. I'm going to produce a working solution. the support person, the operations person, the subject expert who has been saying for years that something could be better. You guys are
[5:07] going to have more ways to demonstrate what you mean. You can just do it. And that changes what I'd encourage in a team. When someone cares deeply about a frustrating part of the work, give them a chance to show you a better version. You know, their first attempt might reveal an opportunity that a meeting about the problem is never going to uncover because they're so close to the problem space. If that's you, and for most of us, we have something we're that close to, pick something specific. Take the explanation that has always confused
[5:38] people. Take the report that never answers the question. Whatever it is that's driving you nuts, try making an alternative. You'll learn how much of your opinion is actually going to get value and you're going to get a feedback loop going and you're going to sharpen your taste. That's how I would develop taste in 2027. I would lean into that passion and I would put your standards to work and I would start a feedback loop you can go after. Takeaway two is agency. Revisit what you believe you can take on. Agency is the ability to take
[6:10] initiative. It's the ability to act on what you could influence. This year in January when I wrote about high agency, I emphasized the distance between saying you're going to do something and actually doing it. I called it the say do ratio. and you can spend an enormous amount of time feeling interested in a project, talking about it, gathering resources, getting ready to begin and you haven't done a AI gives you more help with the bits that used to stop you. It also gives you an almost unlimited supply of plans. Boy, they
[6:40] like to plan. So, the willingness to make a real attempt, that's on you. That's on you. The important shift isn't what limits you. That's what's changed this year. If you understand a problem, but solving it requires six months of someone else's time, it might not anymore. If it required access to that person, it might not anymore. Once AI helps you do enough of the work yourself, then the leverage is on you. Your willingness to act has a much more direct effect on whether problems get
[7:11] solved today than it did last year. That's what I mean by AI extending the reach and leverage of these skills. the same initiatives has so much more capacity behind it. And so it's up to you to actually translate your intent into impact. This year also changed how I think about being ready. In August, I used an example of a sales leader who wanted a better way to understand which deals needed attention. The leader could not describe the whole system in advance. Most sales leaders can't. An agent can make a rough first version. It
[7:42] can flag deals with no activity for 14 days. can do an approximation and looking at it you start to realize no that's not the signal I want the real concern is this contract that legal has had for 10 days with no next step and that's what I'm frustrated about in other words the first version elicits information from you that you didn't know you had but it's in there that you didn't know how to explain so seeing an imperfect attempt can actually help you to take action now I've been putting a
[8:13] lot of weight on the difficulty of saying exactly what you want. But one of the things that I think this year has taught me is that you can develop that clarity as you iterate through the work with AI. Building helps you find the plan. So take action. That matters if you've been waiting until you understand every step. Right? Don't do that. The first useful attempt is right in front of you. So go after it. So my call to you for 2027 is that you should be willing to increase the scope of what
[8:44] you're going to try. You should be willing to realize that what is possible for you is probably going to grow every single month in the coming year. And you have the chance to recognize that and say, I'm raising my sights. I'm raising the ambition that I have. I'm going to go after something bigger. I have the ability to directly impact the business in a way I didn't before. Whether you're an individual or whether you're leading a team, agency is going to scale for you. You can revisit the projects you've been sitting on and you can say, you
[9:15] know what, what's blocking me? Why can't I start? Why can't I start? And develop the plan along the way. There's an opportunity here to practice agency with real consequence that's growing every single month. So, make the attempt, learn what it t it sort of reveals for you, and let that experience change your next decision. Takeaway number three is focus. Concentrate your effort as more options become possible. And I realize this is a little bit countercultural. I have seen prominent folks in AI say we
[9:45] can try many many many many things because AI gives us the ability to do that. And it's certainly true that you can try many many things. But that doesn't mean that you maximize your leverage by trying many things. There is a lot of value to focus in a world that is telling you that you can make anything possible. Focus and determination enable you to put more leverage, more intent, more of your passion to get something done behind any one thing. Back in April of last year, I
[10:15] was writing to companies about strategy to I I was telling them you got to understand the problem. You got to choose where to concentrate. You got to take actions that reinforce one another. That part all makes sense, right? But one of the things that has really popped out to me as I've been thinking about what strategy looks like in the age of AI is that you have to have an intensity of focus to go with whatever strategic plan you have. You have to go after something and kind of load up the bases in baseball terms and be determined to
[10:45] just get that particular thing done. And I think that's been a little bit underweighted in the conversations that I've been in about AI. How much sustained attention can you put on a problem? How long can you stay interested in something that is difficult and boring? And how passionate are you about seeing it all the way through? You are going to have the world's best tools. The tools that only OpenAI has and only Anthropic has now are going to be in our pockets in a year. And so this year, the advice I
[11:15] would have to you is to become more demanding as fewer ideas are ruled out by the cost of attempting them. Have higher standards. Being anti-slop and being passionate about pushing stuff through and being passionate about getting exactly the right thing done and making your dent in the universe because you're obsessively focused and you're just staring it down. That's rare. That's rare. Consider two people with the same AI tools. You kind of don't have to cuz we all have the same tools now, right? Everyone has this. One of them is going to spend a month beginning
[11:46] 15 projects and moving on whenever the work gets awkward. It's the infamous like vibe code project start problem. The other one is going to choose a customer problem and just laser focus on it and go after it. The second person has given their attention somewhere to accumulate to build. they understand the problem better and other people will have started to experience the benefit of what they're working on and be giving them feedback. You can have a lot of wide interests and you can still make
[12:16] the choice to focus. Exploring lots of directions, yeah, that can be good for discovering a problem, but once you decide to pick something, I'm telling you, the benefits to focus are disproportionate in 2027. It's going to matter. AI can also make distraction look extremely productive. I can tell you that by five o'clock you might have lots of prototypes. You might have three plans. You have an very elaborate testing version of an idea that you weren't pursuing when you woke up. You did things right. But how much of that work got far enough to make a difference
[12:47] to anybody? So for 2027, I expect that deciding where to concentrate is becoming a bigger and bigger part of leadership. A team may bring you more credible proposals than ever next year. And whether you are an individual or whether you're a people influencer or a manager, picking where you focus is going to matter more and more. If you keep adding projects because each one looks easier and easier and easier, it's easy to consume all the extra capacity
[13:17] that AI gives you and still leave the important work unfinished. And so when you're considering a new opportunity, ask what deserves your attention? What would you keep working on after it stops getting exciting? What are you willing to leave alone so you can focus on what's important? Those are decisions about how you're going to spend your life, about how you're going to spend your team's time. AI gives them much, much more consequence because you can do so much more with all of that practical power. Takeaway four, resourcefulness.
[13:50] Get comfortable going through a full project. Resourcefulness is the ability to find a way forward when you don't know how to proceed, right? It's being scrappy. It's getting in there and getting your elbows dirty. Last year when I wrote about taste, I talked about people being flexible tool users. I think that's a great description of us as a species. This year, the people building on OpenBrain, the knowledge system I shared, gave me a concrete example of what flexibility looks like. I recommended that people try Slack and Notion and Zapier and an AI model all
[14:20] together. And people started sending in different versions and hooks and one used Discord and one worked it into Obsidian and one worked it into Mac Whisper and another worked within software their organization allowed. Regardless of what they did, they understood what the intent of the system was for. Capture information and make it useful again when you need it. Right? That understanding led them to change the components out in ways that worked for them. And so when they got stuck, they asked other people who had encountered similar problems in the community and worked with AI to
[14:50] implement the next step. All I did was give folks a starting point. They found ways to make the idea fit their circumstances. That experience is a good example of why I would bet on resourcefulness for 2027. The tools, they will keep changing. I can't tell you what model will be like in 2027 except it'll be better, right? The ability to get yourself through unfamiliar territory is going to remain useful across all of those changes. I expect more people to become willing to take on projects with unfamiliar parts
[15:22] because they can learn close to the moment they need the knowledge. We all can. The next project becomes possible partly because of the learning of finishing the last one. And AI really changes the experience of what it's like to be a beginner. Bring the confusing error in the project you're working on. Ask the AI to explain what's happening. Ask whether there's a simpler approach. The explanation has somewhere to go because you needed to accomplish something you care about. Right? It goes back to focus and intent. If you care
[15:53] about doing this, you will be scrappy and you will use AI to go after it. For a leader, that means the skills listed on someone's resume may not tell you about how helpful they're going to be in 2027. You want to be paying attention to how people approach unfamiliar problems. Do they find a useful question? Do they reframe? Do they come back with another attempt they made on their own? How do they move the work forward when it's unclear? And you can practice that yourself. Choose something that's worthwhile that's hard for you that you
[16:24] don't understand yet. Learn enough to chew on it, to make progress on it. Each time you do that, you start to get evidence of what works and what doesn't, and you can start to expand your ambitions. Takeaway five is follow-th through. Carry the work through until someone benefits. No more prototypes. Follow through is finishing what you committed to and making sure it helps the person it was meant for. In September of last year, I wrote about AI execution in terms of improving the business. That principle still holds.
[16:54] This year what has changed is how much of a useful result one person is able to carry. I talked about expanding the scope of our ambition. Right? We are getting very very good at getting things started with AI. The first version can appear very quickly. It can look really good. It can give you a little taste of the finished result. We sometimes call that vibe coding. And then you have to stay involved long enough to get it to matter to somebody. Let's go back to that customer support example. You cared about the confusion. You made a better
[17:26] guide. Now a new customer actually has to get through the bit that was confusing them before. Perhaps the guide needs to show up sooner. Perhaps a colleague needs to change the handoff. Perhaps the explanation still makes an assumption that the customer doesn't it doesn't quite register right. They don't quite get it. I've been in those rooms. Your followth through in that circumstance getting back after the problem and saying, "Yeah, I released something, but now does the customer really get it? Did it really help? The follow through is what connects the thing you made to something that the
[17:57] customer actually cares about. Our sales leader example is the same way. The sales leader has to figure out how to actually change the way pipeline is measured in a way that's useful for them. They can't just get done by brainstorming with AI. That's not enough. Think about what that principle means when you make commitments to people. You may previously have been able to offer an analysis or a recommendation. Now, you may be able to go farther on follow-through. You can deliver something that really works, that really helps people get something done they
[18:27] couldn't before, and you can stay involved while people discover what still needs changing and give you feedback and help you get better. Your ability to impact how people solve problems, your ability to deliver value just got bigger, which means the value of follow-through went up. There is effort involved in understanding what the person receiving your work actually needs in staying interested after you've demonstrated the idea is possible. The effort to keep paying attention and care often determines whether people are glad
[18:58] you took the project on or whether it's just another boneyard project that dies. My expectation for 2027 is that impressive first versions are going to be increasingly blasze and ordinary. People will have more of them than they know what to do with. the person who repeatedly carries that start through, they're going to earn a different kind of trust. If you lead a team, you know what that trust feels like. Someone will take on a problem and actually finish it and solve it. That's going to be priceless. You learn that commitment
[19:31] from watching others do it and from staying attentive and passionate about the problems you care about and choosing to do it even when it's boring. Before you start your next project, ask yourself, who is this for? What will that person better be able to do? What would I need to do to make sure that I did the whole job? I don't know what that looks like for you. I've seen lots of examples in coding. I've seen them in product management. I've seen them in support. They generally look like going past the edges of your job boundaries. I have talked with engineers who are like,
[20:02] "Well, if it ships to production, I'm done." But better engineers know that's not the end of the job. They know that they have to make a piece of software that's actually useful to people. And that kind of intent has so much more value now. It means that you actually care enough in a world when so many people are using AI to substitute for care. It's like the opposite of the AI written emails that nobody pays attention to. It's the choice to care about how your work impacts people.
[20:32] That's what has maximum leverage in 2027. That's what you should be paying attention to. That's what I'm calling followthrough. And really, these five skills connect like taste, agency, focus, resourcefulness, follow-through. If you feel connection, it's there, right? Together, they help you recognize worthwhile improvements, go after them, give them sustained attention, learn what you need along the way, and carry that result all the way through to the end. That is the skill group that I would develop for 2027. I expect more
[21:03] and more and more of the work between an idea and a useful outcome to become accessible to all of us to anyone who understands the problem and is willing to stay with it regardless of your job family. I can't tell you exactly which project is going to become doable for you right now, but I can tell you that that group of five skills is going to stick around. I have very high confidence in that. AI has expanded what we can do faster than many of us have changed what we're willing to attempt.
[21:34] And so maybe take a minute, look at your plans, your career dreams for the next year. Think about a project you've had in mind for a while and then think about that skill group. Think about the five skills that I talked about in this video and give yourself a chance to discover whether you can do more than you thought you could and whether you can actually follow all the way through and deliver impact. Let me know what you're thinking about building. Uh I'll be curious to see in the comments.