This morning I was reading about Lovable's annualized revenue crossing $600 million. A company that helps people build software by describing it in plain English. And my first thought wasn't "wow, that's impressive." It was "here we go again." Every few months there's a new number that's supposed to make us panic about automation. A new tool that's going to replace us all. A new headline asking whether AI will take our jobs. And every time, the conversation goes the same way. Someone says AI is coming for everything. Someone else says no, it's just a tool. And nothing gets resolved because both sides are arguing about the wrong thing. Here's what I actually think, and I've been thinking about this for thirty years, long before it was fashionable. The real question isn't whether AI will replace workers. It's whether workers will learn to use AI to do things they couldn't do before. That distinction matters more than any prediction about job losses. That's the definition I want you to hold onto. Automation doesn't eliminate work. It changes what work is worth doing. The people who understand this will thrive. The people who don't will spend a lot of time being angry at the wrong target. Let me tell you a story. In 1995, I started a company called Viaweb with Robert Morris. We were building software that let people create online stores. At the time, if you wanted to sell something on the internet, you had to hire a programmer. You had to know HTML. You had to understand servers. It was a nightmare. We made it so anyone could do it. And you know what happened? A lot of web developers lost their jobs. Not all of them. But the ones who made a living building simple storefronts for small businesses? They were toast. But here's the thing. The total number of people working on the web exploded. Because suddenly every small business could have an online store, not just the ones who could afford a developer. The work didn't disappear. It moved. It changed. It became something different. I think about that a lot when I read these panicked articles about AI. The jobs that get automated are the jobs that were already pretty mechanical. The ones where you're basically following a script. And the jobs that grow are the ones where someone needs to decide what script to write in the first place. Now, I'm not naive. I know this is hard for the people in the middle of it. When your specific job gets automated, "the work moved" doesn't pay your rent. I get that. I've seen it up close. But the solution isn't to fight the automation. It's to understand what it's actually doing. Let me give you a concrete example. A few years ago, if you wanted to build a web application, you needed to know a programming language. You needed to understand databases. You needed to know how to deploy things to servers. That was the job. Now, with tools like the ones Lovable is building, you can describe what you want and get something working. Does that mean programmers are obsolete? No. It means the barrier to entry is lower. It means more people can build things. And it means the programmers who remain need to be better at the things that are still hard. What's still hard? Taste. Deciding what to build. Knowing when something is good enough. Understanding the problem well enough to know what the solution should look like. I've written about this before. I call it the difference between writing and thinking. When you write something down, you're not just recording what you already knew. You're discovering what you actually think. The act of putting words on paper forces you to be precise. It exposes the gaps in your reasoning. AI can write. It can generate text that sounds like thinking. But it can't do the discovering part. It can't be surprised by what it finds. And that's the part that matters. So when people ask me if AI will take their jobs, I want to reframe the question. The real question is: what part of your job is thinking, and what part is just executing? If your job is mostly execution—following a process, filling in templates, doing the same thing over and over—then yes, AI is coming for you. And it's probably already there. But if your job involves deciding what to do, figuring out what matters, knowing when something is good enough—that's harder to automate. Not impossible. But harder. Here's the thing that nobody wants to hear: the answer isn't to become a programmer. That's just another form of execution. The answer is to become the person who decides what the programmer should build. I've been thinking about this in the context of my own work. I write essays. I've been doing it for thirty years. Could AI write my essays? It could write something that sounds like me. But it couldn't have the ideas. It couldn't be surprised by what it found while writing. It couldn't change its mind halfway through because it realized the original premise was wrong. That's the part that's still human. And I think it will stay human for a long time. But I'm not here to give you false comfort. There are real disruptions happening. Real people losing real jobs. And the transition is going to be messy. The tech industry is already seeing this. According to tracking by Yahoo Tech, layoffs in 2026 have hit Xbox, Apple, Oracle, Uber, TikTok, Meta, and others. Some of that is AI. Some of it is just the normal churn of an industry that's always changing. The question is what happens to the people who get laid off. Do they retrain? Do they find new roles? Or do they get left behind? I don't have a clean answer. Nobody does. But I know what doesn't work: pretending it's not happening, or pretending it's all going to be fine. The truth is, automation is a mirror. It reflects what we value. If we value efficiency above all else, we'll automate everything we can and wonder why people feel empty. If we value creativity, judgment, and taste, we'll find ways to use AI to amplify those things rather than replace them. I know which world I want to live in. And I know it's not guaranteed. So here's my advice, for what it's worth. Don't try to compete with AI on execution. You'll lose. Instead, focus on the things that are hard to automate. Taste. Judgment. The ability to know what's worth doing. And if you're in a job that's mostly execution? Start building the skills that aren't. Not because you're going to lose your job tomorrow. But because the ground is shifting, and you want to be standing on solid rock when it does. I've been wrong about a lot of things. I thought Viaweb would be a way to make money while I figured out what I really wanted to do. It turned into something else entirely. I thought YC would be a small experiment. It became an institution. But I've been right about one thing: the people who thrive are the ones who adapt. Not the ones who fight the change. Not the ones who pretend it's not happening. The ones who look at what's changing and figure out how to use it. That's what I'd tell anyone worried about AI taking their job. Don't worry about the job. Worry about whether you're becoming the kind of person who can do the things that still matter. The rest is just details. FAQ Q1: What impact will AI have on the job market? AI will automate jobs that are primarily execution-based—following scripts, filling templates, doing repetitive tasks. Jobs that require judgment, taste, and deciding what to do will be harder to automate. The net effect will be a shift in what skills are valuable, not a total elimination of work. Q2: Will automation replace human workers? Automation replaces specific tasks, not entire humans. The workers who thrive will be those who use automation to amplify their judgment and creativity. The workers who struggle will be those who compete with automation on execution alone. Q3: What's the most important skill to develop in an AI-first world? Taste. The ability to know what's worth doing. AI can generate options, but it can't tell you which option is good. That's still a human job. And it's only going to get more valuable.