How to Choose Your Career in the AI Era (Skip the Passion Advice)

Three circles, and why most people start in the wrong one

The simplest way to think about a career is a Venn diagram with three circles: what you are good at, what pays well, and what you love (or could grow to love). Ideally you land somewhere in the middle.

The mistake most people make is starting with "what do I love?". We are told from childhood to follow our passion and everything else will fall into place. I see some truth in that, but I do not fully believe it, especially now. The world is more competitive and AI is making basic skills much cheaper. The market does not care how passionate you are. It cares whether you are useful.

So I would start with the first circle: what are you good at? Not world-class at 18, but how your brain naturally works. Are you strong at maths and logic, or visual thinking? Good with people? Creative? Organised? A natural technical problem-solver? I studied accounting and finance, then taught myself to code. I liked numbers, strategy and building things, and that combination eventually led me to startups. If you are naturally weak at something, choosing one of the most competitive careers in the world because it looks fun is a risky bet.

What does the market actually pay for?

This circle matters enormously, because careers follow supply and demand. The jobs people think are fun are, by definition, jobs lots of people want. Gaming, acting, music, content creation: millions would do them for free because they are fulfilling. When that many people want a role, employers do not have to pay a premium for it.

Compare that with work that is technically demanding, stressful or just difficult. Nobody wakes up dreaming of hard things, so the economics are very different. You do not need to love your job. If you are good at it, it pays well, you do not dread mornings, and it gives you a good life, that is a genuinely good outcome. Your career does not have to meet every emotional need. You can love things outside work.

If all three circles overlap, brilliant, you have found the ideal role. But most people should not expect that on day one. Two out of three at the start is very good. If you want a structured way to test your fit, try our career compass or read which tech career fits me.

Do not get attached to a job title

Even if you find the perfect career today, the career itself will change. Look at software engineering. Ten years ago it mostly meant manually writing code. Today you might spend more time describing what you want, reviewing AI-generated code, debugging, designing systems, talking to users and making higher-level decisions. Someone who got into tech because they loved typing code by hand has watched the job evolve under them.

So do not become emotionally attached to a title. Become attached to your ability to adapt. Industries, technologies, companies and markets all change. What you love doing at 23 may barely resemble the same career at 33.

AI is removing the bottom rung of the ladder

Traditionally you graduate, join as a junior, do the simpler work, a senior checks it, and over three to ten years you become the senior. Now think about what juniors do. A junior lawyer researches cases and drafts. A junior marketer writes copy and basic campaigns. A junior analyst builds spreadsheets. A junior developer ships small features and simple bug fixes. AI is getting very good at exactly these things.

That does not mean every junior job vanishes tomorrow. But the number of people needed for the same output can fall sharply. If one experienced developer with a coding agent produces what several used to, companies will think much harder about hiring. That is the brutal part of markets: companies do not employ people because society needs entry-level jobs. They employ people to solve problems. So the bar for being useful goes up. I wrote more on this in AI took the bottom rung.

Which creates a catch: how do you get experience when nobody hires inexperienced people? Increasingly, you have to close part of that gap yourself.

Closing the gap with proof

In software, instead of finishing a course or a CS degree and saying "I know React, please give me my first job", build something. A real, production-ready application. Deploy it, get users, solve an actual business problem, talk to those users, fix the bugs, maybe get someone to pay for it. Then in an interview you are not saying "I have no experience, I am here to learn". Everyone says that. You are saying: "I built this from zero. Here are the problems I hit, here is how I solved them, here is how people use it."

The same applies everywhere, because AI is now in every industry like the internet. Want to be a marketer? Grow something. Sales? Sell something. Designer? Design and build a portfolio. AI and automation? Build automations for a real business, because most businesses are full of boring admin that can be automated. Hiring is becoming proof-based rather than certificate-based, and AI makes proof cheaper to create than ever. Our guide to getting a tech job without experience goes deeper.

Which careers are harder to automate?

I do not think any career is completely AI-proof. But some categories are harder to automate:

University, shovels and small teams

This changes the university calculation too. The question becomes: what exactly will this degree buy me? For medicine, dentistry and other regulated professions, you need the degree. But spending three or four years and a lot of money on something employers barely value deserves serious thought, especially if the entry-level jobs at the end are getting scarcer. I have a degree myself. I am not saying university is useless. I am saying the bar for justifying it is higher, and "everyone else is going" is a weak reason.

There is also the gold-rush lesson: you do not have to dig for gold, you can sell shovels. You do not need to build the next OpenAI to benefit from AI. Help a traditional SME implement it. Build automations. Run a two-person agency that used to need ten. Become a far more productive engineer, or a marketer testing 50 creatives instead of five. Sometimes the easier opportunity is simply riding the wave by becoming very good at using the technology.

AI also makes it possible to run a business with five or six people that once needed 20 or 30. Expect many more small startups, remote teams, freelancers and solopreneurs. But I will not romanticise it. In a three-person startup there is nowhere to hide, income can be unpredictable, and some people hate that. Know your personality.

If I were 18, I would decide in this order

First, what am I naturally better at than most people around me: quantitative, verbal, people, technical, creative, entrepreneurial? Second, which careers use those strengths? Third, which of those have the strongest economics? Does the market need this, and what happens if I become one of the best at it? Use AI to research, but do the thinking yourself.

Fourth, look at AI, not as "will AI touch this?" because it will touch everything, but as: if I become excellent at this and use AI well, does AI make me more powerful or take away my value? Finally, from what remains, choose what you could enjoy, or at least not hate, and could learn to love.

That is very different from starting with passion. Competence usually comes before passion. When you get good, you get results, respect, autonomy and money, and very often you end up enjoying it. The lesson from AI is not that young people are doomed. It is that the bar keeps rising: being average at something AI can also do will be low value, while being excellent, solving hard problems, taking ownership and using AI well will be worth more.

Whatever you choose, some building blocks pay off everywhere: understanding the principles of how software works, how APIs connect things, and how to learn new tools quickly. You do not need to go deep, but understanding how things work is becoming basic literacy. If you want a hands-on, free way to test whether technical work suits you, start with our free trial.