The best way to protect your career in the age of AI is not to compete with the machine on what it does best. It is to cultivate what made the human race so successful through the ages: drive, curiosity, imagination, judgement, taste and, yes, personal responsibility. Those qualities were always valuable. As AI absorbs more of the work that used to fill our days, they are becoming the whole job.
This is the last piece in a series on what I call the capacity for greatness. It started with the idea that your ceiling is higher than you think. It ends with what that means for your career in a world where machines can do more every month.
A human beat the machine. For now
In July 2025, OpenAI entered its own AI into the AtCoder World Tour Finals, the most prestigious invitation-only programming contest in the world, open to only 12 top-ranked competitors a year. The Heuristic round is a 10-hour marathon on a single, fiendishly complex optimisation problem, with no libraries, no documentation and no outside help (TVP World, 2025).
The machine took the lead early with a straightforward strategy. Przemysław Dębiak, a 41-year-old programmer from Gdynia known as Psyho, answered with a far more sophisticated approach and slowly overtook it. The AI briefly reclaimed first place around the eighth hour. Dębiak's final optimisations won it, and the official scores put him 9.5% ahead of OpenAI's model, which finished second (Do IT in Poland, 2025). Sam Altman congratulated him publicly: "Good job Psyho."
There is a nice irony in it. Dębiak was one of the early engineers on OpenAI Five, the system that beat world champions at Dota 2.
I would not read this as proof that humans will keep winning contests like this. They probably will not. I read it as a picture of what the human edge actually is. The machine was fast and competent from the first minute. The human won with ten hours of drive, imagination and judgement about which ideas were worth pursuing.
The skills that are expiring, and the ones that are not
The World Economic Forum's Future of Jobs Report 2025 expects that, on average, 39% of workers' existing skill sets will be transformed or become outdated between 2025 and 2030. That is two in every five things you are good at today.
The same report is clear about what is rising. AI and big data top the list of fastest-growing skills, which should surprise nobody. But alongside them, creative thinking, resilience, flexibility and agility, and curiosity and lifelong learning are all expected to grow in importance. Analytical thinking remains the core skill employers rate most essential, with seven in ten companies considering it essential.
The honest reading is not that soft skills beat hard skills. It is that you need both. You need enough technical depth to direct the machine and check its work. But the technical part is no longer the scarce part. I learned this the hard way, after seventeen years as a CTO: the engineering was the easy part. Judgement decided the outcomes.
Six qualities the machine does not supply
None of these is new. Every one of them made our ancestors successful long before there was such a thing as a career.
Drive. A model does what it is asked. It does not wake up wanting more, and it does not keep going at hour nine because it refuses to finish second. Drive decides which problems get attacked at all.
Curiosity. AI answers questions brilliantly. Choosing which questions are worth asking, and noticing what nobody has asked yet, is still a human habit, and a learnable one.
Imagination. Picturing an outcome that does not exist yet, and a path to it, is where every product, company and career move starts.
Judgement. Deciding what is worth building, which trade-off to accept, and whether the output in front of you is actually right. When I went back to building software myself, I found the core skill had moved from recall to judgement.
Taste. Knowing the difference between good enough and good. As machine output becomes abundant, the ability to tell them apart, and the standards to insist on the better one, become rarer and more valuable.
Personal responsibility. A model cannot be accountable. Someone has to own the decision, stand behind the outcome and answer for it when it goes wrong. That someone is becoming the job.
What this means for your career
This is about you, not about how your organisation designs its workflows. That is a different question, and I have written about how to design humans into AI systems deliberately. Here the question is personal: what do you invest in?
Everything in this series points the same way.
Raise your ceiling. Most people stop long before their real limit, held back by pessimism, complacency or past success. In a market that reprices skills every few years, the people who keep climbing are the ones who stay relevant.
Widen your circle of responsibility. Your job description is the smallest circle of your impact. As AI takes over well-defined tasks, the work that remains is the work nobody defined: spotting the risk, connecting the teams, owning the outcome rather than the ticket.
Never settle for the minimum. Doing the minimum is a habit, and habits compound. Stanislav Petrov could have passed a false missile alert up the chain and nobody would have blamed him. He owned the judgement instead. More ownership, more imagination, more care about the outcome. Not more hours.
If you lead people, measure yourself with a different stick. Your job is to grow these qualities in others, and to make sure their precious time at work goes into meaningful outcomes rather than work a machine could do. People do not need to be taught to give up. They need to be shown that their effort matters.
If your growth scares your boss, take it somewhere it will be welcomed. Check your own part first, then leave well, without bitterness. The brightest people are often the first to go, and they are usually right.
The edge that compounds
Technical skills now expire faster than ever. The qualities in this piece do the opposite. Every year you practise curiosity, judgement and responsibility, they get stronger, and they transfer to whatever the next wave of technology turns out to be.
That is the real career plan. Not a list of tools to learn before they change again, but a set of human qualities that make you the person others trust to decide, to build and to answer for the result. The machine will keep getting better. So can you.
The philosophy behind this series, in my own words, is here.
If you are working out what AI means for your own team's skills and careers, I would be glad to compare notes. Let's talk.
Related: Your ceiling is higher than you think · After 17 years as a CTO, I learned the engineering was the easy part · One person and a fleet of AI agents now build like a whole team · Fractional CTO services
