Will AI take my job? A better question to ask

For 30 years I've built software for a living. Today I'm going to say something most people in my industry won't: the market for what I used to do is disappearing.

Not slowly, either. The kind of software I spent decades building can now be produced by almost anyone willing to describe the problem clearly. The value of it is heading towards zero. I've watched it happen in my own business, in proposals that didn't land and clients who realised they could do it themselves.

So when people ask me, "Will AI take my job?", I won't give them the comfortable answer. For many white-collar roles, part of the job is already gone. AI taking jobs isn't a future scenario; it's happening one task at a time. Pretending otherwise doesn't help anyone.

The part that gets missed, though, is this.

What AI is actually replacing

AI is extraordinarily good at production: writing the code, drafting the contract, building the report, summarising the file. When people list the jobs AI will replace, they're really listing jobs made mostly of production. If your value lived mainly in producing those things, that value is under real pressure.

But production was never the whole job. When production becomes cheap, value moves to everything around it:

• Deciding what's worth doing. Anyone can build now. Very few people know what to build, or whether they should.
• Judgement. Is this output right? Does it fit this business, these people, this moment? What's missing?
• Accountability. Someone still has to own the decision and live with the consequences.
• Trust. People still buy from, follow and confide in people.
• Helping humans change. Most projects never failed on technology. They failed because people didn't adopt, resisted, or lost heart.

Those things aren't getting cheaper. They're getting scarcer, and so more valuable. If you're looking for jobs safe from AI, this is where to look: not at job titles, but at the parts of any job that are made of judgement.

The real question to ask

So I think the useful question isn't "Will AI take my job?" It's this:

Which parts of my work were production, and which parts were judgement?

Most people have never had to separate them. The two were bundled together, and we were paid for the bundle. Now they're being pulled apart, and the future of work belongs to people who can see clearly which part is theirs.

The harder part: identity

The harder part isn't skills. It's identity.

Many of us don't just do our work, we are our work. "I'm a developer." "I'm a lawyer." "I'm an analyst." When the market for that craft shifts, it doesn't feel like a career problem. It feels personal. That's why so much of the anxiety about AI job loss sits deeper than a skills gap.

I know that feeling well, because I'm in it. After three decades, I've had to ask myself what I'm actually good at, beyond what I've produced. The answer turned out to be the parts I used to treat as side effects: reading a situation, making calls under uncertainty, and helping people through change.

That shift is slower and messier than any reskilling course suggests, especially if you're weighing a career change at 40 or 50. A bit of grief is part of it, and that's fine. What matters is not letting the loss of a craft become the loss of your footing.

Steady judgement

AI isn't taking you. It's taking the part of your work that could always, in principle, be done without you.

What's left is the part that was always yours: your judgement, your relationships, your willingness to own a decision, and your ability to help other people through change. That part is worth building a working life around.

That's the work I do now: helping founders, business owners and professionals find their footing when the ground shifts. Steady judgement in a changing world.

Common questions