Ask someone who they are and they’ll tell you what they do.
“I’m a designer.” “I’m an engineer.” “I’m a writer.”
That’s not a shallow answer. Work gives you structure. It gives you a reason to get better, people who count on you, and a place in the world.
So when machines start doing big parts of that work, something real changes. Not just your income. A piece of what holds your identity up starts to move. That deserves honesty, not a quick “just become a creator.”
Here’s what actually changes.
The work doesn’t disappear. It moves.
David Heinemeier Hansson, the creator of Ruby on Rails, just said he barely writes code by hand anymore. He calls himself retired from programming. He says he’s a maker now.
That’s where a lot of knowledge work is heading. Less time producing first drafts. More time deciding what you want, checking what comes back, and fixing what’s wrong.
The job shifts from doing to judging. Is this right? Is this good? Is this what we actually meant?
Those questions were always part of the work. Now they’re becoming most of it.
But not everyone lands softly
Some roles will shrink. Some incomes will drop. Some people who spent a decade mastering a skill will watch it become cheap in a year.
The people who say this change is easy are usually the ones already set up to benefit from it.
Adapting is possible. It isn’t free.
The problem nobody has solved
Judgment doesn’t come from nowhere.
Jazz musicians build skills by transcribing and playing other people’s solos. Writers get good by writing hundreds of weak pages. Engineers learn by breaking things and fixing them at midnight.
The boring work is how you learn to recognize good work. If machines take the drafts, the junior tasks, and the repetitions, where does the next generation get those hours?
Today’s experts in many domains can manage AI well because they did the work by hand for years. Nobody knows how someone starting today gets there.
That might be the biggest open question of the next decade.
What gets automated, and what doesn’t
Machine translation is instant, free, and good enough for most everyday needs.
People still learn languages. Not to move information around. That part got automated. They learn to talk to their partner’s family, to get the joke, to feel at home somewhere new. To become someone who can.
The transactional part of a skill is easy to replace. The human part usually isn’t, because it was never about the transaction.
Expect that pattern to repeat across many fields. Not everywhere, and not without real damage along the way.
A few things stay human
Not in a poetic sense. In a practical one.
Deciding what matters. Tools are great at answering questions. They’re bad at choosing which questions deserve your life.
Taking responsibility. When something goes wrong, “the model did it” won’t satisfy a customer, a patient, or a court. Someone has to own the outcome.
Wanting other people. Recordings are free and people still go to concerts. Information is free and people still pay teachers. We keep valuing what comes from a person.
What to do about it
Nothing dramatic. Keep doing some things the hard way, so your judgment stays sharp.
Notice which parts of your work you’d keep doing even if they were automated. Those are probably the parts that matter most.
Get better at saying exactly what you want. When execution becomes cheap, vague goals become expensive.
And be skeptical of anyone who’s certain, in either direction.
AI will change what your work looks like, how you learn, and where your value comes from.
But the things that make you human were never really about the work. The work was always a way of expressing what you care about. That part is still yours.
Dan Berges is the founder and managing director of Berges Institute, an online Spanish language school, and lead developer of Berges AI, a text assistant built on open-weight models that gives direct, concise answers. He also publishes content in Spanish about descriptive grammar, semantics, and pragmatics on Instagram, YouTube, and TikTok.