AI doesn't take work away from you.

It just shows you how much work would be possible.

Diesen Artikel auf Deutsch lesen

There’s a comforting story going around right now: AI makes us more productive, saves time, and frees us from work. That’s the official version. The unofficial one is considerably less pleasant: many people are working more than before. Not because they have to. Because they can.

Work used to be constrained by things nobody actually liked, but that turned out to be surprisingly useful: time, friction, complexity. These weren’t bugs in the system — they were a kind of built-in circuit breaker. They forced you to stop at some point. Eventually you’d hit the moment where the effort simply wasn’t worth it anymore. “Good enough” wasn’t a philosophy. It was a necessity.

That moment has disappeared. Or at least been pushed very far back.

You can think faster, build faster, discard faster. You can chase ideas without committing to them. You can exist simultaneously in multiple versions of your own thoughts. And something happens in that space that feels like freedom at first, but is actually something else: the external limits disappear, and suddenly you are the only limit.

The problem is: humans weren’t built for that.

We’re remarkably good at optimizing within constraints. But when those constraints fall away, clarity and calm don’t automatically follow. Escalation does. Not because anyone forces you, but because every additional possibility feels like an invitation. One more feature, a better version, one more thought. Not necessary — but possible. And that’s exactly where the shift begins: possibility gets mistaken for necessity.

AI amplifies this pattern perfectly. It doesn’t tell you what you should do. It only shows you what you could do. And that’s the decisive difference. “Must” has an end. “Can” has none.

This changes the nature of work itself. The bottleneck used to be production. Building things, executing ideas, developing systems — all of that was hard. Today, it’s trivial. In minutes, you produce what used to take days. But the real difficulty hasn’t disappeared — it’s just moved. It now lives in evaluation.

You can generate more than you can meaningfully understand.

When production becomes free, judgment becomes the bottleneck.

This is where a principle older than any AI debate suddenly gains new relevance: Brandolini’s Law. The simple observation that it takes far more energy to refute nonsense than to produce it. Before, at least producing nonsense was costly enough to keep things roughly stable. Today, one prompt is enough. The production side has become practically free. The evaluation side hasn’t.

The result is a system that generates faster than humans can meaningfully filter.

This dynamic isn’t limited to developers. It runs through entire organizations. Suddenly, people with no previous proximity to implementation can produce things themselves: code, concepts, drafts. Not necessarily correct, but fast enough to be taken seriously. And each of those artifacts creates work on the other side. Someone has to review it, contextualize it, correct it — or often just explain why it won’t work.

Work is no longer created by scarcity. It’s created by abundance.

And that shifts where value comes from. It used to live strongly in the ability to produce. Now it’s moving toward the ability to decide. To decide what matters. To decide what can be ignored. To decide what not to build. Because when production gets cheap, judgment becomes the scarce resource.

That’s not a small shift. It’s a fundamental one.

And here’s the real paradox: AI saves you time. And you don’t use that time to work less — you use it to do more. More ideas, more variants, more options, more polish. Every efficiency gain gets refilled immediately. Not consciously. Systematically.

This isn’t new behavior. It’s a very human one.

The Industrial Revolution reduced physical labor, and we created new work. The digital revolution accelerated information work, and we produced more of it. Now AI is making cognitive work more efficient — and we’re reacting exactly the same way. We don’t optimize for less work. We optimize for more output.

The problem isn’t technological. It’s cultural.

We have no clear sense of what “enough” means. No metric for “done.” No reward for deliberately not doing something. But we have highly sophisticated systems for “more,” “faster,” and “better.” And as long as that stays true, every new technology will amplify this dynamic rather than dissolve it.

That’s why the most important skill of the future may be shifting somewhere unexpected. It has less to do with tools, less to do with models, less to do with the next generation of systems. It’s about the ability to limit yourself. To say: this is enough. This is good enough. I’m not doing that one too.

Because AI won’t stop you. It will always show you a better version, a faster variant, one more option. And every time, the decision is yours — whether you follow that thread or not.

In the end, AI doesn’t take work away from you.

It just removes the natural limits that used to protect you from yourself.

What remains is you.

And your ability to stop.

Or not.