The things technology should not optimize

Technology is very good at answering one kind of question: how can we get more? More output, more engagement, more speed, more growth. Give a system a clear objective and it will search relentlessly for ways to improve it.

That ability has transformed the world. It helps planes consume less fuel, doctors detect diseases earlier, and small teams accomplish things that once required entire organizations. Optimization is powerful when the goal is clear.

The problem begins when we let the system choose the goal.

A metric is meant to represent something we care about. Over time, the representation can quietly replace the thing itself. Time spent becomes a proxy for attention. Engagement becomes a proxy for interest. Productivity becomes a proxy for progress. Growth becomes a proxy for success. The numbers may improve while the underlying experience gets worse, because time spent is not the same as time valued, engagement is not the same as meaning, and a productive day is not the same as a worthwhile one.

I spent much of my career building products inside large technology companies. Metrics were indispensable. They made vague conversations concrete, exposed weak assumptions, and showed us what intuition could not. But a metric can only tell us whether we are moving efficiently toward a destination. It cannot tell us whether the destination is worth reaching.

That choice remains ours.

Some of the most valuable parts of life resist measurement precisely because they were never designed to produce an outcome. A long dinner with friends. A walk without a destination. A book that changes you slowly. An afternoon spent making something no one asked for. From the perspective of an optimization system, these moments look inefficient. Their value often comes from the absence of a measurable return.

The same is true of attention. The internet treats attention as inventory to be captured, and every notification, headline, and recommendation competes for another moment of consciousness. The systems behind them are becoming extraordinarily good at predicting what will keep us watching. From the perspective of the system, this is success. From the perspective of the person inside it, the answer is less obvious: a platform gains another minute while a person loses the empty space in which an original thought might have appeared.

This rarely feels dramatic. It happens in fragments. A screen opened while waiting for the elevator. A conversation interrupted by a notification. A moment of boredom removed before it has the chance to become curiosity. Eventually, constant stimulation begins to feel normal and silence starts to feel uncomfortable.

Technology has optimized our attention. It has not necessarily improved it.

We increasingly apply the same logic to ourselves. We optimize our mornings, our sleep, our diets, our careers, and our relationships. We measure steps, focus, recovery, output, and progress. Some of this is useful. Measurement can expose the distance between our intentions and our behavior. But a day can meet every target and still feel empty. A career can advance exactly as planned toward a life we no longer want.

No dashboard warns us when the person setting the goals has changed.

As AI becomes more capable, this distinction between means and ends will matter more. Machines will help us move faster, generate more options, and pursue objectives with greater precision. They will increasingly answer the question of how.

They cannot decide why.

An AI system can help a company maximize engagement. It cannot determine whether more engagement is good. It can help us work faster, but it cannot decide what deserves our time. It can produce infinite variations, and it cannot tell us when enough is enough. Those decisions require context, values, restraint, and responsibility.

The most important question of the next decade may not be what technology can optimize. It may be whether we remember to choose the objective ourselves.

Technology should help us use our time.

It should not decide what our time is for.