The Comfortable Cage We Built

"Getting lost used to be structural. Now it has to be intentional." 

What do I cook tonight? Where do I get the best croissants in Paris? Which is the most picturesque place on the hilltops of Darjeeling? What is the best-reviewed museum in DC? What are the ten most exciting things to do in India? What are the top 10 best-selling books of 2025?

These are good questions. They are also, quietly, dangerous ones — not because of what they ask, but because of where we now ask them.

How We Used to Get Lost

There was a time — not so long ago — when answering these questions required a different kind of effort. You asked someone who had actually been there. A colleague who had spent a week in Paris. A cousin who had done the Darjeeling loop. A friend who had lived in DC for three years and had opinions about every museum.

Their answer came with texture. It came with caveats — 'The Louvre is extraordinary but go on a Wednesday morning quite early and exit to rue de Rivoli.' It came with the specific weight of memory, which is different from the specific weight of data.

And then you went. And you got a little lost. You walked down the wrong street and found a café that wasn't in anyone's notes. You ended up at a museum you hadn't planned because it drew you in. That lostness was not a failure of planning — it made for great stories and memories.

"Getting lost used to be structural. Now it has to be intentional."

When Optimization Moved In

At some point, the language of optimization — which had always lived in workspaces and spreadsheets and quarterly targets — crossed into everything else. Into what the kids should do this weekend. Into what we eat, how we exercise, what we read, where we holiday, how we rest.

Now we ask the internet.

The algorithm answered. And the answers were good. Genuinely good. The restaurant it recommended was excellent. The museum it surfaced was worth the visit. The efficiency was real.

But we lost something. Serendipity — which had always been a structural feature of how humans discovered things — became optional. And optional things, in optimized lives, tend to disappear.

The Problem Isn't the Answers. It's What the Questions Are Doing to Us.

Here is what I have been thinking about: exploration is not just a method of finding things. It is a way of becoming someone.

When you wander without a ranked list, you encounter things your past self would never have predicted you'd enjoy. A hill in Darjeeling where the tea garden workers gather at dusk and no travel blog has ever pointed a camera. A book you picked up because the spine was an interesting colour, which then altered how you thought about the next five years of your life — and the impromptu picnic where you stopped because your legs were too tired to wander any further.

Algorithms cannot surface these. Not because they are unintelligent, but because they are optimizing against a model of you that is always, by definition, out of date. The algorithm knows who you were last Tuesday. It cannot know who you are in the middle of becoming.

"A life lived through recommendations is a life lived in the shadow of who you already were — never the person you're in the middle of becoming."

Algorithms are quietly convincing us that being known is the same as being understood. They are very good at the former. They cannot do the latter.

What We Are Trading Away

The cost is not dramatic. No one tragedy is caused by asking Google for the best restaurant in Paris. But across a life, across thousands of such queries, a pattern sets in:

•       The unexpected meal becomes rarer. Not impossible — rarer.

•       The accidental friendship with a stranger on a train who recommends a book you'd never have found becomes a story we tell as if from another era.

•       The discovery that you, surprisingly, love a style of music or art or food you never knew existed — because you never wandered into the part of the city where it lives — becomes statistically less likely with every trip that is fully mapped before departure.

•       Children grow up with the sense that the answer to every question exists somewhere, retrievable, ranked, already sorted. That exploration is a style choice, not a way of knowing.

None of this is the algorithm's fault. It is answering the question we asked. The question is the problem.

An Invitation to Get Lost

I am not suggesting we throw the phones into the Ganga. The recommendation engine, used well, is a genuinely useful tool. The problem is not having it — it is confusing it for the whole of discovery.

What I am suggesting is something smaller and stranger: leave one question unanswered. Let the season and local markets decide what's for dinner tonight. Let a friend surprise you with her neighbourhood restaurant. And mostly, let the family time be unplanned.

Because the aha moment — the inexpressible, private joy of discovering something for yourself — cannot be optimized for. It can only be stumbled into. And the stumbling requires you to have left the path.

The best version of your trip, your dinner, your reading life, your Saturday afternoon is not waiting in a ranked list. It is waiting around a corner you have not yet decided to turn.

If this resonated, pass it to someone you would take on a trip. The best recommendation is still a human being who knows you and wants to surprise you anyway.

Namaste

Savitha