When did we stop trusting our own taste?

Hello

It’s Thursday, 22^nd October 2026. Welcome to Bold Efforts!

Think about the last time you chose a restaurant. Chances are, you checked Google Maps, looked at the ratings, read a few reviews, and perhaps searched Instagram to see what the food looked like. I do this all the time, and it generally works. I've avoided plenty of disappointing experiences because someone else took the trouble to leave a review.

We've developed similar habits across much of our lives. Before watching a film, we check IMDb. Before buying a book, we look at Goodreads. Before visiting a new city, we consult countless lists of the best places to visit. Sometimes, even after enjoying a film, we'll look up what critics thought of it, almost as though we're checking whether our enjoyment was justified.

Recommendations are useful. We have more choices than we could possibly explore, and our time is limited. But I've been wondering whether we've become so accustomed to external validation that we're gradually losing confidence in our own preferences. A restaurant with a 4.8 rating might serve excellent food, but that doesn't necessarily mean you'll enjoy it more than the little place around the corner.

There's been considerable discussion lately about taste becoming one of the most valuable skills in the age of AI. When anyone can generate code, images, designs, or writing with a few instructions, the ability to produce something becomes less of a differentiator. What matters increasingly is knowing what should be produced and recognizing quality when you see it.

Taste isn't limited to aesthetics. A good editor knows which sentence to remove. A good product designer recognizes which features don't belong. A good founder knows which opportunities to pursue and which to ignore. These are different disciplines, but they all require the ability to make informed distinctions.

Consider two people using AI to design a website. One accepts the first professional-looking result, makes a few adjustments, and publishes it. The other questions the layout, rethinks the typography, removes unnecessary elements, and keeps refining until the design communicates exactly what they intended. Both have access to the same technology, but their results could be remarkably different. The difference is their judgment.

As AI makes it cheaper and easier to produce acceptable work, the ability to distinguish the exceptional from the ordinary could become increasingly valuable.

What's interesting is that we're celebrating taste as a valuable human capability at precisely the moment technology is making it easier to avoid exercising it.

We already rely on recommendation systems to decide what to watch, listen to, buy, and read. Now AI assistants can help us evaluate ideas, critique designs, recommend strategies, and make increasingly complex decisions. Often, their advice is genuinely useful.

But there's a difference between asking AI to challenge a judgment you've formed and asking it to make that judgment for you. Both approaches might produce a good outcome today, but they don't necessarily leave you equally capable of making a decision tomorrow.

Good judgment develops through experience, observation, and mistakes. A writer develops taste by reading widely, writing badly, receiving criticism, and gradually understanding what makes something work. A founder builds product judgment partly by launching things customers don't want. We learn to recognize quality by repeatedly encountering it, comparing alternatives, and forming our own opinions.

If we routinely outsource these decisions, we risk missing out on the very experiences that help develop judgment.

This extends beyond work. Imagine an AI assistant that knows your favourite cuisines, travel history, reading habits, and lifestyle preferences. It could make remarkably personalized recommendations about where to eat, what to read, and even where to live.

That sounds useful, but our preferences aren't fixed. They evolve as we encounter unfamiliar experiences and discover things we didn't know we'd enjoy. A system that becomes exceptionally good at predicting our existing preferences might not always help us develop new ones.

I don't intend to stop checking reviews or asking AI for suggestions. There's no virtue in making uninformed decisions simply to demonstrate independence. The question is whether we use these tools to become more discerning or merely more dependent.

Developing taste requires paying attention to our own reactions, learning from people who know more than we do, and being willing to disagree. Sometimes that means choosing a book because its premise intrigues you rather than because it's popular. At work, it might mean questioning an AI-generated recommendation instead of accepting something that looks reasonable.

The irony is that just as technology makes our taste more valuable, it also makes it easier to stop developing it.

Perhaps the challenge isn't learning how to get better recommendations. It's becoming better at deciding what to do with them. Thank you for reading.

Best,
Kartik

I write Bold Efforts every week to think clearly about where work and life are actually headed. If you want these essays in your inbox, you can subscribe here.