[Eng] A World Programmed to Be Perfect

There are five levels for evaluating user search query results: Perfect, Excellent, Good, Acceptable, and Off-topic. When a query clearly expresses its search intent, these levels are fairly easy to distinguish. But with vague queries—which users often input—the system relies on popularity to offer a so-called “perfect” answer.

For example, if a user types "Ed Sheeran Perfect lyrics", the intent is crystal clear: they’re looking for the lyrics to Ed Sheeran’s song “Perfect.” But for a query like “perfect song”, the system must rely on what is most frequently searched and then suggest a “perfect” result—usually the most famous song titled “Perfect”.

Tech companies use human evaluators to assess these kinds of queries, aiming to create the best version of their AI. In doing so, they can more easily grasp the psychology behind users’ searches. Gradually, a world of “perfection” begins to form—where every query comes with a neatly packaged answer.

In this world, people no longer have to synthesize information themselves to arrive at the answer they seek; everything comes more easily. AI helps users save tremendous amounts of time reaching their destination. But what if that destination isn’t what the user truly wanted?

Every search has both a primary and a secondary intent. Contrary to what many might assume, it is often the secondary intent that leads people to explore new dimensions of knowledge—these intents are spontaneous, unexpected.

For instance, when someone searches “best hunting dog,” they might only want information about fast and agile breeds. But a secondary intent could lead them to learn about the history of purebred dogs, the ethics of breeding, or even regional hunting dog cultures. These angles are often ignored by AI because they don’t directly serve the primary search goal. Over time, users risk losing the ability to broaden their thinking if they rely too heavily on machine-generated answers.

AI does not possess human curiosity. That is a fact. AI lacks creativity. This, too, is intrinsic to the nature of machines—and it’s what sets humans apart. Curiosity is what drives people to explore, to ask follow-up questions, and to go further than the original goal. AI, on the other hand, runs on logic, aiming for precision, compression, and efficiency. It doesn’t understand the productive kind of digression—the type that sometimes leads to surprising, creative, out-of-the-box discoveries.

It becomes dangerous when users begin to believe that AI’s answers are the final word. Doubt disappears. Users stop questioning, stop digging deeper. They become too satisfied with the suggested answer and forget to ask, “What else is out there?” This can lead to narrow, dependent thinking—lacking both depth and expansion.

“Doubt is the origin of wisdom.” – René Descartes

Artificial intelligence is an unstoppable trend—21st-century humans are not Luddites. But we still have the right to question, even when every result is wrapped in a shell of “perfection.” No machine can replace human curiosity. Don’t stop at the first answer. Doubt it. Explore further. Let those “secondary intents”—those unexpected detours—lead you to discoveries that no algorithm can ever program.