The Weight That Algorithms Cannot Replace: Why We Shouldn't Become Believers in the Cult of Science

There’s an imagination that goes like this:

If we could analyze every technical chart pattern, the system could automatically issue buy and sell signals. If we could analyze a user’s age and assets, the system could automatically recommend the most suitable insurance product. If we could collect enough behavioral data, the algorithm could predict what you’ll do next.

“Systematize decision-making”—this has been the basic faith of countless engineering-minded startup teams over the past decade or more.

But this is less logic than it is religion. They are believers in the Cult of Science.

That Unbridgeable Chasm

From organizing data to making a decision is not a continuous spectrum but a zoned process, with a cliff in the middle.

Let me break this process down. Data comes in three forms: structured, semi-structured, and unstructured. Structured data is the numbers in a spreadsheet—revenue, user counts, conversion rates. Semi-structured data is information with a certain format but requiring interpretation—customer feedback emails, meeting notes, market reports. Unstructured data is the stuff floating in the air—a glance, an unspoken word, a gut feeling that “this direction doesn’t feel quite right.”

What platforms and algorithms can handle perfectly is basically only structured data. You give it a table, and it can sort, filter, find outliers, and draw trend charts for you. This is useful, but it’s still far from “decision-making.”

Because a decision is a subjective leap of thought—data can’t give you that final step.

You’ve read ten market reports, and the data all points toward direction A. But at a dinner yesterday, you heard an industry veteran offhandedly mention something that left you with a vague sense that direction B is the right one. You can’t quantify the weight of that remark, and you can’t even explain to your team why you “feel” B is better. But in that instant, you simply made the decision.

What happens in that instant is something algorithms cannot see and cannot simulate.

The Faith Structure of the Cult of Science

I call the attitude of over-believing that algorithms can replace human decision-making the “Cult of Science”—not to insult rational thinking, but because its structure bears a striking resemblance to religious faith.

The core of religious faith is: there exists a transcendent being (God, fate, the law of cause and effect) that can eliminate uncertainty for you. As long as you believe, you have the answer.

The core of the Cult of Science is: there exists a transcendent system (big data, AI, algorithms) that can eliminate uncertainty for you. As long as you feed it enough data, you have the optimal solution.

The psychological driver common to both is the same: avoiding the pain of decision-making.

Making a decision is painful. That present-moment ambiguity, equivocation, and uncertainty is a weight everyone must face alone. Career choices, investment judgments, the trade-offs of relationships—these are all unstructured subjective judgments, carrying the warmth, bias, and willfulness of human nature.

Believers in the Cult of Science think that algorithms can shrink the vast chasm between semi-structured and unstructured. But the chasm hasn’t shrunk. It’s just that, because they “believe,” they feel the problem has disappeared.

A Warning from 2017, Worth Heeding Even More in 2026

I first wrote down these ideas in 2017. Back then, AI didn’t yet have the halo of ChatGPT, big data was the trendiest buzzword, and every founder was saying “data-driven.”

Today, nine years later, AI’s capabilities have indeed made a qualitative leap. GPT can write articles, Claude can analyze contracts, and various AI agents can automatically execute complex tasks. But the core problem hasn’t disappeared—on the contrary, it has grown sharper.

Because as AI increasingly looks like “something that can make decisions,” people become all the more inclined to offload the responsibility of deciding onto it.

In running a company, I’ve seen too many situations like this. A team uses AI to do market analysis, AI says a certain market opportunity scores highest, so they decide to head in that direction. No one asks: “Who designed AI’s scoring model? Are the scoring weights reasonable? Are there factors the model can’t see?” This isn’t AI’s fault—it faithfully produces results based on the framework you give it. The problem is that people have abdicated the responsibility of questioning whether the framework itself is correct.

This is the same issue I discussed in The Post-Code Era: When Taste Becomes Humanity’s Key Competitive Advantage. AI can find the optimal solution within the coordinate system you define, but defining the coordinate system itself—what is “good,” what is worth pursuing, what risks are tolerable—these have always been human responsibilities.

The Real Role of Information Platforms

Let me be clear: I’m not anti-technology. Quite the opposite—I use AI every day.

But I’m very clear about AI’s role in my decision-making process: it is a “decision-support tool,” not a “decision-dependency system.”

This distinction is extremely important.

Decision support means: AI helps me organize structured data, performs preliminary classification of semi-structured information, and provides options and analysis within the framework I set. Then, I bear the responsibility for that final “leap.”

Decision dependency means: AI tells me what to do, and I do it. If the outcome is bad, it’s AI’s fault.

If a platform claims it can help you make better decisions and judgments—use this feature to buy stocks that will rise more, use that model to select the most suitable employee—I said nine years ago that such a platform is either a fool or a fraud, and today my judgment hasn’t changed.

Not because the technology can’t do it. Because the essence of decision-making isn’t an optimization problem. It’s a process of making choices under incomplete information, carrying value judgments. You can have better information, but you can’t eliminate the weight of the act of “choosing” itself.

Decision’s Pain Through the Lens of Theology

In my theological training, I learned a concept that I later found tremendously helpful for understanding decision-making: finitude.

Christian theology has a core premise: humans are finite beings. Your knowledge is finite, your vision is finite, your understanding is finite. This isn’t a flaw—it’s a fundamental condition of existence. Accepting finitude isn’t giving up the pursuit of better; it’s acknowledging the fact that “it’s impossible to know everything,” and then, on the foundation of that fact, making the most responsible judgment you can.

The problem with the Cult of Science is precisely that it refuses to accept finitude. It assumes that as long as you have enough data and good enough models, you can approach “omniscience.” But omniscience is an attribute of God, not of humans. Projecting this attribute onto algorithms is essentially a form of idolatry—dressing up a longing for certainty in the garb of technology.

In The Algorithm as Judge, I explore this problem further: when we let algorithms judge a person’s worth and allocate a person’s opportunities, what exactly are we trusting?

The pain of decision-making won’t disappear just because we have better tools. Tools can let you see more, calculate faster, and simulate more scenarios. But that final instant of “I’ve decided” remains your responsibility alone.

The Prerequisite for Using AI Well

Back to the practical. If you’re an operator currently using AI to assist with decisions, I suggest you stick this sentence next to your screen:

The quality of AI’s output will never exceed the quality of the question you input.

AI won’t tell you that you asked the wrong question. It will only faithfully answer the question you asked, no matter how absurd that question is. So your responsibility isn’t to learn how to use AI, but to learn how to ask the right questions. And asking the right questions requires deep understanding of the domain, sensitivity to human nature, and the humility to admit you might be wrong.

Thank goodness so many people in the world are willing to believe that algorithms can solve everything. Their existence gives those who truly understand the essence of decision-making an inimitable competitive advantage.

Because in a world where everyone has the same AI tools, only one source of differentiation remains: how you use it, and whether you dare to choose B when AI says A, because of your own judgment.

There is no law of averages. There is no unified equation for happiness. The weight of decision is the weight of being alive.