The One Who Refuses is the One Who Decides

What does it mean to be “strong”?

Most believe it means the ability to push things through. Being able to haggle, forcing others to accept your terms, or manipulating people to move in your direction. But when I look at those who are actually strong, it feels different. The strong ones are those who can refuse.

Writing that you shouldn’t buy it

There is a grocery store chain called OK. Founded in 1967, they now have stores across Kanto and parts of Kansai.

In their stores, you’ll find things called “Honest Cards.” As the name suggests, these are cards placed right next to products that list the product’s flaws.

The company provides examples like these:

The grapefruits currently on sale are from South Africa and are a variety with strong acidity. Delicious Florida grapefruits are scheduled to arrive in December.

Due to prolonged rain, the quality of the lettuce is lower than usual and prices have soared. We recommend substituting this with other products for the time being.

The store is explicitly telling customers not to buy what is currently on the shelf. Furthermore, they tell them exactly when better stock will arrive. The store is systematically destroying the reasons to buy.

The company explains their reasoning for this: they wanted to recreate the experience of a neighborhood greengrocer who would tell you, “These vegetables aren’t great today, so you’d better pick something else,” but on a supermarket scale.

Looking at each card individually, this is a lost sale. Today’s lettuce will not sell.

What they refused was immediate revenue; what they gained was a state where customers trust them when they do make a recommendation.

Deciding to make people wait

In 1972, a man named Satoshi Sakurada, who had left a securities firm, opened an experimental shop of 2.8 tsubo in front of Narimasu Station in Tokyo. That same year, he opened the first official store: Mos Burger.

From the start, he decided not to pre-make burgers. They make them only after receiving an order—a system they call “after-order.”

In the world of fast food, this is strange. Speed is the product, yet they voluntarily let go of that speed. Customers have to wait. Moreover, because it requires more effort, the price is higher.

How much higher? In 1995, when a McDonald’s hamburger was 80 yen, a Mos Burger was 210 yen. In 2000, even when McDonald’s dropped its weekday price to 65 yen, Mos stayed at 210 yen. The gap grew to more than threefold. In 2002, McDonald’s hit 62 yen, and in 2005, the “100-yen Mac” began.

Mos did not enter this competition.

And they actually suffered for it. The number of stores peaked in fiscal 2000 and continued to decline, dropping from over 1,500 to around 1,200. Three hundred stores vanished.

Even so, they did not start pre-making burgers.

What they refused was speed and cheapness; what they gained was a state where they could decide their own prices.

Deciding to cut gross profit

In November 2006, NVIDIA released a component called the GeForce 8800. It was a semiconductor for rendering images on screens, and at the time, its performance left others in the dust. The company was profitable thanks to this speed.

In the same year, they released a framework called CUDA. This allowed GPUs to be used for calculations other than rendering images.

To make this possible, additional circuitry had to be added to the chip. This increased the surface area and raised the cost of production. Moreover, the people buying these were gamers, for whom this feature had no use.

Normally, you would only include such a feature in the high-end products. Or you wouldn’t include it at all. Since the reason the company was winning was its image-rendering speed, it would be more profitable to double down on that.

They didn’t. They put it in every chip. Even the cheap ones. The founder later said:

We added enormous costs to the chips. There were almost no CUDA customers, but we still made every chip CUDA-compatible. Just look at the gross profit margins. They were bad from the start, and then they got even worse.

The numbers remain. The first CUDA-compatible design took four years and 475 million dollars—roughly one-third of their R&D budget at the time. Gross profit margins dropped from 45% in 2007 to 35% by 2010. Shareholders complained, and some left.

From 2006 to around 2012, almost nothing happened in terms of business. For six years, it was simply a drain on resources.

Even now, gaming cards have CUDA. For people who only play games, it remains a feature with no use.

What they refused was gross profit; what they gained was a state where any card could perform calculations.

Deciding not to sell rankings

In 1998, Sergey Brin and Larry Page, then graduate students at Stanford, published a paper on search engines. This became the foundation of today’s Google.

At the end of the paper, there is a short section called Appendix A. The title: “Advertising and Mixed Motives.”

We believe that a search engine funded by advertising will inherently be biased toward advertisers and move away from the needs of the user.

They even gave an example. At the time, searching for “mobile phone” on their prototype brought up research on the dangers of mobile phones while driving at the top, because it was linked to by many other sites. They wrote that a search engine taking money from mobile phone ads would be unable to explain this result to its advertisers.

They wrote, before even building it, that the moment the thing they were creating accepted money [for rankings], it would break.

What happened next? Alphabet’s revenue in 2024 was 350 billion dollars, and over 70% of that is from advertising. They didn’t refuse advertising. What they refused was selling the rankings themselves. You can pay to get into an ad slot, but the order of the organic search results does not move.

What they refused was the sale of rankings; what they gained was a state where rankings could be trusted.

Without the power to refuse, there is no price and no freedom

In all four cases, things happened in the same order. First, they decided what to refuse, and once decided, they made no exceptions regardless of who the opponent was.

And the sequence often looks reversed to an observer. It wasn’t that they became strong and then were able to refuse. OK is just a local grocery store; Mos Burger lost 300 stores while refusing to lower prices; Google in 1998 was just two graduate students. NVIDIA alone was winning at the time they refused, but they were winning because of rendering speed—the cut gross profit vanished into something entirely unrelated to that success. In every case, at the moment of refusal, there was no obvious reason why it would be beneficial.

Setting a price is an act of refusal. It is saying, “I will not sell unless it is for this amount.” Therefore, a seller who cannot refuse cannot set a price. They are bought at the buyer’s dictated price, and that’s the end of it. Whether a price hike works is the same problem as whether one can refuse.

Freedom takes the same form. If you cannot refuse, it means you are submitting to that party. Mos didn’t match McDonald’s prices because they didn’t have to. Google’s rankings don’t move for money because they are not submitting to advertisers. The scope of what you can refuse becomes the scope of your freedom.

So, why are they able to refuse? Because they possess their own yardstick. An absolute yardstick that differs from others’. What OK measured was honesty, not revenue. For Mos, it was whether the food was freshly made; for NVIDIA, whether calculations were possible; for Google, whether the ranking was correct. Speed, gross profit, and revenue do not exist on those yardsticks.