MarsBit has published a bylined commentary that takes one of the most memorable lines from The Three-Body Problem — "only send the brain" — and uses it to examine business history and organizational design in the AI era. The article argues that the value of the phrase lies less in its science-fiction setting than in the method behind it: under severe technical or resource constraints, keep only the irreplaceable part of the task and leave the rest for a time when conditions improve.
The starting point: the staircase project in The Three-Body Problem
The article opens with the staircase project described in The Three-Body Problem. In the author’s retelling, the original idea was to send a frozen human body as a spy to the Trisolarans, but a full human body was too heavy to reach the required speed. Wade, who was responsible for the plan, proposed a brutal alternative: "only send the brain."
The author treats that line as a rule for hard trade-offs. When a mission is close to impossible, the question is not whether the whole thing can be completed exactly as first imagined. The first step is to identify the one part that cannot be replaced and ask whether that alone can be delivered. In the article’s description, the brain is the sole carrier of consciousness and memory, and the Trisolarans, with technology far beyond that of humans, would be capable of restoring that consciousness.
From there, the article distinguishes between two kinds of thinking. Ordinary tasks follow the logic of preparation first. Near-impossible tasks follow the logic of "only send the brain." The author also refers to an earlier essay titled A Standard Tutorial for Ordinary People to Make a Film for 100,000 and Get It Released Without Losing Money, where a similar approach was used to discuss ultra-low-cost filmmaking. In that example, conventional thinking says a film requires substantial capital and a team; the alternative is to isolate the few parts that are truly non-substitutable.
The piece briefly notes that AI could make filmmaking much "lighter." In the author’s view, once the creative idea, administrative approval, and distribution remain, the production stage that used to be the heaviest part may be compressed sharply. Still, the article says its main topic is not AI itself. The claim is that the spirit of "only send the brain" existed long before the current AI wave and has already shaped commercial history.
Nike and ARM: removing the heavy layers
The article calls "only send the brain" a form of reverse engineering. A standard method starts by listing every condition needed to complete a task. This method works backward and asks which conditions can be removed one by one while preserving the core objective.
The first corporate example is Nike. The article says Nike was founded in 1972. At the time, some apparel brands had already outsourced parts of their manufacturing, but usually only low-margin orders or overflow production. Nike, by contrast, is described as the first shoe company with no factories at all. Its view, as quoted in the piece, was straightforward: "We are not good at manufacturing. We are good at telling others what should be made."
According to the article, that model feels natural today, but it faced heavy skepticism at the time. Nike itself tried owning and operating factories in the late 1970s, then concluded that design and marketing were enough and that it did not need factories.
The article then moves to ARM. If Nike stopped making shoes, ARM went a step farther: it did not manufacture chips and did not even design specific chip products. Instead, it designed licensable processor architectures and CPU core IP. The author contrasts that with NVIDIA. In the article’s framing, NVIDIA designs a generation of GPUs and ends up with that specific GPU, while ARM can license one design to many companies, embed it across countless products, and keep collecting fees.
In the author’s metaphor, ARM has moved beyond "only send the brain" and toward something even more abstract, exporting rules and structures that can be replicated at scale. The article sums up the trajectory this way: from Nike to ARM, the concrete product gives way to visual design, and visual design gives way to abstract rules. In the AI coding era, the author writes, many complex and ambiguous tasks eventually become lines of code.
That leads to one of the article’s central formulations: ordinary things are done with weight; unusual things are done by becoming light.
IKEA: moving the final step to the customer
The article then makes a clarification. Its purpose is not to argue that "light" business models are always better than "heavy" ones. The author explicitly cites Taiwan Semiconductor Manufacturing Co. as a counterexample: in the chip supply chain, TSMC does no design work, stays focused on the heavy manufacturing layer, and has still become the least replaceable company in that system.
The real question, the piece says, is what a company does when it has very few resources and an almost impossible task. In that setting, it can only choose to become lighter.
IKEA is the next example. In 1950s Sweden, the article says, buying furniture was cumbersome. Customers would go into a furniture store, choose a table, then return home and wait for the store to deliver it from the warehouse. For IKEA, which was just starting its mail-order furniture business, the problem was more severe. Furniture took up warehouse space before sale, occupied truck capacity after sale, and was easy to damage in transit.
The turning point came in 1956. According to the article, IKEA designer Gillis Lundgren tried to fit a three-legged table called LÖVET into his car and found that it would not fit. He then did something that looked absurd at the time but later changed the furniture business: he removed the legs. Once the legs were detached, the table became flat enough to transport.
The company, the article says, suddenly recognized that there was no need to deliver a fully assembled table to the customer’s home. If the legs could be detached, the factory could produce components only, place all of them in a thin carton, and let the customer assemble the product at home.
From that point on, IKEA moved the final "assembly" step to the customer. The article adds that this changed product design itself. If customers had to assemble furniture on their own, then products needed to be designed around that fact from the start. Packaging was no longer something considered after the product was finished; fitting the product into a compact, flat package became part of the design problem.
Later, the article says, IKEA turned customer assembly into a profit point rather than a cost shifted to consumers. The pitch was simple: take the furniture home immediately at a lower price, instead of waiting several days for delivery.
The author treats this as a new definition of outsourcing: getting the customer to participate voluntarily in the completion of the product. In that reading, Wade in The Three-Body Problem assumes the Trisolarans can make the brain "live" again, while IKEA assumes the customer can complete the final step. Neither side delivers a finished whole; both, in different ways, "only send the brain."
Visa: owning almost nothing but the standard
The fourth section turns to Visa. The article says that in 1969, BankAmericard, Visa’s predecessor, fell into a serious operating crisis, and Dee Hock was appointed to deal with the situation, with abandonment still on the table if a solution could not be found.
Instead of asking how to issue more cards, Hock, in the article’s account, asked a deeper question: the efficiency of a credit-card business depends on the number of consumers and merchants using it. If Visa tried to open banks on its own, how many could it really build, and how much could it outgrow rivals?
The piece says banks around the world were independent competitors. Trying to convert all of them into branches under one organization was not realistic. Hock’s answer was the opposite approach: do not control the banks. Let them compete with one another, but require them to follow one set of rules written by Visa. Any bank willing to join the network could become part of it.
The author describes the result as one of the strangest commercial organizations in history. What does Visa actually own? Not the consumers, who belong to banks. Not the credit-card accounts, which also belong to banks. Not the goods, which belong to merchants. It does not even directly control the money, which only moves between banks. The article says the only thing Visa truly owns is a standard: rules, a network, a settlement system, the technology that links those pieces, and the trust of all participants.
From that perspective, the article argues, many top companies may "sell standards," but they usually attach those standards to products in order to increase value. Visa, by contrast, is presented as a company that sells only the standard. It tries to own as little as possible while controlling the most important connections.
That is why Visa ultimately became a network jointly participated in by member financial institutions rather than a simple centralized corporation. The article links this to what later came to be called a "chaordic" organization, one that combines competition and cooperation. It also says Hock’s original question remains relevant: if nothing in the world were constrained, what kind of organization would best fit tasks with very high demands for innovation?
From Visa to AI agents: chaordic design as a practical issue
The final section focuses on the possible relevance of chaordic organization in the AI era. The article says Hock believed, while designing Visa, that industrial-era organizations tended to use mechanical structures built around clear hierarchies, concentrated power, and command-and-control methods. Complex systems, however, do not necessarily need to operate that way. They can combine the stability of order with the creativity of chaos.
After extended thought and practice, Hock introduced the word "Chaordic" in his 1999 book Birth of the Chaordic Age, combining chaos and order.
The article notes, though, that the idea was ahead of its time. Visa could write the rules, but banks still had to send people to carry them out. Earlier chaordic organizations could reduce ownership, yet they still faced high coordination costs.
In the author’s view, AI agents may change exactly that point by turning execution into a capability that can be called on demand. If so, Hock’s chaordic thinking would move from a marginal management theory to a concrete organizational design question for the AI era.
That would also alter the role of the boss. The article says the industrial-era leader’s main skill was organization: organizing talent, resources, and capital. Chaordic theory puts more weight on rule design. Under AI conditions, the article pushes that one step further, toward defining goals, principles, and boundaries. Humans become less like managers and more like constitution writers; AI becomes less like a conventional employee and more like an executor inside the organization, a temporary organ assembled when needed.
The piece also refers to a 2026 paper in Business Horizons. As summarized in the article, the paper argues that the next stage of AI may not simply involve adding more agents. It may instead position AI as an "organizational coordination layer" that reallocates decision rights, coordinates different agents, and embeds governance rules.
In that sense, the author argues, "only send the brain" is not merely a stopgap idea born from technological weakness. The company itself is described as a body humans once built for their own brains: factories are a body, employees are a body, offices are a body, supply chains are a body, and sales networks are also a body.
What "only send the brain" points to, in the article’s closing argument, is a future in which a commercial subject no longer needs to fully own that body. It can call capabilities temporarily, combine them, split them, borrow them, or let others complete them. The article ends by suggesting that the most important organizational shift from AI may not be the often-discussed replacement of employees, but a lower need for the organization itself.
The original article was sourced from the WeChat public account "lig0624" (ID: tongyipaocha) and written by 思想钢印.