“Novices stare at their two hole cards. Veterans look up and read the whole table.”
That line sets the tone for “The Power of the Board,” a long essay published by MarsBit and credited to Lao Yu of the WeChat account Lonely Brain. The piece uses Texas Hold’em to think through a much larger question: how people should understand the relationship between what they personally possess and what the world around them already provides.
Lao Yu calls those external conditions “community cards.” A small home may still sit near a park worth walking in. A modest bookshelf may still be backed by a library. A person may not know how to do a task alone, yet still have access to tools, tutorials, and people willing to help. In his framing, those are all part of the hand.
The essay was prompted in part by a recent article the author says he read, one that argued AI had become so powerful that life itself felt empty by comparison. He rejects that idea outright. AI, he writes, may indeed be the strongest community card of this era, but that does not make human life void. He compares it instead to sunlight, air, clean water, and the trees on a street corner: something powerful, widely available, and capable of improving ordinary life.
From there, the article lays out two basic ways to think about “community cards.” First, even people who are not trying to compete still benefit from them. Second, if someone does want to build something, the mistake is to compare two private cards against five community cards as if they were rivals. In poker, the board is not the enemy. It is what helps complete the hand.
Lao Yu then asks three questions that run through the rest of the piece: Which free, everyday resources around us have we ignored for years? When should a person enter a new game and combine with the opportunities on the table, and when should they fold? And when large models turn countless old crafts into zero-threshold public resources, what private cards remain truly irreplaceable?
Poker math as a way to think about life
The essay begins with the mechanics of Hold’em. In a 52-card deck, each player starts with two private cards. That creates 1,326 possible combinations, or 169 categories if grouped by rank and suitedness. The five community cards that appear on the table have 2,598,960 possible combinations. If a player’s two hole cards are excluded, the remaining 50 cards can still produce 2,118,760 possible boards.
Players choose the best five-card hand out of seven cards in total. Pocket aces are the strongest starting hand, and the chance of being dealt them is about 0.45%. If an opponent holds random cards and everyone reaches showdown, AA wins about 85% of the time heads-up. At a nine-handed table, that falls to roughly 35%, including split pots.
Lao Yu uses that structure to make a broader point. A person’s background, health, experience, and savings matter, but so do the era, city, institutions, and technologies around them. Beginners focus on the two cards in their hand. Stronger players ask what those cards can become once the board is revealed.
He then turns to a pure “play the board” scenario. If the table shows A, K, Q, J, and 10 of the same suit, every player still in the hand can use the board’s royal flush and split the pot. In that case, pocket aces and the worst unsuited 7-2 are functionally the same. For Lao Yu, that is a useful analogy for moments when progress turns what used to be elite advantages into ordinary infrastructure. Some opportunities are so large that once you are seated at the table, the outcome is not especially sensitive to your own cleverness.
Still, the same board card can help one player and hurt another. A new card may complete one person’s straight while making another player’s top pair worthless. New technology works the same way, he argues. It can save one person hours of repetitive work while making another person’s livelihood cheaper. The real question is not whether a new card is strong in the abstract, but what it combines with in your own hand.
Drawing hands, adjacent possibilities, and engineering preparation
One of the essay’s early sections focuses on “drawing.” If a player holds two hearts and the flop brings two more hearts, the flush is not complete yet. But among the 47 unseen cards, nine hearts remain that can complete it. In poker, those future cards are called outs.
Lao Yu links that idea to Stuart Kauffman’s concept of “the adjacent possible,” and to Steven Johnson’s example in Where Good Ideas Come From. Gutenberg’s development of movable metal type in mid-15th century Europe was not a sudden miracle, the essay says. It came from combining and adapting existing technologies such as screw presses, metalworking, oil-based ink, and paper into a workable printing system.
The lesson is practical. A project that feels one step short of viability may not need to be abandoned. It may simply be waiting for one or two external conditions to arrive: cheaper compute, a missing open-source component, or some other enabling piece. The goal is to prepare until you are one card away from completion. Then, when the right card appears, you can actually catch it.
Lao Yu describes that as a powerful engineering mindset and says he sees it in The Martian, where the protagonist survives by solving one adjacent problem after another.
Borrowed scenery, property, and public goods
The article then leaves the card table and moves into daily life. Chinese classical gardens and Japanese dry landscape gardens, Lao Yu writes, offer a model for using external resources well. Ji Cheng’s 1631 work Yuanye devoted its final volume to “borrowed scenery,” calling it central to garden design and dividing it into several forms, including distant, neighboring, upward, downward, and seasonal borrowing.
He points to Sogenchi Pond at Tenryu-ji in Kyoto, which borrows Arashiyama and Kameyama from outside the temple grounds. He also mentions Entsu-ji, where gardeners did not plant large trees inside the compound but instead used a low hedge to block nearby clutter and frame Mount Hiei several kilometers away. The point is not aesthetic trivia. It is that people can organize their lives by asking what already exists nearby instead of trying to own every function outright.
That same logic appears in his reading of Thoreau. In Walden, published in 1854, Thoreau wrote that he nearly bought the Hollowell farm. The deal fell through before it was signed, yet he felt he had already received its best woods and views for free. Even the woodland where he built his cabin at Walden Pond belonged to Emerson. Ownership has value, Lao Yu says, but so does use and appreciation. A landscape can restore a person whether or not their name is on the deed.
Public infrastructure, in his view, exists partly to compensate for what individuals do not privately hold. He cites the design of New York’s Central Park by Frederick Law Olmsted and Calvert Vaux in 1858, at a time when wealthy New Yorkers were building private estates in the Hudson Valley. The project employed large numbers of workers, moved millions of cubic yards of earth and stone, and eventually created the 843-acre park known today. Meadows, paths, and lakes no longer had to exist only behind private walls. He places Hangzhou’s West Lake in the same category of civic intelligence.
That leads to a simple test for evaluating where one lives: not just how many square meters are inside the apartment, but what lies around the corner. Parks, hospitals, public transit, and libraries all shape daily life.
The essay also revisits the economic idea of free-riding. Pure public goods are usually defined by two traits: non-excludability and non-rivalry. If a wealthy homeowner spends millions landscaping an exterior wall, or a city spends heavily building a riverside walkway, passersby can enjoy the visual and environmental benefits without paying. In business and governance, free-riding is often treated as a problem. In personal life, Lao Yu argues, legally and frequently consuming the positive externalities of public goods is simply smart living.
Modern life contains advantages kings could not buy
Another section of the essay argues that many ordinary “community cards” are so familiar that people no longer notice them. Lao Yu mentions the story that the Roman general Lucullus had snow brought down from the Apennines to eat chilled desserts in summer. He also points to the death of U.S. President Calvin Coolidge’s 16-year-old son in 1924 after a blister became infected and led to sepsis. Penicillin would not be discovered by Alexander Fleming until 1928, and mass production would not arrive until the 1940s.
He then cites economist Don Boudreaux’s thought experiment: would you trade places with Rockefeller in 1916 if you also had to accept the technology of that era? Modern medicine, long-distance communication, and audio-visual services available to ordinary people today were not things even the rich could reliably buy then.
The same logic appears in his discussion of Isaac Newton’s famous line about standing on the shoulders of giants, written in a 1676 letter to Robert Hooke and traceable further back to Bernard of Chartres in the 12th century. No one today needs to reinvent writing, calculus, or printing. Admitting that your private cards are limited is not shameful if you know how to climb onto the knowledge and tools built by others.
Warren Buffett appears in the essay as another example. Lao Yu recalls Buffett’s “ovarian lottery” remarks at Berkshire Hathaway’s 1997 shareholder meeting, where Buffett stressed the role of birthplace, talent, and social environment in shaping outcomes. He also cites Buffett’s 2018 annual letter, which said that the $114.75 Buffett first invested in stocks in March 1942 would have grown to $606,811 before taxes by Jan. 31, 2019 if it had simply tracked the S&P 500 with dividends reinvested and no fees, a gain of more than 5,200 times. The point is not to deny individual choice, but to place it beside the era and system that made those returns possible.
AI is already on the table
AI sits at the center of the essay. Lao Yu quotes William Gibson’s line from a late-1990s interview: “The future is already here — it’s just not evenly distributed.” He notes that Gibson had already used the word “cyberspace” in the 1982 short story Burning Chrome, and that when Neuromancer was published in 1984, Gibson still did not own a computer and wrote the book on a manual portable typewriter.
That becomes a direct analogy for the present. AI, Lao Yu writes, is already face-up in the middle of the table. The unevenness lies not in whether the card exists, but in how quickly different people connect to it. Those who look up earlier can start making new combinations even if their own private tools are ordinary.
He extends the metaphor through sailing. Seneca wrote in Letter 71 of the Moral Letters that if a person does not know which port they are sailing toward, no wind is favorable. Early square-rigged ships struggled to sail against the wind, but later sailors learned to use triangular sails and tacking. Even with wind coming from 45 degrees ahead, adjusting the sail angle could generate forward motion through Bernoulli lift. In engineering terms, Lao Yu says, that means accepting the input and changing the transfer function. People cannot choose the era’s wind, but they can still adjust the sail.
Open doors, third spaces, and access over ownership
The essay also argues that people need channels for outside signals. Lao Yu cites Richard Hamming’s 1986 talk “You and Your Research,” where the Turing Award winner observed that scientists at Bell Labs who kept their doors shut often produced a lot in the short run but drifted away from the most important problems over a decade. Those who worked with doors open and were interrupted by hallway traffic were more likely to hear which questions were becoming important. Hamming, the essay notes, also admitted he could not prove the causality statistically.
That observation leads into a discussion of “third places,” a term from Ray Oldenburg’s 1989 book The Great Good Place. Home is the first place, work the second, and cafes, libraries, parks, and community courts the third. Lao Yu mentions Lloyd’s Coffee House in late-17th-century London, where captains and merchants could buy coffee for a penny and read shipping and disaster notices posted on the wall. He also notes that J.K. Rowling’s official site records that she wrote Harry Potter in Edinburgh cafes while her daughter slept beside her in a stroller.
His point is that the first space depends heavily on housing budgets, while the third depends on how well a person can use public and commercial resources outside the home. He adds a personal note: his daughter likes to go to a cafe far from home to write, and he tries to drive her there without asking why she does not simply use her own study. She has made good friends there.
On the economics of use, Lao Yu cites Kevin Kelly’s 2016 book The Inevitable, which listed access over ownership as a major technological trend. Compute, software, books, and even specialized equipment can increasingly be used like utilities. The relevant comparison is total cost: buying and maintaining something, borrowing it from a public institution, or renting it by the month.
Knowing what you want matters more than what you hold
One section of the essay turns to entrepreneurship. Lao Yu recounts a story from Meituan’s internal incubation of Maoyan Movie. When the team leader complained about lacking money, people, and resources, Wang Xing replied that real entrepreneurial spirit means knowing the goal and thinking about how to reach it regardless of what you currently possess. Lao Yu also quotes Wang Xing’s three recurring questions: What do you have? What do you want? What are you willing to give up?
He pairs that with Harvard Business School professor Howard Stevenson’s definition of entrepreneurship as the pursuit of opportunity beyond resources currently controlled. The table is full of external cards every day, Lao Yu writes. If you do not know what hand you are trying to make, even a strong board looks like noise. If you know exactly what you need, you can identify the right card even with a weak starting hand.
The essay then moves into Daniel Kahneman’s “focusing illusion,” citing a 1998 study by Schkade and Kahneman comparing students in the U.S. Midwest and Southern California. Both groups overestimated how much happier Californians must be because of the weather, yet self-reported overall life satisfaction was nearly the same. When people fixate on one variable, Lao Yu says, it occupies more space in the mind than it does in life. A cramped apartment or one persistent problem can dominate attention while several stabilizing cards on the table go unnoticed.
James J. Gibson’s concept of affordance appears next. In The Ecological Approach to Visual Perception from 1979, Gibson described affordances as the action possibilities an environment offers. The same half-meter stone can be a seat for a tired traveler, cover for a hunter, or an obstacle for a child. Lao Yu applies that to ordinary settings: one park can be a running route, a place to teach children about plants, or a walking meeting room.
He then borrows from Richard Dawkins’ 1982 book The Extended Phenotype. A hermit crab does not grow a hard shell; it uses an empty one found on the beach. A beaver’s teeth and fur are only part of its private equipment. The dam it builds, and the lake created behind it, become part of its effective phenotype. For people, Lao Yu says, a well-organized knowledge base, a familiar sports field, and reliable collaborators all extend the self beyond the body.
That idea flows into the “extended mind” thesis of Andy Clark and David Chalmers from 1998. Their thought experiment about Otto’s notebook argues that an external memory aid can become part of cognition if it is used consistently and reliably. Lao Yu says AI tools and note systems work the same way. Installing ten apps on a phone means little. What matters is whether a person can pull up the right information immediately, verify it, and connect it to the next action.
Reuse, folding, and staged commitment
The article’s engineering sections are especially direct. Citing Synopsys’ 2024 Open Source Security and Risk Analysis report, Lao Yu notes that 96% of the 1,067 codebases analyzed contained open-source components, and 77% of the scanned source code and files came from open source. An engineer who refuses to use libraries and insists on writing an operating system and compiler from scratch may never ship usable software at all.
That becomes a general rule for work. Reuse mature frameworks when writing plans. Reuse proven checklists when organizing events. Save your limited energy for the 20% that actually requires unique judgment, while still having a human decide which module fits the task and whether the license terms are acceptable.
Another major theme is folding. At a full nine-handed Hold’em table, top professionals often have a VPIP, or voluntarily put money in pot rate, of only 15% to 22%. In other words, they fold roughly 80% of hands after a quick look at their cards or the flop. Amateur players often go broke not because they cannot play strong hands, but because they feel compelled to get involved in every deal.
Lao Yu reinforces that with Ted Williams’ 1971 book The Science of Hitting, where the baseball legend divided the strike zone into 77 baseball-sized cells. He estimated that swinging only at pitches in the sweet spot could produce a .400 batting average, while chasing low outside pitches would drop it to .230. Buffett has cited that chart repeatedly. Life and investing, Lao Yu says, are kinder than baseball. There is no strikeout rule forcing a swing.
He also stresses that the board is revealed in stages. The flop comes first, then the turn, then the river. Players can fold, call, bet, raise, or check after each stage. That structure becomes a way to think about uncertainty. Lao Yu cites MIT economist Stewart Myers, who introduced the idea of real options in 1977. A major commitment does not always need to be made all at once. Sometimes the rational move is to spend a small amount to buy the right to stay in the game and see more cards before deciding whether to commit or walk away.
He illustrates that with Stanley Druckenmiller. According to the essay, in the fall of 2022 one of Druckenmiller’s younger partners noticed that top engineering students at Stanford and MIT were moving from crypto toward AI. Druckenmiller, then nearing 70, still thought Nvidia was mainly a gaming graphics company, but he bought an initial position before fully understanding large models. A month later, ChatGPT arrived — the equivalent, in Lao Yu’s metaphor, of the key turn card — and Druckenmiller sharply increased the position in early 2023.
Position, pot odds, and bluffing off the board
Position matters in poker, and Lao Yu says it matters in technology adoption too. The player under the gun must act first with less information. The button acts last and sees everyone else’s response to the board. Professional players often enter pots from the button at two to three times the rate they do from early position.
He connects that to a 1993 study by Tellis and Golder covering 50 product categories and about 500 brands. In 36 comparable categories, market pioneers had a failure rate of about 47% and an average market share of about 10%. “Early leaders” that emerged during the early growth phase had a failure rate of about 8% and an average market share of about 28%. The first entrant is not always the winner. Lao Yu mentions Altair 8800, Archie, and SixDegrees as examples. Apple was not first in MP3 players, smartphones, or tablets. Tencent rarely insists on being first into a new track.
He then turns to pot odds. If a hand has only a 20% chance to win, a player focused only on raw probability may fold. But if the pot already contains 1,000 and the player needs to call only 100 with no further betting, the pot odds are 10-to-1. In that case, a 20% win probability makes the call positive expected value.
Lao Yu uses Alexander Fleming’s 1928 observation of mold killing staphylococci on a culture plate at St. Mary’s Hospital in London as an example of asymmetric upside. The cost of noticing the anomaly was tiny. The upside, once recognized, was enormous. In daily life, he says, buying a book, trying a new tool for half a day, or having coffee with someone from another field can all be high-pot-odds actions.
The essay also warns against “board bluffs.” In poker, a player may shove chips when an ace appears on the board, pretending to hold one as well. Lao Yu links that to Michael Spence’s signaling theory from 1973 and Amotz Zahavi’s handicap principle from 1975. A signal’s credibility depends on how hard it is to fake. Verifiable delivery and real operating results matter more than cheap talk. Whenever a dazzling public card appears, some people will use it to bluff and manufacture anxiety.
AI still leaves bottlenecks, and expertise still matters
One of the essay’s more technical sections invokes Amdahl’s law. In 1967, IBM System/360 architect Gene Amdahl argued that the serial portion of a system limits the total speedup available from parallel processing. Even if half the work becomes 10,000 times faster, if the other half does not change, the whole system can improve only so much.
Lao Yu maps that directly onto writing and decision-making today. AI and search are the massively accelerated parallel component. Human taste, fact-checking, and final judgment are the serial part that cannot be outsourced. If a draft appears instantly but the author still cannot decide what problem matters or which claims are trustworthy, the process remains stuck.
He then cites the 1990 paper on absorptive capacity by Wesley Cohen and Daniel Levinthal. A firm’s ability to recognize, assimilate, and use external knowledge depends heavily on the related knowledge it already possesses. Applied to AI, Lao Yu says, people with real domain understanding and real business problems know what to ask and can spot flaws in generated output much faster. The stronger the public card becomes, the more valuable private depth and taste become.
That leads to a distinction between absolute improvement and relative competition. When a new tool becomes as common as running water, almost everyone gets faster at old tasks. But if every competitor receives the same card, the tool alone does not guarantee an edge. One question is whether life gets better than before. Another is whether you become stronger than others. The first can improve dramatically through public technology. The second still depends on the quality of the private hand.
When generation becomes free, what remains scarce?
The final third of the essay focuses on what large models do to creative and knowledge work. Lao Yu begins with photography. After Louis Daguerre introduced the daguerreotype in 1839, painter Paul Delaroche reportedly exclaimed that painting was dead. Before photography, rendering a face accurately was a hard-earned private skill. Photography removed the middle layer and made likeness cheap.
He pairs that with Richard Sutton’s “The Bitter Lesson,” which argues that hand-built wrappers and domain tricks may look useful for a while, but end-to-end methods tend to crush them once general compute scales. Lao Yu says the same thing is happening in AI software. Jasper, once valued at $1.5 billion, and many wrapper companies built around one-click document reading or copy polishing looked impressive when large models first broke out. As the base models improved, many of those intermediary functions were absorbed.
For creators and white-collar workers, he says, the same reversal is underway. Expanding articles, translating materials, writing basic code, and producing standard reports used to be paid middle-layer work. Large models have turned much of that processing into a public utility.
He then uses another poker analogy. If the board contains four tens, everyone still in the hand has quads. The deciding factor becomes the kicker. In creator markets, Lao Yu argues, that kicker is no longer polished standard content. The music industry already showed the pattern. Before streaming, artists sold records and used concerts to promote them. Once perfect recordings became easy to access, the recording became the free promotional layer and the live, embodied, non-replicable performance became the scarce product.
Today, he writes, a large model plus deep search can produce a clean industry overview or reading note in ten minutes. People who made a living collecting information, rewriting it, and reselling it as processed knowledge now hold cards that overlap almost completely with the board. When passable synthetic content multiplies online, the scarce question becomes trust: who should a reader believe?
Lao Yu’s answer is blunt. AI has never lost real money in a bear market. It has never driven a daughter to a distant cafe late at night so she could write. It does not bear consequences for its advice. A real person’s lived experience, accumulated credibility, and social field are the private card that cannot be taken away.
Writers, non-writers, jagged intelligence, and enterprise moats
The essay cites a 2024 article by Paul Graham titled “Writes and Write-Nots.” Graham argued, in Lao Yu’s retelling, that society once forced professors, executives, and founders to write for themselves. Large models now make polished prose a public utility, splitting the world into a small minority who still write and a large majority who outsource expression. Lao Yu quotes Leslie Lamport: “If you’re thinking without writing, you only think you’re thinking.” The danger is not just laziness. It is the atrophy of the very process that turns vague intuition into clear judgment.
He also brings in Andrej Karpathy’s description of large models as possessing “jagged intelligence.” Biological intelligence evolved in the physical world and has relatively smooth boundaries. Large models, by contrast, are “ghosts summoned from internet text,” capable of solving Olympiad-level problems or restructuring 100,000 lines of code in seconds while also making absurd mistakes on simple real-world edge cases. That, for Lao Yu, is where human value remains clear: setting boundaries, spotting hallucinations, and hitting the brakes at critical moments.
The business sections follow the same logic. Lao Yu says Bret Taylor, OpenAI board chair and former co-CEO of Salesforce, argued while building Sierra that large models are undermining the software industry’s long-standing per-seat pricing model. If one AI agent can do the work of ten customer service reps, charging by employee account stops making sense. Taylor said at a Sequoia event that AI business models are shifting from selling software tools to charging for outcomes. Sierra’s own materials, Lao Yu notes, describe billable outcomes such as resolving inquiries, driving purchases, and preventing cancellations. If the issue is not resolved or must be escalated, there is often no charge, though the exact terms are agreed in advance and may be hybrid.
He then moves to Satya Nadella’s comments on the BG2 podcast in late 2024. Traditional enterprise software, Nadella argued, is often just a thin interface and business logic layer on top of CRUD databases. In an agentic era, AI may bypass the interface and read or write the underlying data directly, weakening the old concept of standalone SaaS.
That, Lao Yu says, helps explain why standard software vendors feel exposed while Palantir, long criticized in Silicon Valley as heavy, slow, and services-like, has surged in this AI cycle. Palantir sends forward-deployed engineers into Airbus factories, hospital wards, and military command rooms to do difficult integration work. They clean up bad data scattered across old systems, map sensor signals, permissions, and physical actions into executable digital twin rules, and build what Palantir calls an ontology layer. OpenAI, Claude, or open-source models may be smart, but they do not know which turbine blade is missing from a specific Airbus aircraft, which warehouse should release inventory, or whose approval is required. In that architecture, the foundation model becomes the replaceable public card. The ontology and execution layer built over years in the mud becomes the hard private hand.
Staying in the game
The essay closes by returning to the table itself. Even the best poker player in the world, Lao Yu writes, can bust out if they shove all their chips into one hand and lose. Once they are gone, no future royal flush matters. As long as a player remains seated, time keeps dealing.
That is the final argument of “The Power of the Board.” A person who looks only at their two private cards can start to see life as a sealed verdict: background cannot be changed, the past cannot be changed, and a weak hand feels like destiny. But once the view widens to the whole table, the situation looks different. The trees in a neighborhood park, the books in a library, the medicine and science accumulated across generations, the code in the open-source world, and AI itself are all cards already face-up on the table.
Lao Yu ends with a simple instruction: keep refining your two private cards, and use the five community cards on the table generously. What you truly have, he writes, is far greater than what you formally own.
The article was originally published on the WeChat account Lonely Brain under the byline Lao Yu.