How to Balance "Long-Termism" with "Present Action"? | 5Y View

Tong Ji

Senior Investment Manager, 5Y Capital

Every day, individuals and companies face countless choices and trade-offs. Behind each choice lies the possibility of gain and loss.

How can we make better choices? This article offers one approach: at every moment, invest yourself in doing the most important thing.

Of course, no theory can replace the passion in your heart. May you find and choose what you love, and go all in at every moment. Let's do this together :)

Source: Lonely Brain (Gudu Danao)

Author: Lao Yu

What exactly is long-termism?

No one seems to have explained it clearly.

George Soros qualifies as a long-termist when it comes to "survival" and "investment returns," yet he's also a classic "reversalist" who frequently shifts with the wind.

Li Lu, whom Charlie Munger regarded as "the smartest person in the room," is a devout believer in value investing and unquestionably a long-termist — yet he also says investors should be like golfers, playing "memoryless" shots.

It all seems a bit contradictory, a bit schizophrenic.

Long-termism is not only hard to define, but even harder to practice.

Once, a friend asked:

How do you reconcile "long-termism" with "present action"?

Long-termism pursues constancy, while action requires adaptability. How do we reconcile the two?

This article will use Steve Jobs's most important decision-making principle as an example, and a unique "dual-track model," to answer this question.

This "dual-track model," in plain language, is: Always do the most important thing for the rest of your life.

Although Jobs's decision-making principle may sound a bit like chicken soup, the underlying principles are actually identical to those of AI.

As I've said: every master is a human AlphaGo.

Its greatest value lies in offering a "system" that anyone can practice, achieving both:

The ability to adhere to long-termism while also playing the best move in the present moment.

To summarize:

1. Long-termism refers to a person or organization's long-term evaluation system — vision, values, ultimate goals.

For AlphaGo, its evaluation system is the probability of winning at the endgame.

This naturally creates a sense of the bigger picture.

2. The present move means using this evaluation system to identify the move with the highest probability of winning at the endgame.

In Go, due to computational limitations and the human habit of "only looking at what's in front of us," we confuse local (short-term) optimality with global (long-term) optimality.

AI only cares about the probability of winning at the endgame. Every move resets to zero, recalculating the probabilities from scratch.

Why introduce probability?

Because even AI cannot exhaust all variations to precisely calculate the endgame win rate, so it simply uses probability assessment to choose the relatively better move.

3. These two points are the two "eyes" of long-termism.

4. The so-called "memoryless shot" is a Markov decision process.

A Markov process (Markov chain) is a memoryless stochastic process. While it appears to only care about the "current state," its purpose is actually to maximize expected global returns.

5. This explains why long-termist masters appear "fickle."

Jobs's decision-making closely resembled AI playing Go — sometimes appearing utterly unpredictable, suddenly abandoning a local position to play elsewhere, discarding when necessary without a trace of hesitation.

6. Long-termism is not simply "persistence" or "continuity."

A person's continuity is evaluated based on the continuity of their goals, not the continuity of their short-term behavior. Though the two often appear consistent.

7. Long-termism is also a Bayesian updating process.

Decision-makers pursue high-probability reliability, not absolute reliability, and this probability continuously optimizes over time.

Long-termism as a Bayesian updating process is both advancement and evolution.

The essence of long-termism is self-growth.

8. Long-termism is faith in truth-seeking, while for the "present move," one dares to adjust one's beliefs at any time.

9. Only thus can one achieve the compounding of time, space, capital, and self in an uncertain world.

10. Many believe "long-termism" is about perseverance — it's actually about insight and computational power.

01

Life has multiple turning points of different significance. After several unreliable relationships, the playboy Steve Jobs was finally tamed. By then, he had been ousted from Apple's board for five years, with six more years before he would be invited back to revive the company.

In 1990, Jobs gave a lecture at Stanford University. Laurene Powell, a new graduate student at the business school, was dragged along by a classmate. Arriving late with no seats available, she walked with her friend to the front row and sat in two reserved guest seats.

When Jobs arrived, he was guided to the seat beside her. He was drawn to Laurene on his right and chatted with her for a few moments.

Jobs later recalled: "I was in the parking lot, with the key in the ignition. I asked myself, 'If this were the last day of my life, would I rather attend a business meeting or spend it with this woman?' I ran across the parking lot and asked if she'd have dinner with me. She said yes. We walked into town together, and that was the beginning of our life together."

People later said Laurene gave the world a better Steve Jobs.

02

In this romantic legend, Jobs essentially conducted a thought experiment:

He compressed the rest of his life into a single day, enabling more focused reflection on what mattered most in the present moment.

Immediate optimality or long-term optimality? Local optimality or global optimality?

This may be one of the greatest differences between extraordinary people and ordinary ones.

Chess grandmaster Garry Kasparov said:

"The strategist sets a distant goal, formulates a strategy to achieve it, and then works backward to develop concrete measures. He establishes intermediate goals necessary for reaching the distant goal.

When a grandmaster plays, he doesn't rely on simply sifting through thousands of response schemes. Instead, he first determines a position he wants to reach in 10-15 moves. He evaluates all possibilities, sets a goal, and then advances toward it step by step.

Yet the human brain was not naturally designed for long-term thinking.

Our brain was cobbled together through漫长的 evolutionary patches, burdened with too much baggage from ancient and jungle times.

The prefrontal cortex governs human reasoning, planning, and decision-making — it freed us from the fate of running at the sight of tigers and gave us the capacity for long-term thinking.

03

That AI defeated the world's top Go players through long-termist "strategy" rather than locally precise "tactics" was surprising.

Demis Hassabis, father of AlphaGo, believed humanity had made a grave mistake about Go over 3,000 years. Go has the saying "golden corner, silver side, straw belly" — traditionally, players believed the third and fourth lines held greater value in the opening, while the center seemed too ethereal.

But in Game 2 against Lee Sedol, AlphaGo played on the fifth line with its 37th move, advancing into the center. Compared to the fourth line, this line sits closer to the central region. This may mean that for millennia, humans had underestimated the center's importance.

To understand this, let me briefly recap how AlphaGo works:

1. Policy network: Imitates human "intuition" to identify the 5-10 best possible moves for the current position;

2. Value network: Evaluates the win probability of these candidate moves. What win probability? The probability of winning at the endgame.

3. Reinforcement learning: AI's "intuition" isn't accurate at first, and its win probability estimates may be imprecise due to computational depth limitations. But through reinforcement learning and massive training, it continuously evolves to surpass humans.

The key point here:

No matter how long a game, AlphaGo judges every move by one criterion: the probability of winning at the end.

Human self-learning and evolution closely resemble artificial intelligence.

As Hassabis said:

From a biological perspective, animals and humans — the human brain is controlled by dopamine, executing reinforcement learning behaviors. Therefore, whether from a mathematical or biological perspective, reinforcement learning is an effective tool for solving AI problems.

04

Impulse is the devil, yet impulse also enables the "autopilot" of earthly life. This is the wonder of nature's algorithms — the Creator achieved this largely through dopamine.

Dopamine is a chemical that helps cells transmit impulses, a type of neurotransmitter.

This transmitter is primarily responsible for the brain's desire, sensation, and conveying messages of excitement and pleasure.

In short, dopamine governs instant gratification, immediate pleasure, seizing opportunities, and fleeing threats.

Later, humans fortunately developed the "prefrontal cortex," acquiring the concept of "future" and becoming willing to suppress present impulses for future goals.

How does the prefrontal cortex control dopamine?

The prefrontal cortex acts like an information hub in the brain. When multiple dopamine-related signals arise simultaneously, they converge at the prefrontal cortex, which — like a jury chair — selects the "most beautiful" signal.

What is the standard of beauty? The prefrontal cortex's "imagination" of the future. If the prefrontal cortex deems an impulse unfavorable to the future, it suppresses that dopamine signal.

One can imagine that if dopamine-related activity becomes excessively active, or if the prefrontal cortex cannot form a clear definition of the future and execute it firmly, then dopamine will prevail.

05

You might say: everyone knows long-termism is good, easier said than done. Why must success come at the cost of being "anti-human"?

Moreover, long-termism seems distant, immediate benefits are right before our eyes, and the middle ground is模糊 — a bird in the hand is worth two in the bush.

And if you can't even afford food, what long-termism is there to speak of?

Kasparov once played a move in actual competition with calculation depth of 18 moves, stunning the chess world.

I recall once when Kato Masao launched a direct attack, 50 moves without a single error, strangling the opponent's large group — truly admirable.

Yet even so, the vast majority of the time, even the greatest players have limited calculation depth, especially when facing an equally matched opponent across the board.

Artificial intelligence is no exception — it must stop calculating at appropriate times.

But even so, its evaluation system remains consistent: play the move that maximizes the probability of winning the entire game.

Planning without action is armchair strategy; action without planning is self-destruction.

Emerson said: "A man who knows how to do something will never be unemployed. A man who knows why he does something will always be his own master."

Like great rivers, when blocked, they detour — yet always flow toward the sea.

There's no evidence that Jobs possessed exceptionally high IQ. On the contrary, his personality flaws gave him a kind of resolute "judgment," like a powerful prefrontal cortex.

Though he himself seemed more like a dopamine-driven person.

Throughout Apple's history, those great product inventions and brilliant business concepts were essentially conceived by the brilliant people Jobs assembled.

Jobs's algorithm was simple: I only want the best. Thus he made decisive decisions again and again — ruthless enough.

Tartakower said: "Tactics is knowing what to do when there's something to do. Strategy is knowing what to do when there's nothing to do."

In the information wave beginning in the 1970s, Jobs's偏执 constituted a powerful form of leadership.

06

However, long-term thinking easily breeds "perfectionism" — disdaining small tasks while the grand vision remains distant, drifting aimlessly.

The solution: Focus on doing the most important thing right now.

One of Jobs's biographies noted a particular trait:

He would become intensely focused on something for a period, then suddenly shift attention to something else. At work, he would focus on what he wanted to do when he wanted to do it, becoming unresponsive to everything else, entirely regardless of how hard others tried to engage him.

This was the other side of Jobs's personal algorithm: only wanting the best, only caring about doing one thing before him to the utmost.

Almost all extraordinary people possess this talent for instantaneous immersion. Cybernetics founder Norbert Wiener once returned home from work without even recognizing his daughter waiting in the hallway.

"Steve would go to extremes, sometimes highly focused as if she were the center of the universe, and at other times showing a cold distance, focused on work," Smith said.

"He had the ability to focus like a laser. When his light shone on you, you basked in his attention. But when his light shifted to other concerns, you felt very, very dark. This greatly confused Laurene."

Chess player José Raúl Capablanca said: "I see only one move ahead, but it is always the correct one."

07

Let's return to dopamine and the prefrontal cortex.

A Harvard University study tracked over 700 people throughout their lives, aiming to discover what determines whether a person lives happily.

The ultimate conclusion:

Only good social relationships — including harmonious, warm family, friendship, and work relationships — bring people happiness and joy.

No one ever regretted having too little sex or bungee jumping in their lifetime. A palliative care nurse surveyed patients in their final days, revealing life's most common regrets. At the top, especially for men: "I wish I hadn't worked so hard."

Bronnie Ware, an Australian nurse who cared for patients with only 12 weeks to live, recorded their deathbed epiphanies in her book The Top Five Regrets of the Dying.

The five regrets she identified:

1. I wish I'd had the courage to live a life true to myself, not the life others expected of me.

2. I wish I hadn't worked so hard. (This was the top regret for men.)

3. I wish I'd had the courage to express my feelings.

4. I wish I had stayed in touch with my friends.

5. I wish that I had let myself be happier.

In that moment of deciding to run through the parking lot after Laurene, Jobs's not-exceptionally-brilliant brain employed his characteristic simple algorithm: If I compressed the rest of my life into one final day, what would I choose?

When Jeff Bezos decided to leave Wall Street to found Amazon, he too used the "regret minimization framework."

Only when we recognize what matters most in life can we send the clearest instructions to our prefrontal cortex, enabling us to make decisions aligned with long-term interests amid the rising and falling stimulation of dopamine.

In her Harvard commencement address, Angela Merkel offered heartfelt advice to thousands of graduates:

"Ask yourself again and again: Am I doing this because it is the right thing to do, or simply because it is something I can do?"

Often, we buy things not because we truly need them, but simply because they're on sale;

We choose to do something merely because it offers short-term satisfaction, willingly sacrificing long-term interests;

We make important decisions not because they're right, but to please certain people.

To build a powerful command center in your prefrontal cortex — one that, like AI, has only one evaluation criterion: maximizing the probability of winning this game of life.

A person's life is, after all, just a game of Go.

You need not be exceptionally intelligent or wealthy — simply do as Jobs did: Always do the most important thing for the rest of your life.

08

Mark Zuckerberg asks himself this question every day:

"Am I doing the most important thing I could be doing?" Only with an affirmative answer does he feel at ease, that his energy and time are not wasted.

Like Jobs, Zuckerberg also excels at greedy algorithms.

A greedy algorithm is one that, at each step, makes the locally optimal choice — the best or most advantageous option in the current state — hoping to find a global optimum.

Greedy algorithms are particularly effective for problems with optimal substructure. Optimal substructure means local optimal solutions can determine the global optimal solution. Simply put, problems that can be decomposed into subproblems, where subproblem optima can be recursively extended to the final solution.

Thus, you need a decomposition system to connect your ideals and reality.

Interrogate yourself: Am I doing the most important thing I could be doing?

This is a process of continuous self-refreshing.

The big picture and the local view, persistence and change, long-term and immediate — because of the world's uncertainty and our brain's computational limitations, this is forever an unsolvable problem, forever a process from模糊 to precise.

In fact, extraordinary people are fickle — or rather, they dare to change.

Perhaps the most important thing hasn't changed, but upon evaluation, the present matter no longer aligns with the future's most important thing.

Or perhaps, discovering that what was previously deemed most important was wrong, they bravely cut losses and resolutely stop the bleeding.

Always doing the most important thing for the rest of your life helps you continuously "refresh" yourself.

As Bill Gates wrote in the foreword to Microsoft CEO Satya Nadella's biography Hit Refresh: It (refreshing) doesn't clear everything and start over — it actually keeps some things and replaces others.

09

Long before meeting Laurene, Jobs already had a daughter.

In 1978, Lisa was born. The 23-year-old Jobs refused to acknowledge her as his daughter. Apple had already gone public, and only under court order did Jobs begin contributing partial child support.

When Lisa was nine, Jobs finally acknowledged her, yet remained bitterly harsh — so much so that Lisa began seeing a psychologist at age ten.

In his life's final countdown, this was no longer a genius's thought experiment but truly the final moments. Jobs finally apologized to Lisa — for not being present, forgetting her birthdays, not responding to messages and calls — repeating over and over: "I owe you one."

He ultimately left Lisa the same inheritance as his other children. He said: "If only I could turn back time, I wish I had been a better father."

A friend in Vancouver attended her child's high school entrance parent-teacher conference. Expecting something like our gaokao mobilization rally, she instead heard the principal emphasize only one point:

"Your children are about to become adults. These years will be your most precious time together — cherish it."

In 2018 in Hong Kong, I met a former acquaintance from the real estate industry who kindly calculated how much money I might have lost over the past decade due to moving abroad.

"Do you regret it?" he asked.

Frankly, I haven't had a single second of regret. Because I didn't miss my children's growing years. Even if I wanted to earn more money, that could come sooner or later — but children's growth is irreversible.

I'm grateful that I chose, at that time, to do the most important thing for the rest of my life.

5Y Capital (formerly Morningside Venture Capital) currently manages approximately US$3 billion across USD and RMB dual-currency funds. We believe that if the crazy person others see in you begins to be believed in, the world will be a better place.

BEIJING · SHANGHAI · SHENZHEN · HONG KONG

WWW.5YCAP.COM