How To Get Rich & Die Trying Pt 1: Yin and 양/兩

---

I. Systems Thinking

In Q1 of 1986, Larry Fink had his best quarter ever. $130 million in profit trading mortgage-backed securities. Three months later, he lost $100 million. Rock star to pariah in 90 days [1].

Fink conducted an immediate postmortem, determined to never be blindsided like that again. He concluded that the solution wasn't "better traders." It was "better systems." This determination laid the foundation for the first draft of BlackRock.

Fink officially founded BlackRock two years after this event. Its flagship product is Aladdin: an Asset, Liability and Debt and Derivative Investment Network. It runs thousands of randomized economic scenarios for every security, billions across all portfolios, all day, every day. The machine was built to observe the market and measure risk with surgical precision.

The first iteration of Aladdin worked. And then it kept working. Every time markets got more complex and asset managers got burnt (the dot-com bust, the 2008 crash) they experienced the same trauma Fink had in 1986. Each time, the losses got bigger. Each time, Fink sold the cure to his own disease. Financial institutions took the bait. Aladdin wasn't able to predict these crashes, but after you've lost a few hundred million, the pitch "this will never happen to you again" hits different. The machine BlackRock built with a $5 million loan now manages $21 trillion in assets [2]. 7-8% of ALL global financial assets runs through a system born from one man's worst quarter.

---

II. Agentic Thinking

The playbook Fink wrote is being repeated one layer up.

OpenAI is inking deals to pay private equity firms 17.5% returns plus early access to unreleased models [3]. In exchange, those firms deploy OpenAI's tools across their entire portfolio of companies. The deal is: we pay you to make every company you own dependent on our models. Every interaction funnels data back. Every dependency deepens the lock-in. Anthropic is running the same play, courting the same firms, and inking the same deals.

Both AI firms are making the bet that subsidized adoption today becomes structural dependency tomorrow. Guaranteed returns buy distribution. Distribution buys integration. And the pitch works because the edge is real enough to believe. A millisecond advantage in execution, a slightly better prediction, a system that never sleeps. That's more than an edge. In the world of investing, that's a money printer.

Trading firms are already rebuilding around this [4]. The old model (Citadel, Millennium) hired a thousand researchers. Each one generated ideas, tested them, and deployed what worked. Their edge was headcount. More brains, more ideas, more alpha. It scaled linearly, but it was capped.

The new model scales on itself*,* rather than headcount. It doesn't get tired, doesn't need bonuses, doesn't leave for a competitor. And unlike a thousand researchers working in parallel, the system is recursive; every point of alpha it discovers feeds back into itself, making the next discovery quicker and easier than the last. The cap disappears and the machine acquires more of the market.

If the machines are playing against each other at a scale and speed no human can match, what game is left for everyone else?

---

III. Memetic Thinking

Billions of dollars exchanging hands with the computational capacity that only billions of dollars can buy. Advanced server farms and AI GPU clusters that can see everything, price everything, and trade everything faster than you can blink. So how can someone on the outside have a competitive edge?

When OpenAI, Stripe, or SpaceX finally IPO, they'll debut at valuations that already reflect a decade of exponential growth - growth that was captured almost entirely by private investors who got in at a fraction of the price. By the time retail can buy a single share, the ceiling is the floor. The value they're handed is a $200 billion ticker that needs somewhere to go.

This problem isn't exclusive to markets. There's a bakery in NYC I loved called Magnolia Bakery. The owners built something real over years of effort. Venture capital bought it in 2021. Now it's a slop factory running on the fumes of what made it worth buying in the first place.

The bakery you liked got bought and gutted. The apartment you rented got flipped. The paths to value... building something real, buying something real, they've been strip-mined.
.
.
So what if you just stopped playing the game entirely?
.
.

Build-A-Bear went up 7,400% because its ticker is $BBW [5]. The retail crowd is optimizing for belonging, revenge, and the joke itself - which in a roundabout way can actually end up optimizing for returns; just not through any logic a referential algo could generate itself.

When the landscape is competing against a risk engine that runs billions of simulations before breakfast, GPU clusters that get exponentially smarter, and legislators who set the market rules and trade ahead of them, why wouldn't you just bet on Bryan Johnson's johnson on Polymarket [6]? Everything is made to feel like a coin flip in your world anyway. Why not feel like you have some skin in the game?

Meme-trading is nihilism as financial strategy. It's born from an era where the traditional paths to wealth seem to be walled off, and people are left opting for whatever get rich quick scheme is available. It doesn't follow the models because it was never playing the models' game.

The failure to predict retail chaos is just one example of a deeper problem: the models can't model the players who are modeling the models.

---

IV. Meta Thinking

George Soros built his trading career on the philosophical premise of reflexivity, as applied to markets. He understood that the act of predicting is what changes the thing being predicted. If enough participants believe a currency will fall then they sell it, which makes it fall, which confirms the belief, which causes more selling. The prediction doesn't forecast reality. It creates reality.

Advanced poker works the same way. It's a layer deeper than just playing your hand. You're playing your opponent's model of you. You know they think you're tight, so you bluff. You know they think you've been bluffing, so you value-bet. It's recursive: what do they think I think they think?

Markets are the same game at scale. They're pricing assets while also pricing other participants' models of the market.

When Morgan Stanley opened the ability for clients to invest in crypto four days after Bitcoin hit its cycle peak, it was inventory distribution being sold as democratization [7]. The institution had been positioned for months, maybe years. They needed to rotate out.

The account holder thinks: Morgan Stanley is offering me Bitcoin in my 401k. If they're offering it, it must be safe. This must be the mainstream moment. That's the buy signal for them, and the sell signal for the institution offering it. The chart is secondary. The psychology is primary.

This is the anti-Aladdin thesis. Aladdin processes thousands of quantitative risk factors daily, but it can't read the intent behind a headline. It can't parse the meta-game. It doesn't know that Morgan Stanley's timing is the signal, that the offer itself is the data point that matters most.

Now apply reflexivity back to Aladdin itself. When $21 trillion in assets runs through the same risk framework, the framework's outputs become inputs to the system it's measuring. Aladdin models the market. The market moves on Aladdin's models. After decades of this you have to ask which one is leading.

Does a concentrated number of institutions, using the same predictive system, make the future an algorithmically self-fulfilled prophecy?

This is a battle between predictive analytics and entropy. Yin and 양/兩. How much of the future is actually predictable? And how much of what looks predictable is just the system producing the outcomes it expects?

Larry Fink built Aladdin because he never wanted to be blindsided again. But a machine that manages 7-8% of global wealth doesn't eliminate the blindspot. It becomes the blindspot.

---

References
[1] Larry Fink: The pioneer who fought his way to the top
https://www.euromoney.com/article/27bjsstsqxhkmh1qi81a0/capital-markets/larry-fink-the-pioneer-who-fought-his-way-to-the-top/
[2] BlackRock’s Ambition: Become Inseparable From Asset Management
https://www.institutionalinvestor.com/article/2bstpo0h7569w26vanugw/corner-office/blackrocks-ambition-become-inseparable-from-asset-management
[3] Exclusive: OpenAI sweetens private equity pitch amid enterprise turf war with Anthropic, sources say
https://www.reuters.com/business/openai-sweetens-private-equity-pitch-amid-enterprise-turf-war-with-anthropic-2026-03-23/
[4] https://x.com/ethanrkho/status/2037184701069238615
[5] Why Build-A-Bear Has Been on an Nvidia-Like-Run
https://www.bloomberg.com/news/features/2025-11-12/build-a-bear-stock-price-has-been-on-a-nvidia-like-run
[6] https://polymarket.com/event/bryan-johnsons-average-nighttime-erection-2h-13m-in-december
[7] Morgan Stanley drops restrictions on which wealth clients can own crypto funds
https://www.cnbc.com/2025/10/10/morgan-stanley-drops-crypto-fund-restrictions-for-wealth-clients.html

---