#159 – Richard Craib: WallStreetBets, Numerai, and the Future of Stock Trading

Here is the translated investment research report based on the podcast transcript and your instructions.

At a Glance

This episode features Numerai founder Richard Craib, who discusses the WallStreetBets event, the power of decentralized coordination, and how Numerai aims to disrupt the traditional hedge fund industry through crowdsourced AI models and a cryptocurrency staking mechanism. Richard Craib argues that the key to WallStreetBets' success was its discovery of a "positive-sum game"—profiting from a short squeeze that forced short-selling hedge funds to buy back shares, benefiting the entire group. This is fundamentally different from a zero-sum game that merely creates a bubble.

The Miracle of Decentralized Coordination: The Mechanism and Risks of WallStreetBets

Richard Craib believes that WallStreetBets' success was not accidental, but the combined result of its unique incentive mechanism and a "positive-sum game" structure.

• Mechanism Breakdown: Craib points out that anonymous users achieved large-scale coordination without credible contracts by relying on a "reputation" system within the community. For example, the user "DeepFuckingValue" accumulated "Karma" (community prestige) by posting real trading screenshots and past successful predictions, making his calls credible. This formed a process of "information aggregation," where individual users observed these signals, judged that "something big was happening," and then joined in.
• Positive-Sum Game: Craib emphasizes that WallStreetBets' clever move was choosing stocks with high short interest (like GameStop). Unlike creating a "watermelon bubble" (a zero-sum game where prices eventually fall), a short squeeze is a positive-sum game. Because short sellers (hedge funds) face unlimited loss risk, they are forced to buy back shares to cover their positions as the price rises, which further pushes the price up, allowing the entire long-side group (Reddit users) to profit collectively.
• Risk and Uncertainty: Craib acknowledges the terrifying nature of this decentralized power. He contrasts it with the "Occupy Wall Street" movement, which he considers "sincere," while WallStreetBets has a "nihilistic" and "YOLO" flavor, making it more frightening. He suggests this power could be used to attack good people, and the moral voice within the group can easily be drowned out. He cites an example of a hedge fund manager who claimed he had the ability to make certain trades but chose not to because it would "mess things up," a concern the Reddit army clearly does not share.

Numerai's Disruption: Open-Source Data and the Staking Mechanism

Richard Craib argues that Numerai, through two major innovations—"open-source data" and "cryptocurrency staking"—solves the talent and data bottlenecks of traditional quantitative hedge funds and achieves perfect alignment of interests.

• Open-Source Data: Craib explains that Numerai takes stock data from expensive data providers (like PE ratios, one-year momentum, etc.) and obfuscates and anonymizes it into millions of rows of numbers between 0 and 1, then offers it for free to anyone globally. This allows talented individual data scientists who cannot afford the data to participate in building quantitative models.
• Staking Mechanism: Users do not need to upload model code; they only need to submit predictions and stake Numerai's cryptocurrency, NMR. If the model performs well, the user receives NMR rewards; if it performs poorly, their staked NMR is burned. Craib emphasizes that the burned NMR does not flow to Numerai, ensuring that the interests of both parties are perfectly aligned, rather than being a zero-sum game. This solves the "credible commitment" problem in anonymous collaboration.
• Extrapolation and Validation: Craib believes Numerai's ultimate goal is to "manage all the money in the world." Its path is to increase the data volume tenfold each year and, through the "Numerai Signals" feature, allow users to generate signals using their own datasets (e.g., sentiment data scraped from WallStreetBets), continuously absorbing "edge data" and "weird data." A key validation signal is: when the total amount staked by Numerai users exceeds Craib's own investment in the fund, he can sleep well, as this proves the system is strong and trustworthy enough.

The "Counter-Intuitive" Principle of Quantitative Investing and the Future of AI

Richard Craib points out that a key counter-intuitive principle for applying machine learning to financial data is to "reduce exposure to the strongest features," which is the exact opposite of practices in fields like computer vision.

• Mechanism Breakdown: Craib explains that financial data is a non-stationary time series. The strongest patterns from the past (e.g., "invest in tech stocks") may become ineffective or even disastrous in the future. Therefore, a good model does not maximize exposure to a particular feature but neutralizes or minimizes this exposure to improve the model's generalization ability and avoid overfitting to history.
• Extrapolation: Craib believes that before AGI destroys the world, a Narrow AI will first win all the money in the stock market. He is building Numerai to ensure that this "Narrow AI" uses their data. He cites a conversation with the CEO of Renaissance Technologies, who argued that AI has limited use in finance, but Craib believes this precisely proves the field's challenge, noting that AlphaGo defeated a human Go champion on the same day.

Position Moves

Ticker	Guest's Stance	Key Data
GameStop	Case Study (Bullish on short squeeze mechanism)	Short interest exceeded its own float; stock price doubled in a single day.
BlackBerry	Risk Warning (Targeted for short squeeze)	Stock price rose 100% in a single day.
Tesla	Hold & Observe (As a long position for Numerai)	Numerai did buy Tesla and believes it played a "minor" role in pushing the price.
Dogecoin	Neutral (Discussed as a meme phenomenon)	Its success is highly correlated with Elon Musk's tweets and the internet's "joke" propagation mechanism.

Judgments Worth Remembering

1. "Positive-Sum Game" is the Key to WallStreetBets' Success: Richard Craib argues that unlike a zero-sum game that creates a bubble, a short squeeze is a positive-sum game because short sellers are forced to buy back shares, allowing the entire long-side group to profit collectively.

2. The Counter-Intuitive Principle of Financial Data Modeling: Reduce Exposure to the Strongest Features: Richard Craib points out that in finance, models should not reinforce the strongest features (like "tech stocks") as in computer vision, but should neutralize or minimize this exposure to improve generalization ability.

3. Numerai's Staking Mechanism Solves the "Credible Commitment" Problem in Anonymous Collaboration: Richard Craib explains that by having users bet their cryptocurrency on their own predictions, with poor performance leading to token burning, this is more reliable than Reddit screenshots and Karma, achieving perfect alignment of interests.

4. "Before AGI destroys the world, a Narrow AI will win all the money in the stock market": Richard Craib believes financial prediction is one of the most likely complex domains for AI to conquer first, and Numerai's goal is to be the platform this "Narrow AI" uses.

5. Hedge Funds are Necessary, but There is a Fundamental Difference Between "Evil Shorting" and "Benign Shorting": Richard Craib distinguishes between two types of shorting: one is like Melvin Capital, heavily shorting a stock and trying to pressure the company through the media; the other is like Numerai, hedging market risk by holding hundreds of long and short positions, allowing it to more safely go long on companies like Tesla.

6. "Don't start a company unless you're ready to make it your life's work": Richard Craib advises young people not to start a business with the mindset of "taking a semester off to sell the company," but to be extremely obsessed with an idea and willing to invest years in it.

7. Working at a Quant Fund or Large Company is a "Paid Education": Richard Craib believes that working at a place like Google or a quant fund allows you to learn for free and explore intersections of different fields (e.g., crypto, quant finance, machine learning), which is an excellent phase of knowledge accumulation before starting a business.

8. "Playing the real game" is the Key to Entrepreneurial Success: Richard Craib advises entrepreneurs not to downplay their optimism and ambition to appear "low-risk." They should boldly state what they truly want to achieve to attract the best talent.