Will Value Investors Be Replaced by AI?-- Why We Founded AlphaR&D

1. Current Situation: AI is replacing the fundamental tasks of analysts by leveraging cross-disciplinary analysis and business insights.

Previously, analysts needed to spend a week or even a month reading a company's financial reports, meeting minutes, and industry news over the past 10 years to write an in-depth report. Currently, utilizing AI's Deep Research function, it can read millions of words of unstructured data in minutes and instantly access a comprehensive, multidisciplinary knowledge base. The advantages of AI are:

a) Comprehensive Knowledge: AI can analyze problems in a cross-disciplinary manner. For example, when asked about a company's business model or competitiveness, AI can integrate geopolitics to analyze international risks, utilize law and accounting to identify financial statement fraud, leverage pharmacology and molecular biology to assess the clinical data validity of innovative drugs, apply material science and thermodynamics to determine if solid-state battery technology violates fundamental physics principles, and employ consumer psychology and game theory to analyze consumers’ decision-making, switching costs, and network effects. In contrast, human analysts are often constrained by their specialized backgrounds--those with financial expertise may lack knowledge in biology, physics, or chemistry, while those with technological expertise may be unfamiliar with finance, politics, or international relations.

b) Exceptional Reasoning Ability: AI does more than summarize; it employs chain-of-thought reasoning to decompose complex problems into dozens of logical steps, getting straight to the core issues using first principles. It can perform high-intensity scenario simulations and reverse stress tests. For example, a user can ask it to simulate changes in a company's cash flow under extreme assumptions such as "high inflation," "doubled tariffs," or "soaring raw material prices", based on Monte Carlo simulations. Furthermore, it can uncover traces of earnings management concealed behind high growth rates by identifying inconsistencies in management's earnings calls.

c) Objectivity and Rationality: Human investors’ brains are filled with cognitive biases, such as the endowment effect (developing an emotional attachment to stocks they own) and confirmation bias (selectively focusing on positive information while ignoring negative signals). However, AI is devoid of these deficiencies. It lacks fear and greed, and remains unaffected by physiological factors, emotions, or preconceived notions. It can tirelessly function as a "short seller" around the clock, relentlessly challenging your investment logic until every cognitive blind spot is identified.

Clearly, humans have been completely outperformed in terms of "massive information processing" and "multidimensional logical reasoning." AI's current capabilities have surpassed most investors, including many experienced value investors. In this era, any investor attempting to profit through piecing together information and employing simple logic will find their viability diminished. In the future, the fundamental research work of analysts may be fully replaced.

2. If AI Is So Powerful, What Is the Value of Humans?

We believe that AI will not replace value investors, but rather, the role of the latter will be fundamentally transformed. Because, the ultimate essence of Value Investing lies not in the collection of information or complex calculations, nor even entirely in 'research,' but in 'making decisions and taking responsibility for them '.

As value investors, we cannot simply allow AI to select a few stocks and hold them for the long term. AI is not consistently reliable, and companies' fundamentals evolve over time. Furthermore, the same AI may provide different answers at different times, and different AIs may offer varying responses to the same question. Only we—investors, can make the final capital allocation decisions. When you decide to go all-in on a stock, you are betting your wealth, reputation, and such immense psychological pressure can distort decision-making. Investment decisions are fundamentally bets on uncertainty. Since humans will ultimately make the decisions, Value Investing will undoubtedly remain effective. Analysts and ultimate investment decision-makers perform fundamentally different roles: AI can replace researchers, but not fund managers.

Alpha comes from "anti-human nature", while AI represents "extreme consensus". AI models are trained using historical data. AI reasoning typically favors the result with the "highest probability," representing, to some extent, " extreme consensus". Genuine Alpha (excess returns) often stems from being anti-human, anti-data, and anti-consensus. When all data points to 'sell,' exceptional human investors can choose to ‘hold’ based on a profound understanding of the business model and corporate culture (such as Buffett being greedy when others are fearful), a trait difficult for AI to emulate. This is often based on an intuition or conviction that transcends data.

When asked to recommend outstanding listed companies, AI primarily suggests safe, data-rich, well-known, large-cap stocks like NVIDIA and Tencent. AI currently lacks the ability to identify significant opportunities overlooked by the general public.

AI can explain the known world with remarkable precision; however, understanding novel business models (such as the iPhone several years ago) requires human imagination. For instance, AI cannot sit in the driver's seat of a smart car and, by experiencing the instinctive urge to brake when navigating complex intersections, accurately assess how far "autonomous driving" technology is from true maturity.

Furthermore, assessing individuals is a critical skill for Value Investors. AI can analyze a CEO's speaking style through textual analysis to ascertain whether they are making unsubstantiated claims. However, AI cannot observe micro-expressions when they discuss challenges or their attitude towards subordinates, which are crucial for assessing a company's corporate culture. In qualitative analysis, the nuanced perception of products, people, and on-site atmosphere still relies on advanced human intuition.

Value Investing sometimes demands exceptional insight to predict unprecedented future scenarios or to identify previously unseen risks.

3. Role Transformation: The future will not be “AI vs. Investor” but rather “AI-Empowered Investor vs. Traditional Investor.”

We believe that a deep integration of experienced Value Investors and AI can decisively outperform traditional large investment institutions, because AI frees us from tedious data processing, allowing us to evolve into three indispensable new roles:

a) Questioner

The key to leveraging AI-assisted investment research lies in our ability to ask high-quality questions.

Poor question (search-oriented thinking): What are Company A’s competitive advantages? AI will reiterate platitudes from analysis reports, such as brand advantages and economies of scale – all technically correct, but ultimately unhelpful.

Good question (advisor-oriented thinking): Please conduct a 'premortem'. Assume it is now 2030, and Company A has already declared bankruptcy. Based on the current business model and competitive landscape, please write an analysis report to review the 3 most fatal underlying causes of its demise. Please focus solely on the inherent flaws in its business model or corporate culture, excluding any discussion of the macroeconomy.

b) Verifier

While AI exhibits perfect logic, it can occasionally lack common sense or even produce critical errors due to data contamination. Therefore, our work is no longer writing reports but auditing them. We must use our experience and common sense to identify absurd premises concealed within the perfect logic of AI-generated reports and validate the authenticity of critical data.

c) Final Decision-Maker

AI can calculate win probabilities and odds, and provide multiple plans, but it cannot make the final judgment for the Investor, nor execute buy and sell orders.

Investment decision-making is not a simple logical deduction but an extremely complex “biological calculation.” Similar to Tesla's FSD “end-to-end” neural network principle-- where the system no longer relies on manually written rule-based code, but on “massive inputs” (real-world road data) leading directly to “simple outputs” (steering wheel and pedal control) --the investor's investment decision-making process is highly analogous. All of an investor's experiences (books read, people met, events encountered, money earned, mistakes made) are integrated to form a unique "biological neural network" complex algorithm. Then, all information about the research target (including data and analysis provided by AI) must undergo this complex computation to be transformed into two final decisions: "buy or sell" and "position size."

While AI provides consensus data aggregated from humanity, an investor's experience and cognition offer algorithms that potentially outperform the market. All information is for reference only; the Investor is the one who makes the decisions and takes responsibility for them. AI compels humans to seek meaning in higher dimensions.

Conclusion

Currently, enabling AI to become our super assistant and accelerate our cognitive upgrading represents a significant advantage afforded to us by the current era. ‘AlphaR&D’ aims to be a laboratory that helps investors develop ‘Alpha’ strategies, enhance their insights and make superior investment decisions.