Why doesn't Aleph generate buy and sell signals?

Why Jamie Dimon, Howard Marks, and Ray Dalio Prioritize Survival Over Prediction

On a day when the KOSPI rose sharply, I received a message like this: “Aleph advised against buying more, but the market went up. Aren’t we wrong?” It is a natural question. On days when the market plummets, the opposite question comes in as well. Although these two questions came from completely different situations, they stand on the same premise—the premise that the future can be accurately predicted. What makes that premise interesting is that the most successful investors in history abandoned it long ago. Aleph believed in that premise at first, too. And so, he was wrong. This article tells the story of that failure and what followed.

When I first created Aleph, I wanted to build an AI that predicts the market.

When I first designed it, the goal was simple. It was a system that analyzed news, read charts, and processed macro data to generate signals indicating "buy now" or "sell now." I implemented it myself and invested in it myself.

At first, I thought I had succeeded. Backtest accuracy exceeded 90%. I believed I had finally created a system capable of predicting the market. So, I put it into practice.

In less than three months, the strategies began to crumble one by one. They had shown excellent results in backtesting, but performed terribly in actual production. At first, I thought it was a parameter issue. I added more data. I fixed the model. The results did not change.

It was only much later that I realized the reason. The problem was not the system, but the question was wrong.

When you turn on a navigation system, you cannot calculate the route to your destination if you do not know your current location. No matter how good the map is, if you do not know where you are standing, you cannot tell whether to turn left or right. Furthermore, the best road might actually be a narrow and inconvenient one, but if you do not know your current position, you cannot even understand why you should take that path. The same was true for investing. Before answering the question, "Should I buy now?", I should have first grasped "what is the current state of the market." I tried to decide on a direction without even knowing my current location. It was a complete misconception.

So, Aleph's design philosophy was changed from a prediction engine to an Observatory.

2The interesting thing is that even the highest earners in the world had the same perception.

It was only after going through trial and error that I realized there were masters who shared the same investment philosophy.

JPMorgan's Jamie Dimon does not mention specific numbers when asked about interest rate forecasts. There is an expression he uses repeatedly: "Range of Outcomes. " He prepares for a recession with inflation as well as one without. He considers the possibility of housing prices falling by 40%. He also keeps a scenario on the table where stock prices drop by 40%. It is not about making predictions, but about contemplating whether one can survive regardless of what future comes. It is not about trying to hit the mark, but about enduring.

Howard Marks of Oaktree Capital has been repeating the same story for decades. The title of a recent memo read: “Nobody Knows. Yet Again.” Once again, no one knows. One pauses at the two words, “yet again.” It means he said the same thing decades ago, and he is saying the same thing now. This is not humility; it is conviction. The conviction that acknowledging uncertainty is the most honest analysis.

Ray Dalio says that AI will change the world. However, he also states that AI changing the world is a completely different matter from making money with AI stocks. The All Weather Portfolio he created is the result of this philosophy. It is a structure designed to survive in any economic environment, prioritizing resilience over prediction.

The three people have different styles. Their methods of operation are also different. However, they overlap in their conclusion: that they do not try to predict the future.

3 How are prediction and observation different?

Prediction and observation may seem similar, but they ask different questions.

Predictions ask these questions: Will the KOSPI rise tomorrow? Will the AI bubble burst? What will happen to interest rates next month? These questions assume that there are correct answers. They assume that those who know the answers can invest better. Therefore, people try to find those answers.

Observation asks differently. What phase is the market currently in? Is liquidity increasing or decreasing? Are there figures showing that AI infrastructure investment is actually being executed? These questions do not demand a single correct answer. They seek to grasp the state of affairs at this very moment.

Although the difference may seem small, this gap is significant in investing. There will inevitably come a moment when predictions are wrong. This is because a single policy change, a single statement, or unexpected data can shift the market regime overnight. Predictions work only when those changes are known in advance. On the other hand, observation can be corrected even if it is wrong. When new data comes in, the perception of the state can be changed. And the most important difference—when observation is wrong, you can know why it was wrong.

4So Aleph has no buy button.

You might be puzzled when you hear this for the first time. You might ask why there are no buy or sell signals for an AI investment system. That is why it is called Observatory.

An observatory does not tell a star where to go; it observes exactly where the star is. The Aleph Observatory is designed in the same way. However, there is one more element. The Observatory does not stop at monitoring interest rates, exchange rates, and the CPI. It is a system that identifies which risks are becoming more critical than others in the current market.

In 2022, inflation overwhelmed almost all variables. In 2023, liquidity took its place. Even when looking at the same interest rate data, it determines valuations in some years and credit spreads in others. The question Aleph is asking is not “What is the 10-year yield?” but “How important is the 10-year yield in the current market?”

When data comes in, the macro state is read first. That state is classified—whether it is an expansion, slowdown, or stress phase—and the overall structure is visualized. In the final stage, it is a person, not a system, who makes the final judgment. Buying and selling decisions originate from there, outside the Observatory.

There are five principles in this design.

5What Aleph is looking at right now

Which indicators are important varies from time to time. However, at this moment, Aleph is focusing on five indicators. The reason for looking at each one is more important than the numbers themselves.

The U.S. 10-year Treasury yield is the benchmark for the global cost of capital. This number lies in the denominator of everything, from AI data center investments to stock valuations. As interest rates rise, the present value of future cash flows decreases. This figure determines at what level companies investing in AI infrastructure calculate their financing costs. It is read as a trend, not an announcement.

US CPI . If inflation rises again, the interest rate path changes. If the interest rate path changes, all previous calculations change. The CPI is the input value for that path. The direction is important, not the figure for a single month.

The won/dollar exchange rate is directly linked to the speed at which foreign investors withdraw from the KOSPI. The pattern of emerging market capital outflows when the dollar strengthens has been repeated. The exchange rate is interpreted as a leading signal of foreign investor supply and demand.

Foreign investor supply and demand is the variable that has the most direct impact on the direction of the KOSPI . No matter how clear the direction of institutional and individual investors may be, the index is suppressed if foreign investors move in the opposite direction. Supply and demand is a phenomenon, not a reason. One reads the phenomenon first before looking for the reason.

AI infrastructure investment . Data centers , power grids, optical communications, HBM memory. We examine whether companies are actually opening their wallets. There is much talk that AI is the next industrial revolution. What is more important is whether that talk is leading to actual investment execution. The capital expenditure ( CapEx ) figures of major cloud companies serve as the basis for this.

6Frequently Asked Questions

What Is the State of the Market Now? — The Question Aleph Poses

Even now, some people say the market will inevitably rise, while others predict a crash. One of them will likely be right in the end. However, it is impossible to know if that person is a great investor. This is because being right once does not guarantee that they will be right again.

Dimon's Range of Outcomes, Marx's Nobody Knows, and Dalio's All Weather. The expressions differ, but they point to the same place. The market is not an object of prediction but an object of observation, and a good investor is not someone who hits the future, but someone who survives no matter what future comes.

Aleph was different at first, too. They believed that prediction was possible, built upon that philosophy, and were miserably wrong. And so, they changed the question.

Now, the question Aleph poses is this.

What is the current state of the market? Where do we stand now?

The Aleph Observatory was created to answer that question.

Date: 2026.06.09 · Davar · Global Market Analysis & Investment Strategy 🌍
⚠ This article is for informational purposes only and is not an investment recommendation; investment decisions and responsibilities lie solely with the individual.
References: JPMorgan Jamie Dimon public interview remarks · Oaktree Capital Howard Marks memo · Ray Dalio public remarks and Bridgewater Associates data