How do the world's largest hedge funds view AI?
On Monday, July 3, Greg Jensen, Chief Investment Officer of Bridgewater, systematically talked about Bridgewater's views on AI technology in an interview, sharing his views on how Bridgewater invests in AI, how to use AI investment, and the outlook for AI technology.
How Bridgewater invests in AI
Jensen said:
In restructuring Bridgewater, we also did something that we hadn't done before, which was to get some people to research and invest in things that might not be immediately profitable, but this is our long-term project.
So we set up this AI project, with a team of 17 people, led by me. I am still actively involved in the core work of the Bridgewater Foundation, but the other 16 people are 100% committed to reinventing the Bridgewater Fund through machine learning.
We're going to set up a fund specifically run by machine learning technology, and that's what we're doing now in the lab, and pushing the limits of AI, machine learning capabilities.
Now, there are still big problems in setting up such a fund. If we take large language models, they have two types of problems. First, these models are trained more on the structure of the language, so they usually give back something that looks structured, grammatically correct, but not always an accurate answer. This is a problem. Second, it creates hallucinations, it makes things up because it is more concerned with the structure of the word or concept that appears next than whether the concept that appears next is accurate.
Therefore, Jensen believes that AI can help people conceptualize and theorize about what they observe, but there is still a long way to go to actually use AI to select stocks. Therefore, Bridgewater's real focus is:
But there are other ways to combine this with statistical models and other types of AI. That's what we're really focused on, which is to combine large, poorly accurate language models with statistical models that are good at accurately describing the past but poor at predicting the future.
Bringing that together, we set out to build an ecosystem that I believe can achieve what Bridgewater analysts are doing.
If this ecosystem is in place, we will have millions of investment partners at the upper middle level at the same time. If we have the ability to control the illusions and errors of AI through statistics, we can get a lot of work done quickly. That's exactly what we've done in the lab and proven that the process is feasible.
How does Bridgewater invest through AI?
If it could build an ecosystem that includes AI and other technologies, how would Bridgewater invest in that system?
Jensen believes that statistical AI and large-scale language models can complement each other and play the role of bridgewater's "right and left hand" in investment:
Statistical AI can take theories, go back to whether those theories were true at least in the past and what their flaws were, and refine them, offering suggestions on how to do it differently, and then we can have a conversation with them.
One advantage of a large-scale language model is that it takes a complex statistical model and discusses what it is doing. There are ways to train a language model to do this. The way we simulate this situation is that language models can come up with underlying theories. It's not the most creative thing in the world, but it's the theory of scale, that's for sure. Again, large-scale language models are very good, but we have to adjust the language model somehow, and we can use statistics to control it.
We can then use the language model again to take the results in the statistical engine and discuss it with humans or other AI, and report what was found, what was done, and what type of theory was made. If the conclusions are contrary to what people think, then more tests are conducted.
That's the cycle I'm very excited about, and as I said, so far, statistical AI has been limited because it focuses on market data. The benefit for a language model is that it enables a better understanding of what statistical models don't.
For example, market statistical models do not have the concept of greed, but large-scale language models can almost understand the concept of greed — these models have read all the articles about greed and fear and the like. Therefore, combining the two now produces a human-like thinking pattern.
What does AI mean for human employees?
Over time, computers can do more and more. Jensen argues:
I would say that today, humans are used to only performing roles related to intuition and creativity, and we use computers to memorize and run these rules constantly and accurately. That's only halfway through the transition, and now it's another leap.
There is no doubt that AI will change the role played by investment assistants. To be precise, we'll still need people to work around these things for the foreseeable future, we'll still need a while to build these ecosystems of machine learning agents and so on.
Leveraging AI will be part of the future of work, and I think it's hard in any knowledge industry not to take advantage of these technologies.
In computer programming, we are seeing huge breakthroughs in coding. Now, with AI, people only need to know what they want to encode, not how to code, which is a huge breakthrough. As a result, a group of people who are not well trained or competent in C++, Python, or otherwise can suddenly get what they want faster.
So all of a sudden, the skill sets needed in the workplace are changing, and the way they're changing is surprising to many people, because it's actually a lot of knowledge jobs, such as content creation, that people once thought would be replaced by machines in the distant future, but they're actually just around the corner.
So the most important thing is that there are so many changes that you need to be flexible in the workplace and be able to take advantage of any tool, which is very necessary.
Can investments be managed directly with AI?
There are now a variety of AI investment management tools on the market, and people are concerned that with the great development of AI, will humans only need to hand over investment to AI in the future?
Jensen argues:
I think it's both going to lead to accidents and I'm really excited. Obviously, I'm excited about the power of AI, and I think there are ways to take advantage of it well. But at the same time, AI generates a lot of errors.
Some funds use GPT to pick stocks, but these fund managers don't really have a deep understanding of AI and possible weaknesses.
In one example, in the real estate market, real estate agency platform Zillow used AI technology to predict home prices, assess home prices, and enter the market to start buying houses that AI deems undervalued. However, Zillow has several problems.
One is that although they have a lot of housing data, this data is happening in a relatively short period of time. So while they have a seemingly large number of data points, there is still a macro cycle that influences the assessments they make.
Second, when it's actually an adversarial market, they underestimate the disconnect between theory and practice.
So this was obviously a huge problem for Zillow, who had a big impact on the real estate market and then suffered a huge failure.
Going back to the stock market, very short-term trading, arguably more suitable for machine learning, because there is a lot of data that AI can learn faster with that data.
But on the other hand, in the longer term, the role of AI may not be able to play out. The data is usually like the heart rate data of a person's lifetime. You might think, wow, my heartbeat has been going on for 49 years, which may seem like a lot of data, but when you have a heart attack, the numbers are completely irrelevant. Therefore, even with large amounts of data, it can be misleading, and these issues will lead to huge problems with these technologies.
Therefore, people must understand these tools, what they are good at and what they are not good at, and put them together in a way that brings out the strengths of each tool and avoids the weaknesses.
There's still a lot of work to be done on large language models, and we can certainly train with reinforcement learning to make sure they don't make known mistakes.
Are markets still dominated by optimism?
Jensen believes that the market is still dominated by optimism. He said:
The Fed seems a little more realistic than the market in terms of what it will do. When you look at the reaction of the market, you can see that it is very optimistic.
But we have to note that historically, markets have often tended to be overly optimistic.
This article is from Wall Street News, welcome to download the APP to see more