I remember the exact moment the power of AI really hit me. It wasn’t when I invested in OpenAI or Anthropic, as you might expect. It was a few years ago, sitting in a slightly-too-expensive coffee shop in Palo Alto, listening to a 22-year-old founder pitch me. He wasn't building another social app or a B2B SaaS tool. He was building an AI to trade stocks. His pitch was raw, the deck was a mess, but his code was brilliant. He showed me a backtest where his algorithm turned $10,000 into over $2 million in three years. I didn’t invest—the kid was too green—but I walked out of that coffee shop knowing, with absolute certainty, that the world of finance was about to get completely rewritten.
That kid, and hundreds like him, are the vanguard of a new revolution. Forget the slick, Gordon Gekko types yelling into phones on a chaotic trading floor. The new king of the hedge fund world is an algorithm. A new breed of hedge fund is emerging, one that is run almost entirely by artificial intelligence. The quants are taking over, and their best traders aren't human anymore.
From Human Gut to Machine Brain
For decades, active investment management was about the human element. It was about a star portfolio manager’s intuition, their “feel” for the market, their network of contacts. They’d read the Wall Street Journal, get a hot tip from a source, and make a big bet. Sometimes it paid off spectacularly. More often, it didn't. The data is pretty clear: the vast majority of human-run hedge funds fail to beat the market over the long term.
Then came the quants. These were physicists and mathematicians who saw the market not as a story, but as a giant data set. They built complex statistical models to find tiny, fleeting patterns that human traders could never spot. They started winning. Big names like Renaissance Technologies and Two Sigma became legends, not because of a star trader, but because of their army of PhDs and their fortress of servers.
Now, we're at the third wave. The AI wave. It’s the quant revolution on steroids. Instead of humans building static models, we have machine learning systems that build their own models. They learn, adapt, and evolve in real-time. It’s like having a million quants working 24/7, constantly searching for an edge.
How Do These AI Funds Actually Work?
So what does an AI-powered hedge fund actually do? It’s not Skynet waking up and deciding to short Tesla. The reality is both more complex and more fascinating.
Imagine you want to predict the price of a stock. A human might look at the company's earnings report, read some news articles, and look at a price chart. A first-wave quant might build a model that looks at a dozen variables—price-to-earnings ratio, moving averages, sector performance, etc.
An AI does something different. It might ingest everything. We’re talking about:
- Market Data: Every tick of every stock, bond, currency, and commodity trade going back decades.
- Fundamental Data: Every line of every SEC filing, every earnings transcript, every analyst report.
- Alternative Data: This is where it gets wild. Satellite images of Walmart parking lots to predict retail sales. Credit card transaction data. Social media sentiment analysis from Twitter and Reddit. Even the tone of a CEO’s voice on an earnings call.
The AI sifts through these petabytes of data, looking for correlations a human mind could never conceive of. Maybe it discovers that a certain combination of weather patterns in Brazil, shipping container movements in China, and a specific phrase trending on Twitter is a powerful predictor of a 0.5% move in a particular commodity's price two days from now. It’s not about finding one magic signal; it’s about finding thousands of weak signals and combining them into a high-probability forecast.
These systems are constantly running experiments, backtesting new strategies, and discarding ones that don't work. They are learning machines in the truest sense. My own journey as a software engineer and CTO at multiple startups taught me the power of iteration. You build, you measure, you learn. These AI funds are doing that millions of times a second.
The Risk Isn't the AI, It's the Humans
Naturally, this brings up a lot of fear. What happens when the AI goes rogue? What about a flash crash caused by dueling algorithms? These are valid concerns. The risk is real. But the risk isn't the AI itself. The risk is the hubris of the humans who build them.
An AI is a tool. It will do what you train it to do. If you train it on flawed data, or give it a flawed objective, you will get flawed results. The 2010 Flash Crash wasn't caused by some superintelligent AI; it was caused by a mess of different algorithms, built by different teams with different goals, all interacting in an unpredictable way. The system as a whole was fragile.
That’s why the field of “AI risk management” is becoming so important. It’s not just about having a kill switch. It’s about building systems that are robust, that understand their own uncertainty, and that can explain why they are making a particular decision. It’s about having humans in the loop, not to second-guess every trade, but to set the strategy and the ethical boundaries.
I have strong opinions on this. I believe that any team building an AI trading system without a deep, almost obsessive focus on risk and interpretability is being reckless. The goal isn't to build a black box that spits out money. The goal is to build a glass box, where you can see the machinery inside and understand the logic. That’s the only way to build trust in these systems.
The Future of Investing is Here
So, is this the end of the human fund manager? For most of them, yes. The idea of paying someone 2% of your assets and 20% of the profits to simply guess which way the market is going is already obsolete. You can get better, more reliable returns from a simple index fund. Or, soon, from an AI-powered ETF that gives you access to these sophisticated strategies for a fraction of the cost.
But I don’t think humans will be removed from the equation entirely. The best AI-powered funds will be a partnership—a centaur, to borrow a term from the chess world. The AI will handle the brutal, high-speed, data-driven execution. The human will provide the high-level strategy, the creative insights, and the ultimate oversight.
The AIs can analyze the data, but a human can understand the story behind the data. A human can anticipate a paradigm shift—a new technology, a political upheaval—that isn't reflected in the historical data the AI was trained on. My experience investing in over 200 startups, from the earliest days of AI with companies like Scale AI and Hugging Face to the giants like OpenAI, has taught me that the biggest opportunities come from seeing the world as it will be, not just as it has been.
The kid in the coffee shop never did launch his fund. He ended up joining one of the big quant shops. But the fire he lit in my mind that day is still burning. The rise of AI-powered hedge funds isn't just another trend. It’s a fundamental rewiring of how markets work and how wealth is created. The quants have taken over. The future of active management is not human. It’s a machine. And it’s just getting started.
The Anatomy of an AI-Driven Trade
Let's get more concrete. What does a trade initiated by an AI actually look like? It's a world away from a human trader seeing a stock dip and deciding to "buy the dip."
First, there's the data ingestion pipeline. This is a massive, constantly running system that pulls in data from hundreds of sources in real-time. Think of it as the AI's sensory system. It's not just stock prices. It's parsing thousands of news articles per second, categorizing them, and scoring them for sentiment. It's tracking the flight paths of corporate jets. It's analyzing satellite imagery to count cars in a retailer's parking lot or measure the shadows of oil tankers to estimate their fill levels. My first startup, MovieLaLa, reached 360 million people through network partnerships; the scale of data we handled there pales in comparison to what these funds process before breakfast.
Next, the feature engineering stage. Raw data is useless. The AI needs to transform it into predictive signals, or "features." This is where a lot of the secret sauce is. For example, instead of just looking at the text of a news article, the AI might create a feature for "the rate of change of negative sentiment in articles about a specific company's supply chain." It might create another feature that tracks the geographic location of job postings for software engineers with a specific, rare skill. It's about finding creative ways to represent reality in a way a machine can understand.
Then comes the modeling. This is where techniques like deep learning, reinforcement learning, and natural language processing come into play. The AI takes these thousands of features and uses them to train a model. Or more accurately, it trains thousands of competing models simultaneously. Each model is like a hypothesis about how the market works. The AI is constantly testing these hypotheses against new data, reinforcing the ones that work and discarding the ones that don't. It's a digital version of Darwinian evolution, happening at light speed.
Finally, there's execution and risk management. When the ensemble of models generates a high-conviction signal, it's translated into a trade order. But it's not just a simple "buy" or "sell." The execution algorithm itself is an AI, designed to place the order in the market with minimal price impact, often breaking it up into thousands of tiny "child" orders. Simultaneously, a risk management AI is watching the entire portfolio, ensuring that the new trade doesn't concentrate risk too much in one sector or factor. It's a system of systems, all working in concert.
The Unfair Advantage
When I was building RemoteTeam, which was later acquired by Gusto, we focused on creating a fair and level playing field for hiring global talent. The world of AI-powered finance is the exact opposite. It is fundamentally about creating and exploiting an unfair advantage.
The advantage isn't just speed or data. It's about cognitive horsepower. A human can maybe keep track of a dozen stocks in detail. An AI can track the entire market, every single stock, every single relationship between them, all at once. It doesn't get tired. It doesn't get emotional. It doesn't get scared during a market crash and panic-sell. It just follows the logic of its programming.
This is why my investments in companies like Pika, which is revolutionizing video creation with AI, and Brainbase, which uses AI to manage intellectual property, are so exciting. They are creating these unfair advantages in their respective domains. The same principle applies to finance. The funds that have the best data, the best talent, and the most sophisticated AI models will win. Everyone else will be left behind.
This doesn't mean it's easy. Building one of these funds is incredibly hard. It requires a rare combination of expertise: world-class AI researchers, brilliant software engineers, and savvy financial minds. It also requires a massive amount of capital for data, computing power, and talent. But for those who can pull it off, the rewards are astronomical.
My book, "Becoming Top 1%," talks about the mindset required to achieve outlier success. It's about relentless learning, embracing asymmetric opportunities, and building systems to scale your own abilities. These AI funds are the ultimate expression of that philosophy. They are systems for scaling financial intelligence to a level that was previously unimaginable. The future isn't just automated; it's intelligent. And in the world of finance, intelligence is the only currency that matters.
Frequently Asked Questions
How has this view evolved over time?
My thinking on most topics has changed significantly over the years. Early in my career, I held many conventional views that experience proved wrong. I try to update my beliefs when the evidence changes.
Do all experts agree with this view?
No, and that's fine. The best ideas in business are often contrarian. I share my perspective based on my experience and data, but I encourage you to seek out opposing viewpoints and form your own conclusions.
How can I apply this thinking to my own situation?
Start by identifying the core principle behind the opinion, not the specific example. Then ask yourself: does this principle apply to my context? If yes, test it in a small, low-risk way before going all in.
What experience informs this perspective?
This perspective comes from over a decade of building companies in Silicon Valley, two successful exits (RemoteTeam to Gusto, MovieLaLa to Gfycat), and investing in 200+ startups including Anthropic, OpenAI, and Scale AI. I write about what I've lived.