Episode summary
$242B Investor on AI, Market Bubbles & How the Best Investors Build Portfolios
How I Invest With David Weisburd
· · 1 hr 4 min · Hosts: David Weisburd · Guests: Peter Hecht
Summary
Peter Hecht argues that the right way to think about portfolio construction is through underlying risk, not the number of holdings. At AQR, he says diversification works like a casino taking thousands of small, largely uncorrelated bets, and the goal is to build a portfolio that can deliver an attractive return for a given level of risk whether the client wants market neutrality or full beta exposure. He also says tracking error is not a flaw but a necessary consequence of trying to beat a benchmark.
Hecht says alpha has no single universal definition because it depends on the risk model and the investor’s portfolio. He contrasts simple benchmark-relative views with broader risk-adjusted measures, and says the most important test is whether a new idea improves the customer’s actual portfolio. He also pushes back on older factor stories, saying AQR long ago questioned the small-cap effect and has treated value as a more flexible, evolving concept than a single book-to-price screen.
The conversation turns to AI, where Hecht says AQR uses large language models differently from prompt-based stock picking. The firm uses them to transform documents into numerical features and then tests whether those features predict returns. He also says systematic investing succeeds through judgment, not just data, and that diversification, trend following, and portable alpha can help investors handle equity dominance, inflation shocks, and the behavioral problem of abandoning diversifiers too early.
Key takeaways
- Hecht says tracking error is the price of active management: the more a manager deviates from a benchmark, the more chance to outperform but also the more chance to lag.
- Hecht argues AQR’s diversification works because thousands of small, uncorrelated trades can create return without needing a single big edge.
- Hecht says alpha is conditional on the risk model and the client’s current portfolio, not a universal number.
- Hecht says AQR uses AI to turn text into predictive features, not to ask ChatGPT or Claude which stocks to buy.
- Hecht says most portfolios still have roughly 90% of their risk tied to equity exposure, which is why trend following and portable alpha matter.
- Hecht says trend following can help in protracted growth or inflation shocks, while portable alpha lets investors keep beta exposure and add uncorrelated active return.
Notable quotes
“tracking error is going to give you a sense of how far a manager's relative return, how different it can be from the S&P 500”
“You have to take tracking error to beat your benchmark. The more tracking error you take gives you more possibility for beating the benchmark”
“we believe in diversification. We're militant about thinking about are these trades correlated with each other”
“alpha is always relative to what model? What is your risk model”
“the most potent thing is to actually calculate alpha relative to their current opportunity set, which is the portfolio they hold”
How the episode opens
Pete, at AQR, you have $242 billion assets under management. You're one of the largest hedge funds in the world. And one of the things I wanted to start with is one of my biggest pet peeves in this industry is this concept of a tracking error.
Tell me about tracking error and is it a misnomer? We're familiar with concepts like volatility, which sort of gives you a sense of the range of outcomes for an investment's return. tracking error is the volatility but it's the volatility of the relative return your return versus a benchmark.
Okay, so let's say the S&P 500. Okay, so tracking error is going to give you a sense of how far a manager's relative return, how different it can be from the S&P 500. So like a tracking error of 3% means if you think it's a normal distribution, right?
You're going to have a return that's going to be either 3% ahead of the S&P 500, 3% behind the S&P 500 with roughly a 66% probability. So it gives you a sense of how much active risk am I taking and is it consistent with my risk tolerance in terms of my willingness and ability to tolerate lagging the benchmark.
Can tracking error be a good thing?
Yes. Because if you want to beat the benchmark, if you want to try to add value, you have to take tracking error. So I always tell people you can take bad tracking air, right?
That's with a manager that does not have a good process for determining what's an attractive security and not an unattractive security. Or you can invest with a manager you think has a good process. You have to take tracking error to beat your benchmark.
The more tracking error you take gives you more possibility for beating the benchmark, but it also means you might actually lag the benchmark by more when your manager experiences bad luck.
From the episode’s captions: roughly the first two minutes, lightly cleaned. Keep watching on YouTube →
Topics
- ai portfolio construction
- diversification and tracking error
- alpha definition
- market bubbles and systematic investing
- trend following vs equities
Companies and tickers mentioned
- AQR Capital Management
- OpenAI
- Anthropic
- ChatGPT
- Claude
Watch the episode
Watch “$242B Investor on AI, Market Bubbles & How the Best Investors Build Portfolios” on YouTube
Summary, takeaways and quote selection by Octagon. Quotes are short excerpts; the full conversation is in the episode.
More from How I Invest With David Weisburd
-
Hunter Walk: Why Venture is Getting Harder, Picking Great Founders, and Investing Against Consensus -
True Anomaly’s Even Rogers on Space Warfare, China’s Military Advantage, and the Defense Tech Boom -
Finding the Next Anthropic, Quantifying Venture Capital, and the Science of Picking Winners -
Why 95% of AI Projects Fail, Anthropic still at $2T?, and Why AI Is Still in the First Inning -
Peter Thiel's 27-Year Partner on the AI Bubble, the Venture Capital Crisis, and the PayPal Mafia -
Why Venture Is Eating Public Markets, Picking Great VCs, and the AI Boom w/Brijesh Jeevarathnam
How I Invest With David Weisburd, this episode and its artwork belong to David Weisburd. Octagon is not affiliated with the show; this page is an independent summary.
