We Asked a Hedge Fund Manager With 40%+ Annual Returns Why Too Much Compute Could Break the AI Trade
It offers a manager's account of how research effort, position sizing and shared exposures interact, which is useful context for advisors evaluating long/short managers. It also gives a practitioner's view of where AI helps in research and where he thinks human judgment still matters, including his caution on the AI compute trade.
Summary
By the alternatives in wealth deskExcess Returns host Jack Forehand interviews David Orr of Militia Capital, a former professional poker player, about how he builds long and short theses, sizes positions and manages the risk of being wrong. Orr describes running hundreds of positions while watching the exposures they share, and explains where he sees limits to additional research in improving a decision. The conversation also covers how AI is changing his research workflow, his concern about a potential oversupply of compute, and opportunities in Japan.
- Orr says additional research can increase conviction without improving the quality of an investment decision.
- He spreads bets across hundreds of long and short positions and pays close attention to the exposures those positions share, so small positions can add up to a large shared risk.
- He is concerned about a potential oversupply of compute and frames compute as a possible commodity boom-and-bust cycle.
- He uses AI to translate Japanese filings, prioritize news and monitor portfolio exposures, but says its tendency to repeat consensus views leaves room for human judgment.
Chapters
From the publisher’s description| Time | Chapter | Length |
|---|---|---|
| Poker, imperfect information and investing | 5 min | |
| Why David Orr sees a bigger edge in markets | 5 min | |
| Finding Japanese businesses and limiting research time | 5 min | |
| How AI is changing investment research | 5 min | |
| Human investors, AI and the future of portfolio management | 5 min | |
| AI, jobs and who benefits from technological change | 4 min | |
| The case for a compute oversupply | 5 min | |
| Building long and short theses and managing shared risks | 5 min | |
| Selecting portfolio managers and opportunities in Japan | 5 min | |
| Leverage and adapting as investment edges change | 4 min | |
| Index funds and the cost of picking stocks | 5 min | |
| Testing your investing skill and knowing when to stop | 3 min |
Similar items
Same topic Behind the Ticker: NVPS Bets on Chipmaker Suppliers and Everything Except Nvidia | ECETF.com |
Talk Your Book: The Anthropic IPO | ASAnimal Spirits Podcast |
Single-Manager Hedge Funds, Pod Shops, and the Age of AI – Paul Enright (EP.527) | CACapital Allocators with Ted Seides |
[REPLAY] Paul Enright – Inside Long-Short Equity Investing (Capital Allocators, EP.266) | CACapital Allocators with Ted Seides |
An AI Bubble Could Hit Stocks. Inflation Could Hit Bonds. How Do You Diversify and Keep Both? | ERExcess Returns |
Get the weekly 3 to watch
Free, in the Morning BriefEvery week the Morning Brief picks three videos and podcasts worth your time, with why each one matters.
Free. Every morning. Unsubscribe from any edition in one click.
Videos and podcasts are embedded from their publishers. Selections and summaries are editorial and are not investment advice.




