This sort of can’t last though right? The whole advantage in a hedge fund is being on the opposite side of a trend. If everyone is using AI, those returns must start to diminish quickly.
akacrobat · 1d ago
I think at the hedge funds, AI will play out similar to how quant has in the past, i.e. as a technology that will still have alphas unlockable by good usage.
High-performing funds may "beat the market" at models, data, fine-tuning, RAG or even figuring out which AI tool should be deployed for which tasks.
severusdd · 1d ago
Interesting parallel, but as it is packaged today - "AI" looks much lighter on technical customization than quant.
What can you really meaningfully change on the large models today? a bit of RAG, some prompting - sure. But that sounds like a much lighter lift compared to the tasks done by large teams of top-tier CS grads in older quants firms.
daveguy · 1d ago
I agree. Hedge funds coming out with "hey! look at us!" this late in the game is a huge red flag.
techpineapple · 1d ago
“These tools don’t negate the necessity to think… if anything they should just give you more time to think,” he says.
I have real trouble buying this. On the one hand maybe most people don’t think and just deferring to AI makes better decisions. Maybe the kind of people in important decision making posts at a $10B hedge fund have the discipline not to outsource their thinking to an LLM (when not appropriate).
But I would be shocked if a few years from now we don’t start seeing some analysis/studies/articles talking about how decision making in companies is sort of coalescing around an AI “median” opinion.
Also, it seems inevitable that some set of Hedge Funds will lose existential amounts of money using AI wrong.
For example, see this recent article from Bloomberg- https://www.bloomberg.com/news/articles/2025-02-13/hedge-fun...
High-performing funds may "beat the market" at models, data, fine-tuning, RAG or even figuring out which AI tool should be deployed for which tasks.
What can you really meaningfully change on the large models today? a bit of RAG, some prompting - sure. But that sounds like a much lighter lift compared to the tasks done by large teams of top-tier CS grads in older quants firms.
I have real trouble buying this. On the one hand maybe most people don’t think and just deferring to AI makes better decisions. Maybe the kind of people in important decision making posts at a $10B hedge fund have the discipline not to outsource their thinking to an LLM (when not appropriate).
But I would be shocked if a few years from now we don’t start seeing some analysis/studies/articles talking about how decision making in companies is sort of coalescing around an AI “median” opinion.
Also, it seems inevitable that some set of Hedge Funds will lose existential amounts of money using AI wrong.