Hedge Fund Huddle

The episode produces a handful of genuinely useful practitioner ideas - cross-model LLM evaluation, the 'always critic agent,' and prompt-library governance - but roughly a third of the runtime is consumed by career introductions, platitudes ('AI is exciting'), and basic LLM-101 explanations that informed listeners will already know.

“have another LLM score it. So if I just think of discretionary investors, if you are parsing FOMC statements, say you say you're a retail investor and you're trying to parse an FOMC statement, you can do that pretty easily off the shelf, Maybe say using ChatGPT, have Gemini score it for you”

“one of the agents that we're building internally that I'm very excited about is what I call the always critic agent. So we were talking about conviction and I really am looking for an Agent that is just every idea that gets proposed to it. It tries to point out issues with the idea.”

Cross-model scoring and the critic-agent framing are fresher than typical AI-in-finance content, but the episode largely recycles dominant narratives - AI as junior analyst, fine-tune don't train from scratch, alt-data as moat - that have been circulating for two or three years.

“have Gemini score it for you and have that consistent framework so that you're having different models, different high caliber models sort of evaluate each other because it keeps the output a lot more honest”

“one of the agents that we're building internally that I'm very excited about is what I call the always critic agent”

Nishant is a genuine quant practitioner who has worked at AQR and SAC/Cubist and runs live strategies, lending real credibility; Andrew is a publishing executive and journalist who covers the data industry rather than operates in markets, which dilutes the overall practitioner depth.

“I spent a summer at AQR and then I spent a summer at what was then called SAC Multi Quant but is now called Cubist”

“we have alternative data that we collect on 24 markets, equity markets globally, that's 10,000 stocks. We aggregate those stocks individually and at the country level and we construct signals doing that.”

There are useful concrete markers - 10,000 stocks across 24 equity markets, named models (Claude Opus 4.6, Codex 5.3), dated AI milestones, and specific tools like RoboRev and Motion AI - but the episode never surfaces strategy performance numbers, AUM, or quantified efficiency gains, leaving core alpha-generation claims unverified.

“we have alternative data that we collect on 24 markets, equity markets globally, that's 10,000 stocks”

“2012 we saw the big Vision paper come out of Alexnet. 2018 is when we see Google launch the BERT paper. 2017 I was at a conference where actually the attention is all you need.”

The host brings genuine domain knowledge (his own book-running background) and asks several non-obvious questions - how do you build conviction in an AI output, and how does autonomous execution differ from discretionary AI - but rarely challenges specific claims and allows the regulation and bubble segments to close without meaningful pushback.

“if you're relying on AI, obviously your conviction can only be as high as the conviction you have in the, in the AI platform. So question number one is if conviction is still a big part of the trading, how do you get conviction in AI?”

“I was being specific about discretionary trading there.”