NYC

AI Street Meetup

I’ll be back in New York next month to attend Cornell’s Future of Finance & AI conference on Sept 19. It has a great agenda exploring many of the topics I’ve been covering and includes some speakers I’ve previously interviewed like Snowflake’s Jon Regenstein and Citadel’s Tharsis Souza. Plus, many more speakers I plan to interview.

I’m only in town for a few days, and there are many folks I want to connect with so I’m thinking of organizing an informal AI Street meetup for drinks that Thursday night (Sept. 18.) Reply to this email if you’re interested and I’ll figure out a spot.

RESEARCH

BlackRock Researchers Develop AI Agent System for Stock Picks: Study

LLMs are often treated as one-size-fits-all tools: they can suggest what to make for dinner from a fridge photo or plan a trip to the Marshall Islands.

But AI struggles with tasks requiring mathematical precision. Because it’s probabilistic, the same prompt can yield different answers — and even the model builders can’t fully explain why.

Instead of a single model, BlackRock built three specialized “agents” that mimic different analyst roles, according to a new study.

• Fundamental Agent — parses 10-Ks and earnings reports

• Sentiment Agent — reviews news and analyst ratings

• Valuation Agent — studies prices, volatility, and volumes

Each agent analyzes a stock independently, then enters a round-robin debate. Disagreements are argued until the agents reach consensus on whether to BUY or SELL — a process designed to mimic an investment committee.

The system runs on Microsoft's AutoGen framework using GPT-4o, with custom tools for each agent: document parsing for 10-Ks, news summarization, and volatility calculators.

The agents' recommendations change based on risk tolerance settings. The same volatile stock might get a SELL from a risk-averse agent but a BUY from a risk-neutral one analyzing identical data.

Tested on 15 tech stocks over four months in 2024, the system outperformed both single agents and the benchmark in risk-neutral portfolios on a risk-adjusted basis (Sharpe ratios). In risk-averse portfolios, all approaches lagged the benchmark — since volatile tech names were excluded — but the multi-agent showed smaller drawdowns than single agents.

The authors argue this setup improves analytical rigor and helps mitigate behavioral biases like overconfidence. While limited in scope and not a full portfolio optimizer, the study suggests specialized, debating agents may prove more reliable than general models for quantitative finance.

Takeaway: LLMs can be a Swiss Army knife in daily life, but for mathematical analysis, specialized agents may be the sharper tool.

Paper: AlphaAgents: Large Language Model based Multi-Agents for Equity Portfolio Constructions 

AQR Study Sparks Clash Over AI’s Place in Quant Investing

Quant investing has long favored simple models to avoid overfitting on noisy market data. But AQR’s Bryan Kelly challenged that orthodoxy with a Journal of Finance paper arguing that more complex, machine learning–driven models can outperform.

The study, titled "The Virtue of Complexity in Return Prediction" and co-authored with researchers from EPFL and Yale, showed a US stock market strategy using over 10,000 parameters trained on just 12 months of data beat a simple buy-and-hold benchmark.

The study ignited a backlash. At least six academic papers have pushed back, with critics like Chicago’s Stefan Nagel and Stanford’s Jonathan Berk calling the results “hard to believe” and “virtually useless” for predicting returns, according to Bloomberg. Many argue the model simply rediscovered momentum trading on a short data window. Kelly has defended the work as "proof of concept research."

Takeaway: The fight underscores how unsettled AI’s place in finance is —promising to some, but constrained by data limits that set markets apart from other AI domains.

Related:

• AI Outperforms Factor Models (AI Street)
  

ADOPTION

Selloff in AI Shares Reflects Adoption Hurdles, Not Weak Models

AI-related stocks fell this week after an MIT report said 95% of projects aren’t generating ROI.

But when you dig deeper into the MIT report, it’s not really about the models being broken. The research—150 interviews, a survey of 350 employees, and an analysis of 300 public AI deployments—shows the problem is more about the learning gap between humans and the technology. Companies are still figuring out how to actually use it.

MIT points out that most budgets are being poured into sales and marketing, even though the biggest ROI is showing up in back-office automation—things like cutting outsourcing costs and streamlining operations. And the success rate is much higher when firms buy from vendors or partner (67%) instead of trying to build in-house (33%).

That build-vs-buy theme reminded me of a conversation I had with Pete Harris earlier this year. He compared today’s AI moment to the cloud industry in the early 2010s: “They said the same thing with cloud technology,” Pete told me. “Firms were like, ‘We’ll build a private cloud, sure, but we’ll never rely on public clouds.’ But now? Many financial companies would be lost without it.”

If cloud is any guide, adoption curves take time.

Even so, the headlines were enough to knock down stocks like Nvidia, Palantir, and Arm. Which seems wild, given this technology has only existed for three years. I’m not sure what ROI timelines investors have. This reaction is lost on me.

Takeaway: We’re still in the very early stages of AI adoption in the business world. It’s hard to judge ROI on this short of a timeline.

WHAT ELSE I’M READING

• BofA CIO Sets Sights on AI With $4 Billion Tech Spend (WSJ)

• Energy Infrastructure, Not Software, Will Determine AI Market Winners (Drift Signal)

• Today in YOLO: Gamblers Now Bet on AI Models Like Racehorses (WSJ)

• Startups Revise Stock Pay Amid Delayed IPOs, AI Talent War (Bloomberg)

• 12% of Daily Corporate Bond Trading Happens Near Market Close (The Trade)

• Is AI The Scapegoat Employers Use To Explain Technology Layoffs? (Forbes)

• AI Already Informs Indexing, Portfolio Management (CIO)

• Arizona’s CIO On Using AI For Retirement Fund Management

CALENDAR

Upcoming AI + Finance Conferences

• AI in Financial Services (Arena) – Sept 9–10, 2025 • London

  Focused on AI strategy, implementation, and ROI in finance.

• Cornell Financial Engineering Manhattan 2025 Future of Finance & AI Conference – Sept 19, 2025 • New York (I’ll be attending)

  A one-day forum on AI, quantitative finance, and hedge-fund strategies, attracting leading quants and industry practitioners.

• Bloomberg-Columbia ML in Finance Conf – Sept 25, 2025 • New York

  Academic–industry event hosted by Columbia University and Bloomberg, focused on ML applications in finance including asset pricing, market forecasting, and LLM risk.

• GAIIM Conference 2025 – Sept 30, 2025 • New York

  Forum on practical applications of AI in investing, featuring tools for research, valuation, and portfolio workflows.

• AIFin Workshop at ECAI 2025 – October 26, 2025 • Bologna, Italy

  One-day academic workshop on AI/ML in finance, covering trading, risk, fraud, NLP, and regulation.

• AI in Finance 2025 – October 27–30, 2025 • Montréal

  Academic event covering ML in empirical asset pricing and risk.

• ACM ICAIF 2025 – November 15–18, 2025 • Singapore

  Top-tier academic/industry conference on AI in finance and trading.

• AI for Finance – November 24–26, 2025 • Paris

  Artefact’s AI for Finance summit, focused on generative AI, future of finance, digital sovereignty, and regulation 

• NeurIPS Workshop: Generative AI in Finance – Dec. 6/7 • San Diego One-day academic workshop at NeurIPS focused on generative AI applications in finance, organized by ML researchers.