We scan new podcasts and send you the top 5 insights daily.
Frontier AI paired with specialized retrieval tools can now beat human consensus baselines on complex financial earnings predictions. Even with this benchmark success, substantial error rates remain, underscoring that long-horizon agentic reasoning in finance is still in its early stages.
Financial reports on AI labs, like a recent Wall Street Journal story on OpenAI, are misleading because they rely on lagging data. The industry's rapid shift to an 'agentic' era, where user behavior changes quickly with new model releases, means historical performance no longer predicts future results, leading to flawed market reactions.
On financial analyst benchmarks, top models from Anthropic, Google, and OpenAI are now almost indistinguishable in capability. This convergence suggests the frontier is commoditizing, questioning the return on investment for massive training runs and shifting value up the application stack.
Human wisdom derives from a single lifetime of experience. AI will achieve a superior form of wisdom by simulating billions of potential future scenarios and identifying the statistically optimal paths. This predictive power, already matching elite human forecasters, will be its core advisory function.
A 2022 study by the Forecasting Research Institute has been reviewed, revealing that top forecasters and AI experts significantly underestimated AI advancements. They assigned single-digit odds to breakthroughs that occurred within two years, proving we are consistently behind the curve in our predictions.
Agentic AI is most advanced in software engineering because code provides a constrained, text-based, and verifiable environment. AI agents can now operate for hours, understanding codebases and fixing errors. This iterative reasoning process is a direct preview of how AI will eventually perform long-running, complex investment research tasks.
Frontier models rapidly improve in fields like coding and mathematics because correctness can be automatically validated and hill-climbed. In open-ended scientific and business domains without ground-truth answers, long-horizon agents struggle due to slow real-world feedback loops.
During the time crunch of earnings season, AI excels at synthesizing disparate information. It can instantly compare a CEO's positive guidance against the recently reported cash flow statements of multiple competitors, flagging potential overconfidence or a genuine outlier.
Contrary to the belief that open-source models would quickly catch up, 2024 has shown the opposite. Frontier models are extending their lead, particularly in long-running tasks, which unlocks new enterprise use cases and allows them to capture the vast majority of revenue.
The future of AI in finance is not just about suggesting trades, but creating interacting systems of specialized agents. For instance, multiple AI "analyst" agents could research a stock, while separate "risk-taking" agents would interact with them to formulate and execute a cohesive trading strategy.
When building revenue models, AI can quickly analyze infinite data slices to spot outliers that skew metrics, such as zero-day service renewals or old opportunities creating survivorship bias. This leads to a more accurate model, representing a performance gain, not just an efficiency one.