In this episode of Crazy Wisdom, I, Stewart Alsop, speak with Andrew Einhorn, CEO and founder of Level Fields, a platform using AI to help people navigate financial markets through the lens of repeatable, data-driven events. We explore how structured patterns in market news—like CEO departures or earnings surprises—can inform trading strategies, how Level Fields filters noise from financial data, and the emotional nuance of user experience design in fintech. Andrew also shares insights on knowledge graphs, machine learning in finance, and the evolving role of narrative in markets. Stock tips from Level Fields are available on their YouTube channel at Level Fields AI and their website levelfields.ai . Check out this GPT we trained on the conversation Timestamps 00:00 – Andrew introduces Level Fields and explains how it identifies event-driven stock movements using AI. 05:00 – Discussion of LLMs vs. custom models, and how Level Fields prioritized financial specificity over general AI. 10:00 – Stewart asks about ontologies and knowledge graphs; Andrew describes early experiences building rule-based systems. 15:00 – They explore the founder’s role in translating problems, UX challenges, and how user expectations shape product design. 20:00 – Insight into feedback collection, including a unique refund policy aimed at improving user understanding. 25:00 – Andrew breaks down the complexities of user segmentation, churn, and adapting the product for different investor types. 30:00 – A look into event types in the market, especially crypto-related announcements and their impact on equities. 35:00 – Philosophical turn on narrative vs. fundamentals in finance; how news and groupthink drive large-scale moves. 40:00 – Reflection on crypto parallels to dot-com era, and the long-term potential of blockchain infrastructure. 45:00 – Deep dive into machine persuasion, LLM training risks, and the influence of opinionated data in financial AI. 50:00 – Final thoughts on momentum algos, market manipulation, and the need for transparent, structured data. Key Insights Event-Based Investing as Market Forecasting: Andrew Einhorn describes Level Fields as a system for interpreting the market’s weather—detecting recurring events like CEO departures or earnings beats to predict price movements. This approach reframes volatility as something intelligible, giving investors a clearer sense of timing and direction. Building Custom AI for Finance: Rejecting generic large language models, Einhorn’s team developed proprietary AI trained exclusively on financial documents. By narrowing the scope, they increased precision and reduced noise, enabling the platform to focus only on events that truly impact share price behavior. Teaching Through Signals, Not Just Showing: Stewart Alsop notes how Level Fields does more than surface opportunities—it educates. By linking cause and effect in financial movements, the platform helps users build intuition, transforming confusion into understanding through repeated exposure to clear, data-backed patterns. User Expectation vs. Product Vision: Initially, Level Fields emphasized an event-centric UX, but users sought more familiar tools like ticker searches and watchlists. This tension revealed that even innovative technologies must accommodate habitual user flows before inviting them into new ways of thinking. Friction as a Path to Clarity: To elicit meaningful feedback, Level Fields implemented a refund policy that required users to explain what didn’t work. The result wasn’t just better UX insights—it also surfaced emotional blockages around investing and design, sharpening the team’s understanding of what users truly needed. Narrative as a Volatile Market Force: Einhorn points out that groupthink in finance stems from shared academic training, creating reflexive investment patterns tied to economic narratives. These surface-level cycles obscure the deeper, steadier signals that Level Fields seeks to highlight through its data model. AI’s Risk of Amplifying Noise: Alsop and Einhorn explore the darker corners of machine persuasion and LLM-generated content. Since models are trained on public data, including biased and speculative sources, they risk reinforcing distortions. In response, Level Fields emphasizes curated, high-integrity inputs grounded in financial fact.
