# The Investor's Edge Lives in Their Head — Anekha Sokhal, Moshi

## Executive summary

The presentation argues that while AI and machine learning are powerful tools for handling the 'data grind' in finance (data gathering, cleaning, and processing), the most valuable 'edge' remains the human capacity for synthesis, judgment, and context. The speaker, Anekha Sokhal, proposes that instead of trying to replace the human brain, technology should be built to augment it. Moshi, the company's product, is positioned as a '24/7 quant analyst' that processes massive data streams (e.g., 2.4 billion data points) to provide context and identify blind spots, allowing the human user to focus solely on decision-making and conviction.

## Key takeaways

- The Paradox of Data Overload: The sheer volume of data (e.g., 175 zettabytes in 2025) has created a gap between data availability and human attention, making financial synthesis a critical, difficult skill.
- The Edge is Non-Downloadable: In a world where models and data are accessible to everyone, the true competitive edge is the institutional knowledge and context that resides in an investor's head, which is difficult to replicate or download.
- AI's Role: Augmentation, Not Replacement: The speaker posits that AI should not attempt to replace the human decision-making engine (which runs efficiently on minimal power), but rather feed it with comprehensive, cross-referenced data to improve reasoning.
- The Need for Infrastructure-Grade Reliability: Unlike 'weekend tools,' production systems require robust infrastructure planning, focusing on security, reliability, and managing costs based on API calls, not just tool usage.

## Technical details

- Data Processing and Scale: Moshi functions as a 24/7 quant analyst, processing over 2.4 billion data points and learning new context continuously. It utilizes a team of six deterministic agents to monitor markets and integrate diverse data sources (PDFs, emails, subscriptions).
- Trust and Traceability: A critical technical requirement in finance is trust. The system must measure hallucination, ensure every claim is traceable, and provide the user with the assumptions and reasoning path used to reach a decision.
- System Architecture Shift: The industry is moving from pricing by tools to pricing by API calls, necessitating a focus on building reliable, scalable infrastructure rather than simple models.
- Agentic Risk Management: The speaker warns that handing the decision-making loop entirely to agents is dangerous in finance, as a single small error can exponentially increase, unlike a human who can take accountability for a decision.

## Practical implications

- Build systems that focus on augmenting human judgment rather than attempting to replace it.
- Prioritize data provenance, traceability, and assumption transparency to build trust in AI outputs.
- Design infrastructure to handle continuous, massive data streams (24/7) and manage costs based on granular API usage.
- Focus on creating deterministic, reliable systems that provide a 360° view of information, moving beyond simple LLM debate.

## Topics

Financial Engineering, Quantitative Analysis, Machine Learning Systems, Data Architecture, AI Ethics and Trust, High-Frequency Data Processing, Moshi, AIE CODE summit

Source: https://www.youtube.com/watch?v=UHbODjuviAU