SI419: What Happens When AI Starts Building the Portfolio ft. Nick Baltas

Market volatility and the evolving role of artificial intelligence in asset management dominate the conversation. Trend following strategies have recently benefited from significant fixed income short positions, outperforming other asset classes as global yields rise. The discussion examines the limitations of using volatility-based triggers to adjust lookback windows in trend systems, noting that such methods often lack statistical significance and fail to outperform static models. Furthermore, the integration of agentic AI into institutional portfolio construction offers a new paradigm for decision-making. By utilizing multiple specialized agents—ranging from macro analysis to adversarial diversification—institutions can achieve greater scale, productive dissent, and a transparent audit trail. While these models face challenges regarding reproducibility and look-ahead bias, they provide a framework for more informed, systematic oversight of complex multi-asset portfolios.