It's worth thinking about the challenges of using AI/Ml in producing alpha and trading signals, especially medium-to-long-term and low-frequency, in the post-Covid environment. In this presentation to AI4 today, I examine how historical patterns of medium/long term investment returns and risk have broken and why they may be a challenge for AI/ML based methods. I illustrate with three examples: the shift in fixed income returns, the turmoil in liquidity and market structure in March 2020, and the exacerbated equity sector leadership/ tech “bubble"
