Odd Lots: Goldman’s Argenti on the Improvements in AI (Podcast)

Goldman Sachs CIO Marco Argenti revealed accelerated AI adoption since 2024, shifting from experimental tools to enterprise-wide deployment, driven by advancements in agentic platforms like Claude Code. Agentic AI now automates complex workflows in trading, risk assessment, and client services, reducing latency and operational costs while maintaining compliance—marking a leap from earlier generative AI pilots. Argenti highlighted Claude Code’s role in code generation and debugging, cutting software development cycles by 40% and enabling faster iteration of proprietary trading algorithms. The bank prioritizes hybrid AI models, combining internal datasets with third-party LLMs to mitigate data privacy risks while leveraging external innovation for competitive edge. Despite progress, Argenti cautioned about over-reliance on AI, emphasizing human oversight for high-stakes decisions and ongoing investments in explainability tools to meet regulatory demands.