Samuel J. Cummings III is an award-winning data scientist, keynote speaker, and renowned thought leader in AI, specializing in complex reasoning and memory architecture. In his recent work he has created AI model architecture that runs 94% less tokens than standard LLMs. As Director of Education at Gen AI Works, Sam brings over a decade of expertise in AI and runs a podcast called Gen AI Talks.

In this episode, we sit down with Sam Cummings, Director of Education at Gen AI Works, as he explores the current and future landscape of large language models (LLMs) and their impact on cybersecurity. Sam unpacks the technical and economic limitations of LLMs, highlighting issues such as model cost, scalability, hallucination, and the looming challenges around reasoning and memory management. The conversation delves into the shift from universal LLMs to specialized models, the inevitability of market monopolization by big tech firms, and the environmental cost of massive data centers. Sam also paints a vivid picture of the “arms race” in the cybersecurity sector, predicting a boom in both offensive and defensive capabilities powered by AI, and offers actionable insights for professionals and entrepreneurs looking to thrive in this rapidly-evolving environment.