V. Hagenmeyer

TSFMs can achieve competitive forecasting performance in critical infrastructure applications while also providing interpretable explanations that align with established domain knowledge.

LLMs can ace general cybersecurity, but choke on vendor-specific trivia and formal standards, revealing critical gaps in their real-world applicability.

Foundation models don't always win: task-specific models can rival or even beat them in electricity price forecasting, especially with clever feature engineering or transfer learning.

LLMs know the lingo of industrial control systems, but can't yet reliably hack them without specialized tools.