Position: Universal Time Series Foundation Models Rest on a Category Error

TL;DR

This paper critiques the idea of universal time series models, highlighting their fundamental flaws and proposing a causal control paradigm with specialized solvers and new benchmarks for robustness.

Contribution

It introduces the Autoregressive Blindness Bound, critiques the category error in universal models, and advocates for a hierarchical, causal control approach with new evaluation metrics.

Findings

Universal models degenerate into expensive filters that fail under distributional drift.

History-only models cannot predict regime shifts due to the Autoregressive Blindness Bound.

A hierarchical causal control paradigm improves robustness and adaptability.

Abstract

This position paper argues that the pursuit of "Universal Foundation Models for Time Series" rests on a fundamental category error, mistaking a structural Container for a semantic Modality. We contend that because time series hold incompatible generative processes (e.g., finance vs. fluid dynamics), monolithic models degenerate into expensive "Generic Filters" that fail to generalize under distributional drift. To address this, we introduce the "Autoregressive Blindness Bound," a theoretical limit proving that history-only models cannot predict intervention-driven regime shifts. We advocate replacing universality with a Causal Control Agent paradigm, where an agent leverages external context to orchestrate a hierarchy of specialized solvers, from frozen domain experts to lightweight Just-in-Time adaptors. We conclude by calling for a shift in benchmarks from "Zero-Shot Accuracy" to "Drift Adaptation Speed" to prioritize robust, control-theoretic systems.