Umar Agha

Research interests: thought experiments; philosophy of physics; values in modeling and simulation

Umar is broadly interested in what happens when our most powerful theories reach their explanatory limits. In quantum mechanics, despite unprecedented predictive success, foundational puzzles persist: the measurement problem, information paradoxes in black hole physics, and arguably incomplete or inconsistent axiomatic foundations. Beyond technical issues, these puzzles raise questions about what we can know in principle across fields when experimental verification is impossible or fails. In AI, as we build systems that exhibit seemingly emergent properties, we confront a parallel fundamental challenge: how do we establish valid criteria for consciousness, comprehension, or agency without circular reasoning? (All theoretical frameworks must otherwise ultimately rest on unprovable first principles or face infinite regress.)

Given this foundational indeterminacy, he explores how emergent phenomena can be efficiently represented in simulations where most interactions are sparse but their cascading effects define the system, and how acknowledging what we cannot predict in principle leads to architectures that compute with emergent dynamics rather than simulate underlying mechanics.

Thought experiments have often served as tools of inquiry to bridge such empirical gaps. He is interested in interrogating how their epistemic status has been contested over time, and what this reveals about how different fields establish criteria for valid reasoning in the absence of empirical evidence.

Outside of these themes, he enjoys game design and exploring the use of seemingly impossible puzzles (e.g., Cicada 3301) as selection and sorting mechanisms.

Previous Degrees:

A.M., Harvard University

A.B., Columbia University