What if real-time mod menus let researchers toggle between competing theoretical models during experiments?

This is a thoughtful thought experiment, but it's asking about a hypothetical tool rather than a physics question — and the archive doesn't contain research on real-time theory-switching interfaces during experiments.

That said, the *underlying tension* you're probing — how to test competing models efficiently — does appear in the archive. Here's what's relevant:

On rapid model exploration: The FERMIACC [AG-2026.03-1649] autonomously generates and validates particle-physics hypotheses at scale, which is conceptually similar to what you're imagining: quickly cycling through candidate theories against data. Similarly, meta-design using language models [AG-2024.06-1530] lets researchers explore entire classes of quantum experiment designs in a single pass, then extrapolate without re-optimizing — a kind of lightweight toggling between experimental variants.

On the hidden risk: There's a crucial catch. When you implement a theory in code (your "mod menu"), the implementation quietly smuggles in assumptions that may not match the mathematical model [AG-2024.02-1972]. Even widely-used models can break the causal link between formal theory and results. So a real-time toggle would only be as trustworthy as the fidelity between each model's equations and its code — and that link is often opaque.

On choosing which models to test: Physicists actually *do* systematically pick competing models before experiments settle them — but not randomly. When experiments stall (as in post-Higgs physics), they pursue "ugly" models: either radically novel or easily testable [AG-2025.07-1199]. A smart toggle would need similar heuristics, not just enumerate all possibilities.