Five domains. One question: what is actually measurable here?

The problem of proving what an automated system did, believed, or committed to, at a standard that holds under adversarial and evidentiary pressure. As machine agents take irreversible actions, the gap between what a system reports and what can be shown about it becomes the load-bearing risk.

Operational theories of temporal passage, state-space pruning, and measurement as irreversible constraint. The work treats the arrow of time not as a given but as an accounting problem: what is spent when a possibility becomes an actuality, and whether that expenditure is what we experience as duration.

Finite-time dynamical Casimir simulation, parametric amplification, Bogoliubov mode recovery, and Floquet-Mathieu stability structure, developed as one validated chain. The same mathematics recurs where it should not: across plasma boundaries, a class of deep-network training instabilities, and vacuum fluctuation. This domain holds the lab's one validated numerical testbed.

Upper-bound analysis on biological quantum coherence, optical collectivity in tryptophan networks, filter theories of consciousness from James forward, and falsifiable predictions distinguishing quantum-coherent from classical substrates in artificial systems.

The lab treats consciousness claims as an arena for bound-setting and falsification. It scores its no-go results here above its speculative ones, and makes no claim that any of these mechanisms is realized.

Conservative constraint-extraction from disputed, observer-sensitive, and under-measured empirical corpora; channel-parameter estimation; decomposition of complex claims into testable dependencies and no-go bounds. This is where the lab is most careful and most easily misread.

Studying a corpus is not endorsing its claims. It is subjecting them to instruments they were never built to survive. No result in this domain is promoted above speculative without an instrument, a kill criterion, and a pre-registered prediction.