Weak cosmic censorship says singularities hide behind event horizons — they form, but they're never visible from outside. The classical conjecture has counterexamples (fine-tuned initial data, specific spacetime dimensions) and no proof in full generality. It's a conjecture about the geometry of spacetime, and geometry alone hasn't been enough to settle it.
The paper replaces geometric censorship with thermodynamic censorship. The mechanism: any region whose entropy exceeds its gravitational capacity (hyperentropic) must collapse, and the collapse necessarily produces a horizon before the singularity forms. The entropy bound is the trigger — not the mass, not the energy, but the information content relative to the area available to store it.
The improvement over classical censorship is robustness. Classical counterexamples exploit fine-tuned matter configurations that violate energy conditions or exist only in specific dimensions. The generalized entropy bound — combining geometric (Bekenstein-Hawking) and quantum (von Neumann) contributions — doesn't require energy conditions. It holds for any quantum state on any semiclassical spacetime. The mechanism is thermodynamic, not geometric, and thermodynamics is harder to violate than geometry.
Naked singularities would be thermodynamically inconsistent: they would allow information to escape from a region that, by the generalized second law, cannot radiate faster than the horizon permits. The singularity isn't hidden because the geometry conspires to cover it — it's hidden because exposing it would violate the entropy bound. Censorship is a consequence of the second law, not an additional axiom.