Offline rendering's trajectory from rasterization to pure ray tracing predicts interactive graphics' future

Offline rendering (movies, VFX) went through a similar evolution: early films used rasterization-based algorithms (REYES, A-buffer), then added limited ray tracing for reflections and shadows, then extended to path-traced global illumination, and finally switched entirely to ray tracing for primary visibility too — because once enough rays are available for everything else, adding one more ray for primary visibility is negligible overhead. The instructor draws this parallel to predict that interactive rendering will follow the same path over the next decade: as GPU hardware becomes faster and more ray-per-pixel budget grows, rasterization will be phased out in favour of pure ray tracing. This prediction is grounded in the observation that ray tracing is objectively superior for image quality; rasterization persists only because dedicated hardware makes it far faster currently.

Examples

Pixar’s RenderMan moved from REYES (rasterization-based) to path tracing. When 1,000 rays/pixel are feasible in real-time, the cost of primary-visibility ray tracing (~1 ray) becomes negligible compared to secondary rays already being traced.

Assessment

Explain the analogy between offline and interactive rendering’s historical trajectory. At what ray-per-pixel budget would it make sense to switch primary visibility from rasterization to ray tracing in a game engine?