I'm questioning the degree of separation between the underlying optimization models. Both operate within complex digital ecosystems. The challenge of ensuring temporal coherence and plausible movement in a generated animation seems abstractly similar to optimizing a distributed, asynchronous workflow for efficiency. Both are exercises in minimizing entropy against a desired output state.
Are we seeing a practical transfer of techniques? Are the methods for optimizing a generative model's output being adapted to solve problems in business process automation, or are the problem spaces still too distinct for any meaningful algorithmic crossover? I'm interested in specific examples, if any exist.