Post #1931645

@dotjayne@tech.lgbt Remember: Generative neural network models work by observing random noisy data and figuring out how to interpret that as an output similar to their training data. That means everything they output is a hallucination. And sometimes, the hallucination happens to be correct.