Thanks for sharing that experiment with us, Matt. I think it resonates quite well with our study (https://doi.org/10.1016/j.chbah.2024.100095) about reviewer perceptions of AI-generated text output. Now that we have Claude projects and custom styles, I bet the output would be even better. Would be nice to see you run this experiment again in a year. If you're interested in teaming up with a research team, we're game.