Generative AI is rapidly reshaping software engineering (SE) research. While widely seen as a source of renewed momentum, enabling new tools, revitalizing established problems, and increasing research visibility, it also disrupts long-standing norms around rigor, evaluation, authorship, and responsibility. In this position paper, we analyze open-ended responses from a community survey conducted ahead of ICSE 2026 FOSE to capture how SE researchers perceive which aspects of GenAI adoption are working well and bringing joy, and which are creating stress or friction. Our findings show that researchers do not call for restricting GenAI use, but for clearer norms, SE-specific evaluation standards, stronger expectations around rigor and reproducibility, and greater attention to fairness in access.