A narrowing of AI research? https://arxiv.org/abs/2009.10385 Juan Mateos Garcia Thematic diversity in AI has been decreasing. Is this universal for other breakthroughs?