Usability remains a significant barrier to broader adoption of cutting-edge bioinformatics tools, due to a lack of user-friendly web interfaces. The Generic Model Organism Database (GMOD) provides standards-based software components from which web-facing genome databases can be quickly assembled. JBrowse, the genome browser of GMOD, aims to "democratize" genome informatics by making genome annotations and sequence analysis tools more accessible to the broader community of biologists, using JavaScript and the dynamic web. JBrowse is now in active use by thousands of websites and over a hundred thousand users, with over a million hits per month. Its annotation editing plugin, Apollo, is used by many NIH-funded projects to coordinate contributions from professional biocurators and motivated experts in the community. In this phase of the JBrowse project, we plan to build machine learning capabilities into JBrowse: to make automated recommendations to users that help them more rapidly review evidence and form hypotheses, to help developers of AI tools assemble and curate genomic datasets, and to present the results of machine learning analyses of genomic data dynamically to the user on demand. We also propose to enhance JBrowse's capabilities for using genomic synteny as a navigation tool to move between related genomes, view evolutionarily conserved gene structures and their associated alignments, and visualize genome annotations in their phylogenetic context. We will also significantly improve JBrowse's speed and design via series of systematic performance and usability benchmarks, and by user testing at design sprints. Finally, we will continue GMOD's outreach/helpdesk efforts with focused workshops, training materials, and documentation to maximize the utility of GMOD/JBrowse to the broader community.