source: simon willison: open source ai gap map
level: technical
current ai, a nonprofit founded at the 2025 paris ai action summit with $400m in committed capital, published its gap map v0.1. the map indexes 421 open source ai products in depth, including 266 software tools and libraries, 85 models, 50 datasets, and 20 hardware projects from 228 organizations. these are organized into 14 categories across three layers: model components, product/ux, and infrastructure. an additional 24,400 artifacts remain uncategorized and unscored until further research.
the underlying data is released under an mit license in the currentai-org/os-ai-map github repository. it contains 1,184 yaml files plus notebooks, schemas, and scripts used for data gathering. the project also tracks 16,185 github repos, which can be explored as a csv file using datasette lite. this open data approach allows developers and researchers to build their own analyses on top of the map.
the gap map aims to provide a comprehensive view of the open source ai ecosystem, highlighting both mature tools and gaps where more development is needed. by making the data freely available, current ai encourages community contributions to improve the map's accuracy and coverage over time. the initiative reflects growing interest in public alternatives to proprietary ai systems.
why it matters: it gives ai developers and researchers a structured, open dataset to identify missing pieces in the open source ai stack and guide investment or contribution.