AIs@home runs open-weight language models on volunteers' GPUs, so that open models and the science they serve do not depend on a few large companies. This BOINC project is the way to contribute a 24 GB NVIDIA card: your client runs a pilot that serves model requests for a few hours. It is a private alpha: joining needs an invitation code. Read more, and manage your settings, on aisathome.org.
Already joined? Log in.
